-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy path.gitignore
More file actions
182 lines (175 loc) · 6.79 KB
/
Copy path.gitignore
File metadata and controls
182 lines (175 loc) · 6.79 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
# Build artifacts
build/
build-*/
*.o
*.obj
CMakeCache.txt
CMakeFiles/
cmake_install.cmake
compile_commands.json
# Python venv and bytecode
.venv/
venv/
__pycache__/
*.pyc
# Raw per-iteration benchmark CSVs (1-40 MB each; regenerate with
# scripts/run_*.sh). The aggregated summary.csv files are tracked
# because the README headline numbers cite them. Per-sweep raw CSVs
# follow per-axis naming conventions (W<N>.csv for world sweep,
# P<N>.csv for payload sweep, dev.csv for the per-build benchmark
# dump). machine.json fingerprints are also per-run and not portable.
results/dev*.csv
results/**/dev.csv
results/**/dev_*.csv
results/**/all.csv
results/**/all_after.csv
results/**/W[0-9]*.csv
results/**/P[0-9]*.csv
# Stray run artifacts: stderr dumps, raw combined CSVs, per-grid score
# breakdowns. The aggregated summary.csv / scores.csv stay tracked.
results/**/*.err
results/**/combined_raw.csv
results/**/score_breakdown.csv
# Hot/cold-test raw per-iteration CSVs (40+ MB each). Comparison
# medians are aggregated into a tracked summary file alongside.
results/hotcold_test/*.csv
!results/hotcold_test/summary*.csv
# Topology-smoke-test raw CSVs (2 MB each); summary is tracked.
results/topology_test/*.csv
!results/topology_test/summary*.csv
# Region-A* smoke-test raw CSVs; summary tracked.
# HPA tier-2 smoke-test raw CSVs; summary tracked.
results/region_full_test/*.csv
!results/region_full_test/summary*.csv
results/region_test/*.csv
!results/region_test/summary*.csv
# Hex region-A* smoke-test raw CSVs; summary tracked.
results/hex_region_test/*.csv
!results/hex_region_test/summary*.csv
results/hex_region_test/*_machine.json
# chunk_summary prototype raw CSVs; headline + per-W summaries tracked.
results/chunk_summary_test/*.csv
!results/chunk_summary_test/*_summary.csv
!results/chