need help AMD Radeon™ AI PRO R9700 ROCm debian 12 #20881

aliaj00 · 2026-03-23T03:02:30Z

aliaj00
Mar 23, 2026

hi guys,

has anyone success with the AMD Radeon™ AI PRO R9700

i got an ASUS one from microcenter but it looks like it has issues.

bellow is want i did find so far.

/opt/llama.cpp/build/bin/llama-server --model /opt/Qwen3.5-35B-A3B-Q4_K_M_GGUF/Qwen3.5-35B-A3B-Q4_K_M.gguf -b 1024 -np 1 --fit off --temp 0.6 --top-p 0.95 --top-k 20 --min-p 0.00 --port 8001 --jinja --host 0.0.0.0 --no-warmup
ggml_cuda_init: found 1 ROCm devices (Total VRAM: 32624 MiB):
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32, VRAM: 32624 MiB
build: 8368 (9e2e219) with GNU 12.2.0 for Linux x86_64
system info: n_threads = 16, n_threads_batch = 16, total_threads = 32

Running without SSL
init: using 31 threads for HTTP server
start: binding port with default address family
main: loading model
srv load_model: loading model '/opt/llama-projects/Qwen3.5-35B-A3B-Q4_K_M_GGUF/Qwen3.5-35B-A3B-Q4_K_M.gguf'
llama_model_load_from_file_impl: using device ROCm0 (AMD Radeon AI PRO R9700) (0000:03:00.0) - 32548 MiB free
llama_model_loader: loaded meta data with 52 key-value pairs and 733 tensors from /opt/llama-projects/Qwen3.5-35B-A3B-Q4_K_M_GGUF/Qwen3.5-35B-A3B-Q4_K_M.gguf (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv 0: general.architecture str = qwen35moe
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.sampling.top_k i32 = 20
llama_model_loader: - kv 3: general.sampling.top_p f32 = 0.950000
llama_model_loader: - kv 4: general.sampling.temp f32 = 1.000000
llama_model_loader: - kv 5: general.name str = Qwen3.5-35B-A3B
llama_model_loader: - kv 6: general.basename str = Qwen3.5-35B-A3B
llama_model_loader: - kv 7: general.quantized_by str = Unsloth
llama_model_loader: - kv 8: general.size_label str = 35B-A3B
llama_model_loader: - kv 9: general.license str = apache-2.0
llama_model_loader: - kv 10: general.license.link str = https://huggingface.co/Qwen/Qwen3.5-3...
llama_model_loader: - kv 11: general.repo_url str = https://huggingface.co/unsloth
llama_model_loader: - kv 12: general.base_model.count u32 = 1
llama_model_loader: - kv 13: general.base_model.0.name str = Qwen3.5 35B A3B
llama_model_loader: - kv 14: general.base_model.0.organization str = Qwen
llama_model_loader: - kv 15: general.base_model.0.repo_url str = https://huggingface.co/Qwen/Qwen3.5-3...
llama_model_loader: - kv 16: general.tags arr[str,2] = ["unsloth", "image-text-to-text"]
llama_model_loader: - kv 17: qwen35moe.block_count u32 = 40
llama_model_loader: - kv 18: qwen35moe.context_length u32 = 262144
llama_model_loader: - kv 19: qwen35moe.embedding_length u32 = 2048
llama_model_loader: - kv 20: qwen35moe.attention.head_count u32 = 16
llama_model_loader: - kv 21: qwen35moe.attention.head_count_kv u32 = 2
llama_model_loader: - kv 22: qwen35moe.rope.dimension_sections arr[i32,4] = [11, 11, 10, 0]
llama_model_loader: - kv 23: qwen35moe.rope.freq_base f32 = 10000000.000000
llama_model_loader: - kv 24: qwen35moe.attention.layer_norm_rms_epsilon f32 = 0.000001
llama_model_loader: - kv 25: qwen35moe.expert_count u32 = 256
llama_model_loader: - kv 26: qwen35moe.expert_used_count u32 = 8
llama_model_loader: - kv 27: qwen35moe.attention.key_length u32 = 256
llama_model_loader: - kv 28: qwen35moe.attention.value_length u32 = 256
llama_model_loader: - kv 29: qwen35moe.expert_feed_forward_length u32 = 512
llama_model_loader: - kv 30: qwen35moe.expert_shared_feed_forward_length u32 = 512
llama_model_loader: - kv 31: qwen35moe.ssm.conv_kernel u32 = 4
llama_model_loader: - kv 32: qwen35moe.ssm.state_size u32 = 128
llama_model_loader: - kv 33: qwen35moe.ssm.group_count u32 = 16
llama_model_loader: - kv 34: qwen35moe.ssm.time_step_rank u32 = 32
llama_model_loader: - kv 35: qwen35moe.ssm.inner_size u32 = 4096
llama_model_loader: - kv 36: qwen35moe.full_attention_interval u32 = 4
llama_model_loader: - kv 37: qwen35moe.rope.dimension_count u32 = 64
llama_model_loader: - kv 38: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 39: tokenizer.ggml.pre str = qwen35
llama_model_loader: - kv 40: tokenizer.ggml.tokens arr[str,248320] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 41: tokenizer.ggml.token_type arr[i32,248320] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 42: tokenizer.ggml.merges arr[str,247587] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv 43: tokenizer.ggml.eos_token_id u32 = 248046
llama_model_loader: - kv 44: tokenizer.ggml.padding_token_id u32 = 248055
llama_model_loader: - kv 45: tokenizer.chat_template str = {%- set image_count = namespace(value...
llama_model_loader: - kv 46: general.quantization_version u32 = 2
llama_model_loader: - kv 47: general.file_type u32 = 15
llama_model_loader: - kv 48: quantize.imatrix.file str = Qwen3.5-35B-A3B-GGUF/imatrix_unsloth....
llama_model_loader: - kv 49: quantize.imatrix.dataset str = unsloth_calibration_Qwen3.5-35B-A3B.txt
llama_model_loader: - kv 50: quantize.imatrix.entries_count u32 = 510
llama_model_loader: - kv 51: quantize.imatrix.chunks_count u32 = 80
llama_model_loader: - type f32: 301 tensors
llama_model_loader: - type q8_0: 60 tensors
llama_model_loader: - type q4_K: 165 tensors
llama_model_loader: - type q5_K: 60 tensors
llama_model_loader: - type q6_K: 67 tensors
llama_model_loader: - type mxfp4: 80 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type = Q4_K - Medium
print_info: file size = 19.74 GiB (4.89 BPW)
load: 0 unused tokens
load: printing all EOG tokens:
load: - 248044 ('<|endoftext|>')
load: - 248046 ('<|im_end|>')
load: - 248063 ('<|fim_pad|>')
load: - 248064 ('<|repo_name|>')
load: - 248065 ('<|file_sep|>')
load: special tokens cache size = 33
load: token to piece cache size = 1.7581 MB
print_info: arch = qwen35moe
print_info: vocab_only = 0
print_info: no_alloc = 0
print_info: n_ctx_train = 262144
print_info: n_embd = 2048
print_info: n_embd_inp = 2048
print_info: n_layer = 40
print_info: n_head = 16
print_info: n_head_kv = 2
print_info: n_rot = 64
print_info: n_swa = 0
print_info: is_swa_any = 0
print_info: n_embd_head_k = 256
print_info: n_embd_head_v = 256
print_info: n_gqa = 8
print_info: n_embd_k_gqa = 512
print_info: n_embd_v_gqa = 512
print_info: f_norm_eps = 0.0e+00
print_info: f_norm_rms_eps = 1.0e-06
print_info: f_clamp_kqv = 0.0e+00
print_info: f_max_alibi_bias = 0.0e+00
print_info: f_logit_scale = 0.0e+00
print_info: f_attn_scale = 0.0e+00
print_info: n_ff = 0
print_info: n_expert = 256
print_info: n_expert_used = 8
print_info: n_expert_groups = 0
print_info: n_group_used = 0
print_info: causal attn = 1
print_info: pooling type = 0
print_info: rope type = 40
print_info: rope scaling = linear
print_info: freq_base_train = 10000000.0
print_info: freq_scale_train = 1
print_info: n_ctx_orig_yarn = 262144
print_info: rope_yarn_log_mul = 0.0000
print_info: rope_finetuned = unknown
print_info: mrope sections = [11, 11, 10, 0]
print_info: ssm_d_conv = 4
print_info: ssm_d_inner = 4096
print_info: ssm_d_state = 128
print_info: ssm_dt_rank = 32
print_info: ssm_n_group = 16
print_info: ssm_dt_b_c_rms = 0
print_info: model type = 35B.A3B
print_info: model params = 34.66 B
print_info: general.name = Qwen3.5-35B-A3B
print_info: vocab type = BPE
print_info: n_vocab = 248320
print_info: n_merges = 247587
print_info: BOS token = 11 ','
print_info: EOS token = 248046 '<|im_end|>'
print_info: EOT token = 248046 '<|im_end|>'
print_info: PAD token = 248055 '<|vision_pad|>'
print_info: LF token = 198 'Ċ'
print_info: FIM PRE token = 248060 '<|fim_prefix|>'
print_info: FIM SUF token = 248062 '<|fim_suffix|>'
print_info: FIM MID token = 248061 '<|fim_middle|>'
print_info: FIM PAD token = 248063 '<|fim_pad|>'
print_info: FIM REP token = 248064 '<|repo_name|>'
print_info: FIM SEP token = 248065 '<|file_sep|>'
print_info: EOG token = 248044 '<|endoftext|>'
print_info: EOG token = 248046 '<|im_end|>'
print_info: EOG token = 248063 '<|fim_pad|>'
print_info: EOG token = 248064 '<|repo_name|>'
print_info: EOG token = 248065 '<|file_sep|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = true, direct_io = false)
load_tensors: offloading output layer to GPU
load_tensors: offloading 39 repeating layers to GPU
load_tensors: offloaded 41/41 layers to GPU
load_tensors: CPU_Mapped model buffer size = 272.81 MiB
load_tensors: ROCm0 model buffer size = 19939.68 MiB
..................................................................................................
common_init_result: added <|endoftext|> logit bias = -inf
common_init_result: added <|im_end|> logit bias = -inf
common_init_result: added <|fim_pad|> logit bias = -inf
common_init_result: added <|repo_name|> logit bias = -inf
common_init_result: added <|file_sep|> logit bias = -inf
llama_context: constructing llama_context
llama_context: n_seq_max = 1
llama_context: n_ctx = 262144
llama_context: n_ctx_seq = 262144
llama_context: n_batch = 1024
llama_context: n_ubatch = 512
llama_context: causal_attn = 1
llama_context: flash_attn = auto
llama_context: kv_unified = false
llama_context: freq_base = 10000000.0
llama_context: freq_scale = 1
llama_context: ROCm_Host output buffer size = 0.95 MiB
llama_kv_cache: ROCm0 KV buffer size = 5120.00 MiB
llama_kv_cache: size = 5120.00 MiB (262144 cells, 10 layers, 1/1 seqs), K (f16): 2560.00 MiB, V (f16): 2560.00 MiB
llama_memory_recurrent: ROCm0 RS buffer size = 62.81 MiB
llama_memory_recurrent: size = 62.81 MiB ( 1 cells, 40 layers, 1 seqs), R (f32): 2.81 MiB, S (f32): 60.00 MiB
sched_reserve: reserving ...
sched_reserve: Flash Attention was auto, set to enabled
sched_reserve: resolving fused Gated Delta Net support:
sched_reserve: fused Gated Delta Net (autoregressive) enabled
sched_reserve: fused Gated Delta Net (chunked) enabled
sched_reserve: ROCm0 compute buffer size = 804.02 MiB
sched_reserve: ROCm_Host compute buffer size = 520.02 MiB
sched_reserve: graph nodes = 3729
sched_reserve: graph splits = 2
sched_reserve: reserve took 77.28 ms, sched copies = 1
srv load_model: initializing slots, n_slots = 1
Segmentation fault

/opt/rocm/bin/rocminfo
ROCk module version 6.16.13 is loaded

HSA System Attributes

Runtime Version: 1.18
Runtime Ext Version: 1.15
System Timestamp Freq.: 1000.000000MHz
Sig. Max Wait Duration: 18446744073709551615 (0xFFFFFFFFFFFFFFFF) (timestamp count)
Machine Model: LARGE
System Endianness: LITTLE
Mwaitx: DISABLED
XNACK enabled: NO
DMAbuf Support: YES
VMM Support: YES

==========
HSA Agents

Agent 1

Name: AMD Ryzen 9 9950X3D 16-Core Processor
Uuid: CPU-XX
Marketing Name: AMD Ryzen 9 9950X3D 16-Core Processor
Vendor Name: CPU
Feature: None specified
Profile: FULL_PROFILE
Float Round Mode: NEAR
Max Queue Number: 0(0x0)
Queue Min Size: 0(0x0)
Queue Max Size: 0(0x0)
Queue Type: MULTI
Node: 0
Device Type: CPU
Cache Info:
L1: 49152(0xc000) KB
Chip ID: 0(0x0)
ASIC Revision: 0(0x0)
Cacheline Size: 64(0x40)
Max Clock Freq. (MHz): 4300
BDFID: 0
Internal Node ID: 0
Compute Unit: 32
SIMDs per CU: 0
Shader Engines: 0
Shader Arrs. per Eng.: 0
WatchPts on Addr. Ranges:1
Memory Properties:
Features: None
Pool Info:
Pool 1
Segment: GLOBAL; FLAGS: FINE GRAINED
Size: 261395552(0xf949460) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Recommended Granule:4KB
Alloc Alignment: 4KB
Accessible by all: TRUE
Pool 2
Segment: GLOBAL; FLAGS: EXTENDED FINE GRAINED
Size: 261395552(0xf949460) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Recommended Granule:4KB
Alloc Alignment: 4KB
Accessible by all: TRUE
Pool 3
Segment: GLOBAL; FLAGS: KERNARG, FINE GRAINED
Size: 261395552(0xf949460) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Recommended Granule:4KB
Alloc Alignment: 4KB
Accessible by all: TRUE
Pool 4
Segment: GLOBAL; FLAGS: COARSE GRAINED
Size: 261395552(0xf949460) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Recommended Granule:4KB
Alloc Alignment: 4KB
Accessible by all: TRUE
ISA Info:

Agent 2

Name: gfx1201
Uuid: GPU-b44207ff2cd402f4
Marketing Name: AMD Radeon AI PRO R9700
Vendor Name: AMD
Feature: KERNEL_DISPATCH
Profile: BASE_PROFILE
Float Round Mode: NEAR
Max Queue Number: 128(0x80)
Queue Min Size: 64(0x40)
Queue Max Size: 131072(0x20000)
Queue Type: MULTI
Node: 1
Device Type: GPU
Cache Info:
L1: 32(0x20) KB
L2: 8192(0x2000) KB
L3: 65536(0x10000) KB
Chip ID: 30033(0x7551)
ASIC Revision: 1(0x1)
Cacheline Size: 256(0x100)
Max Clock Freq. (MHz): 2350
BDFID: 768
Internal Node ID: 1
Compute Unit: 64
SIMDs per CU: 2
Shader Engines: 4
Shader Arrs. per Eng.: 2
WatchPts on Addr. Ranges:4
Coherent Host Access: FALSE
Memory Properties:
Features: KERNEL_DISPATCH
Fast F16 Operation: TRUE
Wavefront Size: 32(0x20)
Workgroup Max Size: 1024(0x400)
Workgroup Max Size per Dimension:
x 1024(0x400)
y 1024(0x400)
z 1024(0x400)
Max Waves Per CU: 32(0x20)
Max Work-item Per CU: 1024(0x400)
Grid Max Size: 4294967295(0xffffffff)
Grid Max Size per Dimension:
x 2147483647(0x7fffffff)
y 65535(0xffff)
z 65535(0xffff)
Max fbarriers/Workgrp: 32
Packet Processor uCode:: 128
SDMA engine uCode:: 662
IOMMU Support:: None
Pool Info:
Pool 1
Segment: GLOBAL; FLAGS: COARSE GRAINED
Size: 33406976(0x1fdc000) KB
Allocatable: TRUE
Alloc Granule: 4KB
Alloc Recommended Granule:2048KB
Alloc Alignment: 4KB
Accessible by all: FALSE
Pool 2
Segment: GROUP
Size: 64(0x40) KB
Allocatable: FALSE
Alloc Granule: 0KB
Alloc Recommended Granule:0KB
Alloc Alignment: 0KB
Accessible by all: FALSE
ISA Info:
ISA 1
Name: amdgcn-amd-amdhsa--gfx1201
Machine Models: HSA_MACHINE_MODEL_LARGE
Profiles: HSA_PROFILE_BASE
Default Rounding Mode: NEAR
Default Rounding Mode: NEAR
Fast f16: TRUE
Workgroup Max Size: 1024(0x400)
Workgroup Max Size per Dimension:
x 1024(0x400)
y 1024(0x400)
z 1024(0x400)
Grid Max Size: 4294967295(0xffffffff)
Grid Max Size per Dimension:
x 2147483647(0x7fffffff)
y 65535(0xffff)
z 65535(0xffff)
FBarrier Max Size: 32
ISA 2
Name: amdgcn-amd-amdhsa--gfx12-generic
Machine Models: HSA_MACHINE_MODEL_LARGE
Profiles: HSA_PROFILE_BASE
Default Rounding Mode: NEAR
Default Rounding Mode: NEAR
Fast f16: TRUE
Workgroup Max Size: 1024(0x400)
Workgroup Max Size per Dimension:
x 1024(0x400)
y 1024(0x400)
z 1024(0x400)
Grid Max Size: 4294967295(0xffffffff)
Grid Max Size per Dimension:
x 2147483647(0x7fffffff)
y 65535(0xffff)
z 65535(0xffff)
FBarrier Max Size: 32
*** Done ***

rocm-smi --showuse --showclocks --showpower

============================ ROCm System Management Interface ============================
WARNING: AMD GPU device(s) is/are in a low-power state. Check power control/runtime_status

=============================== Current clock frequencies ================================
GPU[0] : dcefclk clock level: 1: (219Mhz)
GPU[0] : fclk clock level: 1: (582Mhz)
GPU[0] : mclk clock level: 0: (96Mhz)
GPU[0] : sclk clock level: 1: (59Mhz)
GPU[0] : socclk clock level: 0: (417Mhz)
GPU[0] : pcie clock level: 2 (32.0GT/s x16)
GPU[1] : mclk clock level: 0: (1800Mhz)
GPU[1] : sclk clock level: 0: (600Mhz)
GPU[1] : socclk clock level: 1: (1200Mhz)

=================================== Power Consumption ====================================
GPU[0] : Average Graphics Package Power (W): 15.0
GPU[1] : Current Socket Graphics Package Power (W): 0.012

=================================== % time GPU is busy ===================================
GPU[0] : GPU use (%): 2
GPU[1] : GPU use (%): 0

================================== End of ROCm SMI Log ===================================

echo on | sudo tee /sys/bus/pci/devices/0000:03:00.0/power/control

/opt/rocm/bin/rocm-smi

WARNING: AMD GPU device(s) is/are in a low-power state. Check power control/runtime_status

========================================= ROCm System Management Interface =========================================
=================================================== Concise Info ===================================================
Device Node IDs Temp Power Partitions SCLK MCLK Fan Perf PwrCap VRAM% GPU%
(DID, GUID) (Edge) (Avg) (Mem, Compute, ID)

0 1 0x7551, 32448 26.0°C 15.0W N/A, N/A, 0 1Mhz 96Mhz 29.8% auto 300.0W 0% 0%
1 2 0x13c0, 2327 38.0°C 0.012W N/A, N/A, 0 N/A 1800Mhz 0% auto N/A 63% 0%

=============================================== End of ROCm SMI Log ================================================

/opt/rocm/bin/rocm-smi --showproductname --showbus --showuniqueid
ls -l /sys/class/drm/
find /sys -path 'amdgpuruntime_status' -o -path 'drm/card/device/power/runtime_status' 2>/dev/null | xargs -r -I{} sh -c 'echo === {}; cat {}'
dmesg | grep -iE 'amdgpu|kfd|drm' | tail -n 200

============================ ROCm System Management Interface ============================
WARNING: AMD GPU device(s) is/are in a low-power state. Check power control/runtime_status

======================================= Unique ID ========================================
GPU[0] : Unique ID: 0xb44207ff2cd402f4
GPU[1] : Unique ID: 0x0

======================================= PCI Bus ID =======================================
GPU[0] : PCI Bus: 0000:03:00.0
GPU[1] : PCI Bus: 0000:7A:00.0

====================================== Product Info ======================================
GPU[0] : Card Series: AMD Radeon AI PRO R9700
GPU[0] : Card Model: 0x7551
GPU[0] : Card Vendor: Advanced Micro Devices, Inc. [AMD/ATI]
GPU[0] : Card SKU: G287BP00
GPU[0] : Subsystem ID: 0x0626
GPU[0] : Device Rev: 0xc0
GPU[0] : Node ID: 1
GPU[0] : GUID: 32448
GPU[0] : GFX Version: gfx1201
GPU[1] : Card Series: AMD Radeon Graphics
GPU[1] : Card Model: 0x13c0
GPU[1] : Card Vendor: Advanced Micro Devices, Inc. [AMD/ATI]
GPU[1] : Card SKU: RAPHAEL
GPU[1] : Subsystem ID: 0x7e59
GPU[1] : Device Rev: 0xc9
GPU[1] : Node ID: 2
GPU[1] : GUID: 2327
GPU[1] : GFX Version: gfx1036

================================== End of ROCm SMI Log ===================================

hipconfig --full
HIP version: 7.2.26015-fc0010cf6a

==hipconfig
HIP_PATH :/opt/rocm-7.2.0
ROCM_PATH :/opt/rocm-7.2.0
HIP_COMPILER :clang
HIP_PLATFORM :amd
HIP_RUNTIME :rocclr
CPP_CONFIG : -D__HIP_PLATFORM_HCC__= -D__HIP_PLATFORM_AMD__= -I/opt/rocm-7.2.0/include -I/include

==hip-clang
HIP_CLANG_PATH :/opt/rocm-7.2.0/lib/llvm/bin
AMD clang version 22.0.0git (https://github.com/RadeonOpenCompute/llvm-project roc-7.2.0 26014 7b800a19466229b8479a78de19143dc33c3ab9b5)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /opt/rocm-7.2.0/lib/llvm/bin
Configuration file: /opt/rocm-7.2.0/lib/llvm/bin/clang++.cfg
AMD LLVM version 22.0.0git
Optimized build.
Default target: x86_64-unknown-linux-gnu
Host CPU: znver5

Registered Targets:
amdgcn - AMD GCN GPUs
r600 - AMD GPUs HD2XXX-HD6XXX
x86 - 32-bit X86: Pentium-Pro and above
x86-64 - 64-bit X86: EM64T and AMD64
hip-clang-cxxflags :
-O3
hip-clang-ldflags :
--driver-mode=g++ -O3 --hip-link

== Environment Variables
PATH =/root/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin

== Linux Kernel
Hostname :

Linux ******* 6.1.0-42-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.159-1 (2025-12-30) x86_64 GNU/Linux
No LSB modules are available.
Distributor ID: Debian
Description: Debian GNU/Linux 12 (bookworm)
Release: 12
Codename: bookworm

lspci | grep -i vga
03:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Device 7551 (rev c0)
7a:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Granite Ridge [Radeon Graphics] (rev c9)

digitalscream · 2026-03-23T09:01:50Z

digitalscream
Mar 23, 2026

I had a similar problem with my R9700. I can't help you fix it, per se, but I can tell you that the way I solved the problem was to completely uninstall ROCm and just use Vulkan (installed from kisak's repo). Vulkan is far faster than ROCm for the R9700 anyway.

1 reply

aliaj00 Mar 23, 2026
Author

thank you man will give it a try.

JohnTDI-cpu · 2026-03-23T09:39:47Z

JohnTDI-cpu
Mar 23, 2026

I'm running dual R9700s (ASUS ProArt X870E-Creator, Ryzen 9 9900X) and hit the exact same segfault. Your card is stuck in deep sleep — sclk: 1 MHz, mclk: 96 MHz is basically a dead card trying to run inference.
The fix is your software stack. Kernel 6.1 has no proper RDNA4 (gfx1201) support:

Kernel → 6.17+ (Debian backports or manual install from kernel.org)
Mesa/RADV → 25.2.8+ (Debian backports or Kisak/Oibaf-style PPA equivalent for Debian)
Force power state:

echo high | sudo tee /sys/class/drm/card0/device/power_dpm_force_performance_level

Lock mclk to highest level (~1258 MHz):

cat /sys/class/drm/card0/device/pp_dpm_mclk
echo 2 | sudo tee /sys/class/drm/card0/device/pp_dpm_mclk

If still broken, try AMDVLK as fallback instead of RADV.

Also watch out for your Raphael iGPU (gfx1036, GPU[1]) — make sure it's not interfering with power management on the discrete card.
My working setup: Ubuntu 24.04, kernel 6.17.0-14, Mesa 25.2.8, dual R9700 — ~183 t/s on Qwen3-30B-A3B Q4_K_M, no segfaults.

15 replies

digitalscream Mar 23, 2026

I'm on 6.17.0-19-generic, which I think has the fixes. Much weirdness.

JohnTDI-cpu Mar 23, 2026

Test on 6.19

digitalscream Mar 23, 2026

@JohnTDI-cpu - well, that's interesting. Rebuilt my machine with 25.10 and Mesa 26.0.3. Still the same result with the mem clock, but...I'm now hitting 195t/s in Qwen3-Coder-30B-A3B-Q4_K_0. Previously, I could only get 140t/s, or 160t/s with profile_peak.

More interesting-er, is that when I enable profile_peak (which sticks the memory at 1124MHz), I still only get 160t/s.

In any case, at this point the R9700s are beating the RTX 3090s I had before, which is quite something given that they had over 900GB/s VRAM bandwidth.

JohnTDI-cpu Mar 23, 2026

Nice results! You sure that's with RADV? On my dual R9700 setup I get ~194 t/s on Qwen3-30B-A3B Q4_K_M but only with AMDVLK — RADV gave me lower throughput. The tradeoff was that AMDVLK had worse prompt processing speed compared to RADV.

digitalscream Mar 23, 2026

Yup:

$ AMD_VULKAN_ICD=RADV RADV_DEBUG=nocompute ./llama-bench -m /opt/working/models.1/Qwen3-Coder-30B-A3B-Instruct-Q4_0.gguf -ngl 99 -fa 1 -p 512,1024,2048 -n 128 -r 2 --device Vulkan0,Vulkan1
load_backend: loaded RPC backend from /main_data/bin/llama.cpp.vulkan/llama-b8457/libggml-rpc.so
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 2 Vulkan devices:
ggml_vulkan: 0 = AMD Radeon Graphics (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon Graphics (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
load_backend: loaded Vulkan backend from /main_data/bin/llama.cpp.vulkan/llama-b8457/libggml-vulkan.so
load_backend: loaded CPU backend from /main_data/bin/llama.cpp.vulkan/llama-b8457/libggml-cpu-haswell.so
| model                          |       size |     params | backend    | ngl | fa | dev          |            test |                  t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | ------------ | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_0          |  16.18 GiB |    30.53 B | Vulkan     |  99 |  1 | Vulkan0      |           pp512 |     3976.57 ± 378.97 |
| qwen3moe 30B.A3B Q4_0          |  16.18 GiB |    30.53 B | Vulkan     |  99 |  1 | Vulkan0      |          pp1024 |       4279.12 ± 7.62 |
| qwen3moe 30B.A3B Q4_0          |  16.18 GiB |    30.53 B | Vulkan     |  99 |  1 | Vulkan0      |          pp2048 |      4118.39 ± 32.80 |
| qwen3moe 30B.A3B Q4_0          |  16.18 GiB |    30.53 B | Vulkan     |  99 |  1 | Vulkan0      |           tg128 |        195.89 ± 0.88 |
| qwen3moe 30B.A3B Q4_0          |  16.18 GiB |    30.53 B | Vulkan     |  99 |  1 | Vulkan1      |           pp512 |     3957.99 ± 314.64 |
| qwen3moe 30B.A3B Q4_0          |  16.18 GiB |    30.53 B | Vulkan     |  99 |  1 | Vulkan1      |          pp1024 |      4250.37 ± 54.16 |
| qwen3moe 30B.A3B Q4_0          |  16.18 GiB |    30.53 B | Vulkan     |  99 |  1 | Vulkan1      |          pp2048 |      4129.03 ± 21.20 |
| qwen3moe 30B.A3B Q4_0          |  16.18 GiB |    30.53 B | Vulkan     |  99 |  1 | Vulkan1      |           tg128 |        196.86 ± 0.66 |

It's nuts. 140t/s on Qwen3.5 now too, when I was only getting 125t/s-ish before.

I iz happy.

aliaj00 · 2026-03-23T15:06:28Z

aliaj00
Mar 23, 2026
Author

apt install -t trixie-backports linux-image-amd64 linux-headers-amd64

Upgrading:
linux-image-amd64 linux-libc-dev

Installing:
linux-headers-amd64

Installing dependencies:
cpp-14-for-host linux-base-6.19.6+deb13-amd64 linux-binary-6.19.6+deb13-amd64 linux-headers-6.19.6+deb13-common linux-kbuild-6.19.6+deb13
gcc-14-for-host linux-base-amd64 linux-headers-6.19.6+deb13-amd64 linux-image-6.19.6+deb13-amd64 linux-modules-6.19.6+deb13-amd64

Suggested packages:
debian-kernel-handbook linux-doc-6.19

Summary:
Upgrading: 2, Installing: 11, Removing: 0, Not Upgrading: 123
Download size: 187 MB
Space needed: 416 MB / 711 GB available

Notice: Ignoring file 'amdgpu-install_7.2.70200-1_all.deb' in directory '/etc/apt/sources.list.d/' as it has an invalid filename extension
Continue? [Y/n] y

will give this a try maybe it will work

0 replies

aliaj00 · 2026-03-23T15:54:22Z

aliaj00
Mar 23, 2026
Author

still got the segmentation fault and it is on the runtime HIP. switched to vulkan and got 4043 tokens 33s 119.96 t/s

0 replies

aliaj00 · 2026-03-23T16:02:27Z

aliaj00
Mar 23, 2026
Author

followed it till the library but dont know what else to do for the ROCm "llama-server[5227] general protection fault ip:7f82fb269f4b sp:7ffc0152db10 error:0 in libamdhip64.so.7.2.70200[269f4b,7f82fb024000+41c000]"

Mar 23 11:17:50 zeus kernel: [drm] Initialized amdgpu 3.64.0 for 0000:7a:00.0 on minor 1
Mar 23 11:17:50 zeus kernel: fbcon: amdgpudrmfb (fb0) is primary device
Mar 23 11:17:50 zeus kernel: amdgpu 0000:7a:00.0: [drm] REG_WAIT timeout 1us * 100000 tries - optc31_disable_crtc line:145
Mar 23 11:17:50 zeus kernel: amdgpu 0000:7a:00.0: [drm] fb0: amdgpudrmfb frame buffer device
Mar 23 11:17:50 zeus systemd[1]: Starting modprobe@drm.service - Load Kernel Module drm...
Mar 23 11:17:50 zeus systemd[1]: modprobe@drm.service: Deactivated successfully.
Mar 23 11:17:50 zeus systemd[1]: Finished modprobe@drm.service - Load Kernel Module drm.
Mar 23 11:17:51 zeus kernel: snd_hda_intel 0000:03:00.1: bound 0000:03:00.0 (ops amdgpu_dm_audio_component_bind_ops [amdgpu])
Mar 23 11:17:51 zeus kernel: snd_hda_intel 0000:7a:00.1: bound 0000:7a:00.0 (ops amdgpu_dm_audio_component_bind_ops [amdgpu])
Mar 23 11:19:32 zeus kernel: amdgpu 0000:03:00.0: amdgpu: PCIE GART of 512M enabled (table at 0x00000087D6B00000).
Mar 23 11:19:32 zeus kernel: amdgpu 0000:03:00.0: amdgpu: PSP is resuming...
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: GECC is disabled, set amdgpu_ras_enable=1 to enable GECC in next boot cycle if needed
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: RAP: optional rap ta ucode is not available
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: SECUREDISPLAY: optional securedisplay ta ucode is not available
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: SMU is resuming...
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: smu driver if version = 0x0000002e, smu fw if version = 0x00000032, smu fw program = 0, smu fw version = 0x00684b00 (104.75.0)
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: SMU driver if version not matched
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: SMU is resumed successfully!
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: program CP_MES_CNTL : 0x4000000
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: program CP_MES_CNTL : 0xc000000
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: [drm] DMUB hardware initialized: version=0x0A000601
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: [drm] Cannot find any crtc or sizes
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring sdma0 uses VM inv eng 9 on hub 0
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring sdma1 uses VM inv eng 10 on hub 0
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring vcn_unified_0 uses VM inv eng 0 on hub 8
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring jpeg_dec uses VM inv eng 1 on hub 8
Mar 23 11:19:33 zeus kernel: amdgpu 0000:03:00.0: [drm] Cannot find any crtc or sizes
Mar 23 11:22:02 zeus kernel: amdgpu 0000:03:00.0: amdgpu: PCIE GART of 512M enabled (table at 0x00000087D6B00000).
Mar 23 11:22:02 zeus kernel: amdgpu 0000:03:00.0: amdgpu: PSP is resuming...
Mar 23 11:22:02 zeus kernel: amdgpu 0000:03:00.0: amdgpu: GECC is disabled, set amdgpu_ras_enable=1 to enable GECC in next boot cycle if needed
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: RAP: optional rap ta ucode is not available
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: SECUREDISPLAY: optional securedisplay ta ucode is not available
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: SMU is resuming...
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: smu driver if version = 0x0000002e, smu fw if version = 0x00000032, smu fw program = 0, smu fw version = 0x00684b00 (104.75.0)
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: SMU driver if version not matched
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: SMU is resumed successfully!
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: program CP_MES_CNTL : 0x4000000
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: program CP_MES_CNTL : 0xc000000
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: [drm] DMUB hardware initialized: version=0x0A000601
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: [drm] Cannot find any crtc or sizes
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring sdma0 uses VM inv eng 9 on hub 0
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring sdma1 uses VM inv eng 10 on hub 0
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring vcn_unified_0 uses VM inv eng 0 on hub 8
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: amdgpu: ring jpeg_dec uses VM inv eng 1 on hub 8
Mar 23 11:22:03 zeus kernel: amdgpu 0000:03:00.0: [drm] Cannot find any crtc or sizes
Mar 23 11:24:46 zeus kernel: traps: llama-server[5227] general protection fault ip:7f82fb269f4b sp:7ffc0152db10 error:0 in libamdhip64.so.7.2.70200[269f4b,7f82fb024000+41c000]
Mar 23 11:24:58 zeus kernel: traps: llama-server[5720] general protection fault ip:7f9329e69f4b sp:7ffdf99541c0 error:0 in libamdhip64.so.7.2.70200[269f4b,7f9329c24000+41c000]
Mar 23 11:27:47 zeus kernel: traps: llama-server[7169] general protection fault ip:7faf4c069f4b sp:7ffeddc8db70 error:0 in libamdhip64.so.7.2.70200[269f4b,7faf4be24000+41c000]
Mar 23 11:28:24 zeus kernel: traps: llama-server[7516] general protection fault ip:7efc08a69f4b sp:7ffe32ee5130 error:0 in libamdhip64.so.7.2.70200[269f4b,7efc08824000+41c000]
root@zeus:/home/tim# ldd /opt/llama.cpp/build/bin/llama-server | grep libamdhip64
libamdhip64.so.7 => /opt/rocm-7.2.0/lib/libamdhip64.so.7 (0x00007f9706800000)
root@zeus:/home/tim# gdb --args /opt/llama.cpp/build/bin/llama-server --model /opt/llama-projects/Qwen3.5-35B-A3B-Q4_K_M_GGUF/Qwen3.5-35B-A3B-Q4_K_M.gguf -b 1024 -np 1 --fit on --temp 0.6 --top-p 0.95 --top-k 20 --min-p 0.00 --port 8001 --jinja --host 0.0.0.0
GNU gdb (Debian 16.3-1) 16.3
Copyright (C) 2024 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later http://gnu.org/licenses/gpl.html
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
https://www.gnu.org/software/gdb/bugs/.
Find the GDB manual and other documentation resources online at:
http://www.gnu.org/software/gdb/documentation/.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /opt/llama.cpp/build/bin/llama-server...
(No debugging symbols found in /opt/llama.cpp/build/bin/llama-server)
(gdb) run
Starting program: /opt/llama.cpp/build/bin/llama-server --model /opt/llama-projects/Qwen3.5-35B-A3B-Q4_K_M_GGUF/Qwen3.5-35B-A3B-Q4_K_M.gguf -b 1024 -np 1 --fit on --temp 0.6 --top-p 0.95 --top-k 20 --min-p 0.00 --port 8001 --jinja --host 0.0.0.0
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x7fffa9bff6c0 (LWP 9067)]
[New Thread 0x7fffa11ff6c0 (LWP 9068)]
[New Thread 0x7fffa90ff6c0 (LWP 9069)]
[Thread 0x7fffa90ff6c0 (LWP 9069) exited]
ggml_cuda_init: found 1 ROCm devices (Total VRAM: 32624 MiB):
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32, VRAM: 32624 MiB
[New Thread 0x7fffa3fff6c0 (LWP 9070)]
[Thread 0x7fffa3fff6c0 (LWP 9070) exited]
[New Thread 0x7fffa3fff6c0 (LWP 9071)]
build: 8368 (9e2e219) with GNU 12.2.0 for Linux x86_64
system info: n_threads = 16, n_threads_batch = 16, total_threads = 32

Running without SSL
init: using 31 threads for HTTP server
start: binding port with default address family
[New Thread 0x7fffa37fe6c0 (LWP 9072)]
[New Thread 0x7fffa2ffd6c0 (LWP 9073)]
[New Thread 0x7fffa27fc6c0 (LWP 9074)]
[New Thread 0x7fffa1ffb6c0 (LWP 9075)]
[New Thread 0x7fffa09fe6c0 (LWP 9076)]
[New Thread 0x7ffe9bdff6c0 (LWP 9077)]
[New Thread 0x7ffe9b5fe6c0 (LWP 9078)]
[New Thread 0x7ffe9adfd6c0 (LWP 9079)]
[New Thread 0x7ffe9a5fc6c0 (LWP 9080)]
[New Thread 0x7ffe99dfb6c0 (LWP 9081)]
[New Thread 0x7ffe995fa6c0 (LWP 9082)]
[New Thread 0x7ffe98df96c0 (LWP 9083)]
[New Thread 0x7ffe93fff6c0 (LWP 9084)]
[New Thread 0x7ffe937fe6c0 (LWP 9085)]
[New Thread 0x7ffe92ffd6c0 (LWP 9086)]
main: loading model
srv load_model: loading model '/opt/llama-projects/Qwen3.5-35B-A3B-Q4_K_M_GGUF/Qwen3.5-35B-A3B-Q4_K_M.gguf'
common_init_result: fitting params to device memory, for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on
[New Thread 0x7ffe927fc6c0 (LWP 9087)]
[New Thread 0x7ffe91ffb6c0 (LWP 9088)]
[New Thread 0x7ffe910656c0 (LWP 9089)]
[New Thread 0x7ffe908646c0 (LWP 9090)]
[New Thread 0x7ffe8bfff6c0 (LWP 9091)]
[New Thread 0x7ffe8b7fe6c0 (LWP 9092)]
[New Thread 0x7ffe8affd6c0 (LWP 9093)]
[New Thread 0x7ffe8a7fc6c0 (LWP 9094)]
[New Thread 0x7ffe89ffb6c0 (LWP 9095)]
[New Thread 0x7ffe897fa6c0 (LWP 9096)]
[New Thread 0x7ffe88ff96c0 (LWP 9097)]
[New Thread 0x7ffe887f86c0 (LWP 9098)]
[New Thread 0x7ffe87ff76c0 (LWP 9099)]
[New Thread 0x7ffe877f66c0 (LWP 9100)]
[New Thread 0x7ffe86ff56c0 (LWP 9101)]
[New Thread 0x7ffe867f46c0 (LWP 9102)]
[New Thread 0x7ffe85ff36c0 (LWP 9103)]
llama_params_fit_impl: projected to use 25926 MiB of device memory vs. 32336 MiB of free device memory
llama_params_fit_impl: will leave 6409 >= 1024 MiB of free device memory, no changes needed
llama_params_fit: successfully fit params to free device memory
llama_params_fit: fitting params to free memory took 0.23 seconds
llama_model_load_from_file_impl: using device ROCm0 (AMD Radeon AI PRO R9700) (0000:03:00.0) - 32400 MiB free
llama_model_loader: loaded meta data with 52 key-value pairs and 733 tensors from /opt/llama-projects/Qwen3.5-35B-A3B-Q4_K_M_GGUF/Qwen3.5-35B-A3B-Q4_K_M.gguf (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv 0: general.architecture str = qwen35moe
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.sampling.top_k i32 = 20
llama_model_loader: - kv 3: general.sampling.top_p f32 = 0.950000
llama_model_loader: - kv 4: general.sampling.temp f32 = 1.000000
llama_model_loader: - kv 5: general.name str = Qwen3.5-35B-A3B
llama_model_loader: - kv 6: general.basename str = Qwen3.5-35B-A3B
llama_model_loader: - kv 7: general.quantized_by str = Unsloth
llama_model_loader: - kv 8: general.size_label str = 35B-A3B
llama_model_loader: - kv 9: general.license str = apache-2.0
llama_model_loader: - kv 10: general.license.link str = https://huggingface.co/Qwen/Qwen3.5-3...
llama_model_loader: - kv 11: general.repo_url str = https://huggingface.co/unsloth
llama_model_loader: - kv 12: general.base_model.count u32 = 1
llama_model_loader: - kv 13: general.base_model.0.name str = Qwen3.5 35B A3B
llama_model_loader: - kv 14: general.base_model.0.organization str = Qwen
llama_model_loader: - kv 15: general.base_model.0.repo_url str = https://huggingface.co/Qwen/Qwen3.5-3...
llama_model_loader: - kv 16: general.tags arr[str,2] = ["unsloth", "image-text-to-text"]
llama_model_loader: - kv 17: qwen35moe.block_count u32 = 40
llama_model_loader: - kv 18: qwen35moe.context_length u32 = 262144
llama_model_loader: - kv 19: qwen35moe.embedding_length u32 = 2048
llama_model_loader: - kv 20: qwen35moe.attention.head_count u32 = 16
llama_model_loader: - kv 21: qwen35moe.attention.head_count_kv u32 = 2
llama_model_loader: - kv 22: qwen35moe.rope.dimension_sections arr[i32,4] = [11, 11, 10, 0]
llama_model_loader: - kv 23: qwen35moe.rope.freq_base f32 = 10000000.000000
llama_model_loader: - kv 24: qwen35moe.attention.layer_norm_rms_epsilon f32 = 0.000001
llama_model_loader: - kv 25: qwen35moe.expert_count u32 = 256
llama_model_loader: - kv 26: qwen35moe.expert_used_count u32 = 8
llama_model_loader: - kv 27: qwen35moe.attention.key_length u32 = 256
llama_model_loader: - kv 28: qwen35moe.attention.value_length u32 = 256
llama_model_loader: - kv 29: qwen35moe.expert_feed_forward_length u32 = 512
llama_model_loader: - kv 30: qwen35moe.expert_shared_feed_forward_length u32 = 512
llama_model_loader: - kv 31: qwen35moe.ssm.conv_kernel u32 = 4
llama_model_loader: - kv 32: qwen35moe.ssm.state_size u32 = 128
llama_model_loader: - kv 33: qwen35moe.ssm.group_count u32 = 16
llama_model_loader: - kv 34: qwen35moe.ssm.time_step_rank u32 = 32
llama_model_loader: - kv 35: qwen35moe.ssm.inner_size u32 = 4096
llama_model_loader: - kv 36: qwen35moe.full_attention_interval u32 = 4
llama_model_loader: - kv 37: qwen35moe.rope.dimension_count u32 = 64
llama_model_loader: - kv 38: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 39: tokenizer.ggml.pre str = qwen35
llama_model_loader: - kv 40: tokenizer.ggml.tokens arr[str,248320] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 41: tokenizer.ggml.token_type arr[i32,248320] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 42: tokenizer.ggml.merges arr[str,247587] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv 43: tokenizer.ggml.eos_token_id u32 = 248046
llama_model_loader: - kv 44: tokenizer.ggml.padding_token_id u32 = 248055
llama_model_loader: - kv 45: tokenizer.chat_template str = {%- set image_count = namespace(value...
llama_model_loader: - kv 46: general.quantization_version u32 = 2
llama_model_loader: - kv 47: general.file_type u32 = 15
llama_model_loader: - kv 48: quantize.imatrix.file str = Qwen3.5-35B-A3B-GGUF/imatrix_unsloth....
llama_model_loader: - kv 49: quantize.imatrix.dataset str = unsloth_calibration_Qwen3.5-35B-A3B.txt
llama_model_loader: - kv 50: quantize.imatrix.entries_count u32 = 510
llama_model_loader: - kv 51: quantize.imatrix.chunks_count u32 = 80
llama_model_loader: - type f32: 301 tensors
llama_model_loader: - type q8_0: 60 tensors
llama_model_loader: - type q4_K: 165 tensors
llama_model_loader: - type q5_K: 60 tensors
llama_model_loader: - type q6_K: 67 tensors
llama_model_loader: - type mxfp4: 80 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type = Q4_K - Medium
print_info: file size = 19.74 GiB (4.89 BPW)
load: 0 unused tokens
load: printing all EOG tokens:
load: - 248044 ('<|endoftext|>')
load: - 248046 ('<|im_end|>')
load: - 248063 ('<|fim_pad|>')
load: - 248064 ('<|repo_name|>')
load: - 248065 ('<|file_sep|>')
load: special tokens cache size = 33
load: token to piece cache size = 1.7581 MB
print_info: arch = qwen35moe
print_info: vocab_only = 0
print_info: no_alloc = 0
print_info: n_ctx_train = 262144
print_info: n_embd = 2048
print_info: n_embd_inp = 2048
print_info: n_layer = 40
print_info: n_head = 16
print_info: n_head_kv = 2
print_info: n_rot = 64
print_info: n_swa = 0
print_info: is_swa_any = 0
print_info: n_embd_head_k = 256
print_info: n_embd_head_v = 256
print_info: n_gqa = 8
print_info: n_embd_k_gqa = 512
print_info: n_embd_v_gqa = 512
print_info: f_norm_eps = 0.0e+00
print_info: f_norm_rms_eps = 1.0e-06
print_info: f_clamp_kqv = 0.0e+00
print_info: f_max_alibi_bias = 0.0e+00
print_info: f_logit_scale = 0.0e+00
print_info: f_attn_scale = 0.0e+00
print_info: n_ff = 0
print_info: n_expert = 256
print_info: n_expert_used = 8
print_info: n_expert_groups = 0
print_info: n_group_used = 0
print_info: causal attn = 1
print_info: pooling type = 0
print_info: rope type = 40
print_info: rope scaling = linear
print_info: freq_base_train = 10000000.0
print_info: freq_scale_train = 1
print_info: n_ctx_orig_yarn = 262144
print_info: rope_yarn_log_mul = 0.0000
print_info: rope_finetuned = unknown
print_info: mrope sections = [11, 11, 10, 0]
print_info: ssm_d_conv = 4
print_info: ssm_d_inner = 4096
print_info: ssm_d_state = 128
print_info: ssm_dt_rank = 32
print_info: ssm_n_group = 16
print_info: ssm_dt_b_c_rms = 0
print_info: model type = 35B.A3B
print_info: model params = 34.66 B
print_info: general.name = Qwen3.5-35B-A3B
print_info: vocab type = BPE
print_info: n_vocab = 248320
print_info: n_merges = 247587
print_info: BOS token = 11 ','
print_info: EOS token = 248046 '<|im_end|>'
print_info: EOT token = 248046 '<|im_end|>'
print_info: PAD token = 248055 '<|vision_pad|>'
print_info: LF token = 198 'Ċ'
print_info: FIM PRE token = 248060 '<|fim_prefix|>'
print_info: FIM SUF token = 248062 '<|fim_suffix|>'
print_info: FIM MID token = 248061 '<|fim_middle|>'
print_info: FIM PAD token = 248063 '<|fim_pad|>'
print_info: FIM REP token = 248064 '<|repo_name|>'
print_info: FIM SEP token = 248065 '<|file_sep|>'
print_info: EOG token = 248044 '<|endoftext|>'
print_info: EOG token = 248046 '<|im_end|>'
print_info: EOG token = 248063 '<|fim_pad|>'
print_info: EOG token = 248064 '<|repo_name|>'
print_info: EOG token = 248065 '<|file_sep|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = true, direct_io = false)
load_tensors: offloading output layer to GPU
load_tensors: offloading 39 repeating layers to GPU
load_tensors: offloaded 41/41 layers to GPU
load_tensors: CPU_Mapped model buffer size = 272.81 MiB
load_tensors: ROCm0 model buffer size = 19939.68 MiB
..................................................................................................
common_init_result: added <|endoftext|> logit bias = -inf
common_init_result: added <|im_end|> logit bias = -inf
common_init_result: added <|fim_pad|> logit bias = -inf
common_init_result: added <|repo_name|> logit bias = -inf
common_init_result: added <|file_sep|> logit bias = -inf
llama_context: constructing llama_context
llama_context: n_seq_max = 1
llama_context: n_ctx = 262144
llama_context: n_ctx_seq = 262144
llama_context: n_batch = 1024
llama_context: n_ubatch = 512
llama_context: causal_attn = 1
llama_context: flash_attn = auto
llama_context: kv_unified = false
llama_context: freq_base = 10000000.0
llama_context: freq_scale = 1
llama_context: ROCm_Host output buffer size = 0.95 MiB
llama_kv_cache: ROCm0 KV buffer size = 5120.00 MiB
llama_kv_cache: size = 5120.00 MiB (262144 cells, 10 layers, 1/1 seqs), K (f16): 2560.00 MiB, V (f16): 2560.00 MiB
llama_memory_recurrent: ROCm0 RS buffer size = 62.81 MiB
llama_memory_recurrent: size = 62.81 MiB ( 1 cells, 40 layers, 1 seqs), R (f32): 2.81 MiB, S (f32): 60.00 MiB
sched_reserve: reserving ...
sched_reserve: Flash Attention was auto, set to enabled
sched_reserve: resolving fused Gated Delta Net support:
sched_reserve: fused Gated Delta Net (autoregressive) enabled
sched_reserve: fused Gated Delta Net (chunked) enabled
sched_reserve: ROCm0 compute buffer size = 804.02 MiB
sched_reserve: ROCm_Host compute buffer size = 520.02 MiB
sched_reserve: graph nodes = 3729
sched_reserve: graph splits = 2
sched_reserve: reserve took 58.35 ms, sched copies = 1
common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)

Thread 1 "llama-server" received signal SIGSEGV, Segmentation fault.
0x00007fffee469f4b in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
(gdb) bt
#0 0x00007fffee469f4b in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
#1 0x00007fffee46a7d5 in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
#2 0x00007fffee4ba782 in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
#3 0x00007fffee46b31e in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
#4 0x00007fffee487206 in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
#5 0x00007ffff64be344 in rms_norm_mul_f32_cuda(float const*, float const*, float const*, float*, int, int, int, int, long, long, long, long, long, long, unsigned int, unsigned int, unsigned int, unsigned int, long, long, long, unsigned int, unsigned int, unsigned int, unsigned int, float, ihipStream_t*) () from /opt/llama.cpp/build/bin/libggml-hip.so.0
#6 0x00007ffff64bce97 in ggml_cuda_op_rms_norm_fused(ggml_backend_cuda_context&, ggml_tensor*, ggml_tensor*) ()
from /opt/llama.cpp/build/bin/libggml-hip.so.0
#7 0x00007ffff63efec7 in ggml_cuda_graph_evaluate_and_capture(ggml_backend_cuda_context*, ggml_cgraph*, bool, bool, void const*) ()
from /opt/llama.cpp/build/bin/libggml-hip.so.0
#8 0x00007ffff63ebe01 in ggml_backend_cuda_graph_compute(ggml_backend*, ggml_cgraph*) () from /opt/llama.cpp/build/bin/libggml-hip.so.0
#9 0x00007ffff738e067 in ggml_backend_sched_graph_compute_async () from /opt/llama.cpp/build/bin/libggml-base.so.0
#10 0x00007ffff74b21d1 in llama_context::graph_compute(ggml_cgraph*, bool) () from /opt/llama.cpp/build/bin/libllama.so.0
#11 0x00007ffff74b4724 in llama_context::process_ubatch(llama_ubatch const&, llm_graph_type, llama_memory_context_i*, ggml_status&) ()
from /opt/llama.cpp/build/bin/libllama.so.0
#12 0x00007ffff74bb0e7 in llama_context::decode(llama_batch const&) () from /opt/llama.cpp/build/bin/libllama.so.0
#13 0x00007ffff74bc75b in llama_decode () from /opt/llama.cpp/build/bin/libllama.so.0
#14 0x000055555577d265 in common_init_from_params(common_params&) ()
#15 0x00005555556bfdd6 in server_context_impl::load_model(common_params const&) ()
#16 0x00005555556132f3 in main ()
(gdb) threat apply all bt
Undefined command: "threat". Try "help".
(gdb) thread apply all bt

Thread 38 (Thread 0x7ffe85ff36c0 (LWP 9103) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 37 (Thread 0x7ffe867f46c0 (LWP 9102) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 36 (Thread 0x7ffe86ff56c0 (LWP 9101) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
--Type for more, q to quit, c to continue without paging--
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 35 (Thread 0x7ffe877f66c0 (LWP 9100) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 34 (Thread 0x7ffe87ff76c0 (LWP 9099) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=cl--Type for more, q to quit, c to continue without paging--
ockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 33 (Thread 0x7ffe887f86c0 (LWP 9098) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 32 (Thread 0x7ffe88ff96c0 (LWP 9097) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, u--Type for more, q to quit, c to continue without paging--
nsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 31 (Thread 0x7ffe897fa6c0 (LWP 9096) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 30 (Thread 0x7ffe89ffb6c0 (LWP 9095) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

--Type for more, q to quit, c to continue without paging--
Thread 29 (Thread 0x7ffe8a7fc6c0 (LWP 9094) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 28 (Thread 0x7ffe8affd6c0 (LWP 9093) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 27 (Thread 0x7ffe8b7fe6c0 (LWP 9092) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, absti--Type for more, q to quit, c to continue without paging--
me=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 26 (Thread 0x7ffe8bfff6c0 (LWP 9091) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 25 (Thread 0x7ffe908646c0 (LWP 9090) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
--Type for more, q to quit, c to continue without paging--
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 24 (Thread 0x7ffe910656c0 (LWP 9089) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 23 (Thread 0x7ffe91ffb6c0 (LWP 9088) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
--Type for more, q to quit, c to continue without paging--
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 22 (Thread 0x7ffe927fc6c0 (LWP 9087) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 21 (Thread 0x7ffe92ffd6c0 (LWP 9086) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 20 (Thread 0x7ffe937fe6c0 (LWP 9085) "llama-server"):
--Type for more, q to quit, c to continue without paging--
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 19 (Thread 0x7ffe93fff6c0 (LWP 9084) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 18 (Thread 0x7ffe98df96c0 (LWP 9083) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
--Type for more, q to quit, c to continue without paging--
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 17 (Thread 0x7ffe995fa6c0 (LWP 9082) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 16 (Thread 0x7ffe99dfb6c0 (LWP 9081) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wai--Type for more, q to quit, c to continue without paging--
t.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 15 (Thread 0x7ffe9a5fc6c0 (LWP 9080) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 14 (Thread 0x7ffe9adfd6c0 (LWP 9079) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
--Type for more, q to quit, c to continue without paging--
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 13 (Thread 0x7ffe9b5fe6c0 (LWP 9078) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 12 (Thread 0x7ffe9bdff6c0 (LWP 9077) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 11 (Thread 0x7fffa09fe6c0 (LWP 9076) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
--Type for more, q to quit, c to continue without paging--
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 10 (Thread 0x7fffa1ffb6c0 (LWP 9075) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 9 (Thread 0x7fffa27fc6c0 (LWP 9074) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@en--Type for more, q to quit, c to continue without paging--
try=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 8 (Thread 0x7fffa2ffd6c0 (LWP 9073) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x7ffe8c000c20, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x7ffe8c000c20, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x7ffe8c000c00, mutex=0x7ffe8c000c30) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555558b2882 in std::thread::_State_impl<std::thread::_Invoker<std::tuple<httplib::ThreadPool::ThreadPool(unsigned long, unsigned long, unsigned long)::{lambda()#1}> > >::_M_run() ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 7 (Thread 0x7fffa37fe6c0 (LWP 9072) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=0, nr=288) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9b6ad in __syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=0, nr=288) at ./nptl/cancellation.c:75
#3 0x00007ffff6f1dfba in accept4 (fd=, addr=..., addr_len=, flags=) at ../sysdeps/unix/sysv/linux/accept4.c:31
#4 0x000055555589b1a1 in httplib::Server::listen_internal() ()
#5 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#6 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#7 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78
--Type for more, q to quit, c to continue without paging--

Thread 6 (Thread 0x7fffa3fff6c0 (LWP 9071) "llama-server"):
#0 __syscall_cancel_arch () at ../sysdeps/unix/sysv/linux/x86_64/syscall_cancel.S:56
#1 0x00007ffff6e9b668 in __internal_syscall_cancel (a1=, a2=, a3=, a4=, a5=a5@entry=0, a6=a6@entry=4294967295, nr=202) at ./nptl/cancellation.c:49
#2 0x00007ffff6e9bc8c in __futex_abstimed_wait_common64 (private=0, futex_word=0x555555b4e5b4 <common_log_main()::log+84>, expected=, op=, abstime=0x0, cancel=true) at ./nptl/futex-internal.c:57
#3 __futex_abstimed_wait_common (futex_word=futex_word@entry=0x555555b4e5b4 <common_log_main()::log+84>, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0, cancel=cancel@entry=true) at ./nptl/futex-internal.c:87
#4 0x00007ffff6e9bceb in __GI___futex_abstimed_wait_cancelable64 (futex_word=futex_word@entry=0x555555b4e5b4 <common_log_main()::log+84>, expected=, clockid=clockid@entry=0, abstime=abstime@entry=0x0, private=private@entry=0) at ./nptl/futex-internal.c:139
#5 0x00007ffff6e9e158 in __pthread_cond_wait_common (cond=0x555555b4e590 <common_log_main()::log+48>, mutex=0x555555b4e560 <common_log_main()::log>, clockid=0, abstime=0x0) at ./nptl/pthread_cond_wait.c:426
#6 ___pthread_cond_wait (cond=0x555555b4e590 <common_log_main()::log+48>, mutex=0x555555b4e560 <common_log_main()::log>) at ./nptl/pthread_cond_wait.c:458
#7 0x00005555557ae7ab in common_log::resume()::{lambda()#1}::operator()() const ()
#8 0x00007ffff70e1224 in ?? () from /lib/x86_64-linux-gnu/libstdc++.so.6
#9 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#10 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 3 (Thread 0x7fffa11ff6c0 (LWP 9068) "llama-server"):
#0 __GI___ioctl (fd=3, request=3222817548) at ../sysdeps/unix/sysv/linux/ioctl.c:36
#1 0x00007fffb5f3f7e7 in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#2 0x00007fffb5f357e4 in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#3 0x00007fffb5eb35e5 in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#4 0x00007fffb5e8b928 in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#5 0x00007fffb5eedcf1 in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#6 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#7 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 2 (Thread 0x7fffa9bff6c0 (LWP 9067) "llama-server"):
#0 __GI___ioctl (fd=3, request=3222817548) at ../sysdeps/unix/sysv/linux/ioctl.c:36
#1 0x00007fffb5f3f7e7 in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#2 0x00007fffb5f357e4 in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#3 0x00007fffb5e8b53a in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#4 0x00007fffb5eedcf1 in ?? () from /opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
#5 0x00007ffff6e9eb7b in start_thread (arg=) at ./nptl/pthread_create.c:448
#6 0x00007ffff6f1c7f8 in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78

Thread 1 (Thread 0x7ffff6b01340 (LWP 9064) #0 0x00007fffee469f4b in ?? () from /opt/rocm-7.2 inferior info
(gdb) inf
inferior info
(gdb) info
address exceptions all-registers extensions args files auto-load float auxv frame bookmarks frame-filter breakpoints functions checkpoints guile classes handle common inferiors connections line copying locals dcache macro display macros (gdb) info sharedlibrary
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb)
From To Syms Read 0x00007ffff7fc8000 0x00007ffff7fef3d1 Yes 0x00007ffff7eb5fc0 0x00007ffff7f5b7a6 Yes () 0x00007ffff7903f80 0x00007ffff7c7879e Yes () 0x00007ffff7738770 0x00007ffff77e2efe Yes () 0x00007ffff7469a00 0x00007ffff7658945 Yes () 0x00007ffff7e864c0 0x00007ffff7e8adc4 Yes () 0x00007ffff7370df0 0x00007ffff73dc218 Yes () 0x00007ffff70a1240 0x00007ffff71c5a5e Yes () 0x00007ffff727d2c0 0x00007ffff72f9dc6 Yes 0x00007ffff7e5a3c0 0x00007ffff7e7cf95 Yes () 0x00007ffff6e34400 0x00007ffff6f9687d Yes 0x00007ffff770b340 0x00007ffff771e113 Yes () 0x00007ffff6d47700 0x00007ffff6df50aa Yes () 0x00007ffff6ba28b0 0x00007ffff6c9a81f Yes () 0x00007ffff6359460 0x00007ffff6ab8db0 Yes () 0x00007ffff6b46780 0x00007ffff6b7eb7e Yes () 0x00007ffff2f61820 0x00007ffff2fc9910 Yes () 0x00007ffff1e1d420 0x00007ffff2d4bb10 Yes () 0x00007fffee225640 0x00007fffee63f2d9 Yes () 0x00007fffed3fe7b0 0x00007fffee1b9414 Yes () 0x00007fffb6584cb0 0x00007fffb69b8450 Yes () 0x00007ffff2eb04d0 0x00007ffff2eff2ca Yes () 0x00007fffb5e1ead0 0x00007fffb5f4ba90 Yes () 0x00007ffff7e4c080 0x00007ffff7e4c39c Yes () 0x00007fffb4658010 0x00007fffb5c195e4 Yes () 0x00007ffff6b20400 0x00007ffff6b33129 Yes () 0x00007ffff6b08be0 0x00007ffff6b15528 Yes () 0x00007ffff7e3da20 0x00007ffff7e443c9 Yes () 0x00007ffff76fd640 0x00007ffff7702da5 Yes () 0x00007fffaa514e80 0x00007fffb0c35d32 Yes () (): Shared library is missing debugging (gdb) quit
A debugging session is active. "llama-server"):
.0/lib/libamdhip64.so.7
to load title" data-id="1619648284" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/1" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/1/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/1">#1 0x00007fffee46a7d5 in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
to load title" data-id="1619708291" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/2" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/2/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/2">#2 0x00007fffee4ba782 in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
without paging--
to load title" data-id="1619722397" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/3" data-hovercard-type="pull_request" data-hovercard-url="/ggml-org/llama.cpp/pull/3/hovercard" href="https://github.com/ggml-org/llama.cpp/pull/3">#3 0x00007fffee46b31e in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
to load title" data-id="1619723006" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/4" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/4/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/4">#4 0x00007fffee487206 in ?? () from /opt/rocm-7.2.0/lib/libamdhip64.so.7
to load title" data-id="1619829255" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/5" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/5/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/5">#5 0x00007ffff64be344 in rms_norm_mul_f32_cuda(float const*, float const*, float const*, float*, int, int, int, int, long, long, long, long, long, long, unsigned int, unsigned int, unsigned int, unsigned int, long, long, long, unsigned int, unsigned int, unsigned int, unsigned int, float, ihipStream_t*) () from /opt/llama.cpp/build/bin/libggml-hip.so.0
to load title" data-id="1619876927" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/6" data-hovercard-type="pull_request" data-hovercard-url="/ggml-org/llama.cpp/pull/6/hovercard" href="https://github.com/ggml-org/llama.cpp/pull/6">#6 0x00007ffff64bce97 in ggml_cuda_op_rms_norm_fused(ggml_backend_cuda_context&, ggml_tensor*, ggml_tensor*) () from /opt/llama.cpp/build/bin/libggml-hip.so.0
to load title" data-id="1619885485" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/7" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/7/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/7">#7 0x00007ffff63efec7 in ggml_cuda_graph_evaluate_and_capture(ggml_backend_cuda_context*, ggml_cgraph*, bool, bool, void const*) () from /opt/llama.cpp/build/bin/libggml-hip.so.0
to load title" data-id="1619895928" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/8" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/8/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/8">#8 0x00007ffff63ebe01 in ggml_backend_cuda_graph_compute(ggml_backend*, ggml_cgraph*) () from /opt/llama.cpp/build/bin/libggml-hip.so.0
to load title" data-id="1619953138" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/9" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/9/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/9">#9 0x00007ffff738e067 in ggml_backend_sched_graph_compute_async () from /opt/llama.cpp/build/bin/libggml-base.so.0
to load title" data-id="1619969062" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/10" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/10/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/10">#10 0x00007ffff74b21d1 in llama_context::graph_compute(ggml_cgraph*, bool) () from /opt/llama.cpp/build/bin/libllama.so.0
to load title" data-id="1619969641" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/11" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/11/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/11">#11 0x00007ffff74b4724 in llama_context::process_ubatch(llama_ubatch const&, llm_graph_type, llama_memory_context_i*, ggml_status&) () from /opt/llama.cpp/build/bin/libllama.so.0
to load title" data-id="1619971325" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/12" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/12/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/12">#12 0x00007ffff74bb0e7 in llama_context::decode(llama_batch const&) () from /opt/llama.cpp/build/bin/libllama.so.0
to load title" data-id="1619987096" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/13" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/13/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/13">#13 0x00007ffff74bc75b in llama_decode () from /opt/llama.cpp/build/bin/libllama.so.0
to load title" data-id="1620015675" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/14" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/14/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/14">#14 0x000055555577d265 in common_init_from_params(common_params&) ()
to load title" data-id="1620039073" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/15" data-hovercard-type="issue" data-hovercard-url="/ggml-org/llama.cpp/issues/15/hovercard" href="https://github.com/ggml-org/llama.cpp/issues/15">#15 0x00005555556bfdd6 in server_context_impl::load_model(common_params const&) ()
to load title" data-id="1620084611" data-permission-text="Title is private" data-url="https://github.com/ggml-org/llama.cpp/issues/16" data-hovercard-type="pull_request" data-hovercard-url="/ggml-org/llama.cpp/pull/16/hovercard" href="https://github.com/ggml-org/llama.cpp/pull/16">#16 0x00005555556132f3 in main ()
Try "help info".
main selectors tracepoints
mem set tvariables
missing-debug-handlers sharedlibrary type-printers
missing-objfile-handlers signals types
module skip unwinder
modules source variables
os sources vector
pretty-printer stack vtbl
probes static-tracepoint-markers warranty
proc symbol watchpoints
program target win
record tasks xmethod
registers terminal
scope threads
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.
Shared Object Library
/lib64/ld-linux-x86-64.so.2
/lib/x86_64-linux-gnu/libssl.so.3
/lib/x86_64-linux-gnu/libcrypto.so.3
/opt/llama.cpp/build/bin/libmtmd.so.0
/opt/llama.cpp/build/bin/libllama.so.0
/opt/llama.cpp/build/bin/libggml.so.0
/opt/llama.cpp/build/bin/libggml-base.so.0
/lib/x86_64-linux-gnu/libstdc++.so.6
/lib/x86_64-linux-gnu/libm.so.6
/lib/x86_64-linux-gnu/libgcc_s.so.1
/lib/x86_64-linux-gnu/libc.so.6
/lib/x86_64-linux-gnu/libz.so.1
/lib/x86_64-linux-gnu/libzstd.so.1
/opt/llama.cpp/build/bin/libggml-cpu.so.0
/opt/llama.cpp/build/bin/libggml-hip.so.0
/lib/x86_64-linux-gnu/libgomp.so.1
/opt/rocm-7.2.0/lib/libhipblas.so.3
/opt/rocm-7.2.0/lib/librocblas.so.5
/opt/rocm-7.2.0/lib/libamdhip64.so.7
/opt/rocm-7.2.0/lib/librocsolver.so.0
/opt/rocm-7.2.0/lib/libhipblaslt.so.1
/opt/rocm-7.2.0/lib/librocprofiler-register.so.0
/opt/rocm-7.2.0/lib/libhsa-runtime64.so.1
/opt/rocm-7.2.0/lib/libroctx64.so.4
/opt/rocm-7.2.0/lib/librocroller.so.1
/lib/x86_64-linux-gnu/libelf.so.1
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm.so.2
/opt/amdgpu/lib/x86_64-linux-gnu/libdrm_amdgpu.so.1
/lib/x86_64-linux-gnu/libnuma.so.1
/opt/rocm-7.2.0/lib/../lib/libamd_comgr.so.3
information.

Inferior 1 [process 9064] will be killed.

Quit anyway? (y or n) y

0 replies

aliaj00 · 2026-03-23T16:05:33Z

aliaj00
Mar 23, 2026
Author

from chatgpt:

What changed

With the cleaned loader stack:

the /opt/amdgpu userspace mix is gone
the crash still happens
with --no-warmup --fit off -b 256, it gets past warmup and now dies later at:
srv load_model: initializing slots, n_slots = 1

So the failure is not just warmup, not just --fit, and not just the /opt/amdgpu library contamination.

Your earlier gdb backtrace already showed the fault chain:

libamdhip64.so.7
called from rms_norm_mul_f32_cuda(...)
via ggml_cuda_op_rms_norm_fused(...)
inside ggml_cuda_graph_evaluate_and_capture(...)

That means the crash is still in the HIP runtime / ggml HIP execution path, specifically around fused RMS norm / graph capture / first real graph execution, not in generic model loading.

0 replies

shilga · 2026-03-24T08:46:25Z

shilga
Mar 24, 2026

I'm running llama.cpp without issues with ROCm on my R9700. But I suggest to choose a different distro, Debian support for the R9700 is probably really bad. My setup:

Fedora 43 (Linux 6.19.8-200.fc43)
official llama.cpp ROCm container run with Podman

It just worked out of the box.

0 replies

exploding-rn · 2026-03-24T17:42:53Z

exploding-rn
Mar 24, 2026

The Core Problem

Your GPU is gfx1201 (RDNA 4) — this is extremely new silicon and ROCm/HIP kernel support is still catching up. The segfault happens right here:

1
2
srv load_model: initializing slots, n_slots = 1
Segmentation fault

That's after all tensors are loaded to VRAM successfully, at the point where it tries to execute the first GPU compute operation. This strongly suggests the HIP kernels aren't working correctly for gfx1201.

On top of that, you're running Qwen3.5-35B-A3B which is a MoE + Mamba/SSM hybrid — one of the most complex model architectures. The SSM (State Space Model) kernels are relatively new in llama.cpp and may not have gfx1201 codepaths at all yet.

Also worth noting: your GPU reports VMM: no which can cause issues with large allocations.
Things I'd Try (in order)

Reduce context size first

You're defaulting to 262K context which allocates a 5.1 GB KV cache. Try:
bash

1
llama-server ... -c 4096

This alone might get you past the segfault.
2. Try a simpler model to isolate the issue

Before debugging further, test with something basic like a Llama-3-8B-Q4_K_M or Qwen2.5-7B (non-MoE, non-Mamba). This tells you whether:

gfx1201 is broken in general → segfaults on everything
OR just the Mamba/SSM kernels aren't supported yet → simple models work fine

Try the GFX version override hack
bash

1
HSA_OVERRIDE_GFX_VERSION=12.0.0 llama-server ...

or if that doesn't work:
bash

1
HSA_OVERRIDE_GFX_VERSION=11.0.0 llama-server ...

This forces ROCm to treat your GPU as a different (more supported) architecture. It's a common workaround for new AMD GPUs. Results vary — it might work, might crash differently, but it's worth testing.
4. Make sure you compiled with the right target
bash

1
2
cmake -DGGML_HIP=ON -DAMDGPU_TARGETS="gfx1201" -B build
cmake --build build --config Release

If you installed a pre-built binary, it might not include gfx1201 kernels.
5. Try latest master branch

You're on build 8368 — gfx1201 support patches may have landed after that. Pull latest and rebuild.
6. Consider trying ROCm 6.3.x instead of 7.2

ROCm 7.2 is very new itself. Sometimes pairing bleeding edge GPU + bleeding edge driver = double the bugs. ROCm 6.3 might actually have better stability for your use case.
My Best Guess

Your hardware is fine — this is a software support issue. gfx1201 is so new that the ROCm runtime and/or llama.cpp's HIP kernels don't fully support it yet. The Mamba/SSM compute kernels are the most likely failure point since they're the newest code in llama.cpp.

If a simple model (non-MoE, non-SSM) works with -c 4096, then you've confirmed it's specifically the Qwen3.5 Mamba architecture that's broken on your GPU, and it's worth opening a GitHub issue with that finding.

4 replies

shilga Mar 24, 2026

gfx1201 is so new that the ROCm runtime and/or llama.cpp's HIP kernels don't fully support it yet. The Mamba/SSM compute kernels are the most likely failure point since they're the newest code in llama.cpp.

If a simple model (non-MoE, non-SSM) works with -c 4096, then you've confirmed it's specifically the Qwen3.5 Mamba architecture that's broken on your GPU, and it's worth opening a GitHub issue with that finding.

Just wanted to mention that I run the new Qwen modes just fine on the R9700. 35B and 120B. Currently on 8400 build with the official llama.cpp rocm docker container. Just a newer kernel on Fedora 43. I'm running 64k context size currently.

exploding-rn Mar 25, 2026

Hmm interesting it could be OS... as you mentioned that you run a newer fedora kernel @aliaj00 might not have it so... that could be the issue that a newer fedora kernel is compatible but not a older one. or ya'know Windows being Windows.

I wish I could replicate this and help solve this but... I don't have a R9700 or a beefy enough GPU to run anything larger than 30B... Sadly..
if only i had 80,000$ to spare on 10 Mac studios for a AI cluster with kubernetes

so i belive the issue is kernel if @aliaj00 updates kernel to Linux 6.19.8-200.fc43 it might work

shilga Mar 25, 2026

It might just be the context length as you described.

Running MOE models with CPU offload really works quite good. Qwen3-coder-next gives me 30+ t/s with just one GPU. That is really usable in real life coding agent scenarios. No real cluster needed. Although I concur it would be nice to have one.

exploding-rn Mar 25, 2026

It might just be the context length as you described.

Running MOE models with CPU offload really works quite good. Qwen3-coder-next gives me 30+ t/s with just one GPU. That is really usable in real life coding agent scenarios. No real cluster needed. Although I concur it would be nice to have one.

I agree 10 Mac clusters running 3 1T models locally

need help AMD Radeon™ AI PRO R9700 ROCm debian 12 #20881

Uh oh!

/opt/rocm/bin/rocminfo ROCk module version 6.16.13 is loaded

HSA System Attributes

========== HSA Agents

=================================== Power Consumption ==================================== GPU[0] : Average Graphics Package Power (W): 15.0 GPU[1] : Current Socket Graphics Package Power (W): 0.012

=================================== % time GPU is busy =================================== GPU[0] : GPU use (%): 2 GPU[1] : GPU use (%): 0

0 1 0x7551, 32448 26.0°C 15.0W N/A, N/A, 0 1Mhz 96Mhz 29.8% auto 300.0W 0% 0% 1 2 0x13c0, 2327 38.0°C 0.012W N/A, N/A, 0 N/A 1800Mhz 0% auto N/A 63% 0%

======================================= Unique ID ======================================== GPU[0] : Unique ID: 0xb44207ff2cd402f4 GPU[1] : Unique ID: 0x0

======================================= PCI Bus ID ======================================= GPU[0] : PCI Bus: 0000:03:00.0 GPU[1] : PCI Bus: 0000:7A:00.0

Replies: 8 comments · 20 replies

Uh oh!

Uh oh!

aliaj00 Mar 23, 2026 Author

Uh oh!

Uh oh!

Lock mclk to highest level (~1258 MHz):

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

aliaj00 Mar 23, 2026 Author

apt install -t trixie-backports linux-image-amd64 linux-headers-amd64

Uh oh!

aliaj00 Mar 23, 2026 Author

Uh oh!

aliaj00 Mar 23, 2026 Author

Uh oh!

aliaj00 Mar 23, 2026 Author

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

/opt/rocm/bin/rocminfo
ROCk module version 6.16.13 is loaded

==========
HSA Agents

=================================== Power Consumption ====================================
GPU[0] : Average Graphics Package Power (W): 15.0
GPU[1] : Current Socket Graphics Package Power (W): 0.012

=================================== % time GPU is busy ===================================
GPU[0] : GPU use (%): 2
GPU[1] : GPU use (%): 0

0 1 0x7551, 32448 26.0°C 15.0W N/A, N/A, 0 1Mhz 96Mhz 29.8% auto 300.0W 0% 0%
1 2 0x13c0, 2327 38.0°C 0.012W N/A, N/A, 0 N/A 1800Mhz 0% auto N/A 63% 0%

======================================= Unique ID ========================================
GPU[0] : Unique ID: 0xb44207ff2cd402f4
GPU[1] : Unique ID: 0x0

======================================= PCI Bus ID =======================================
GPU[0] : PCI Bus: 0000:03:00.0
GPU[1] : PCI Bus: 0000:7A:00.0

Replies: 8 comments 20 replies

aliaj00 Mar 23, 2026
Author

aliaj00
Mar 23, 2026
Author

aliaj00
Mar 23, 2026
Author

aliaj00
Mar 23, 2026
Author

aliaj00
Mar 23, 2026
Author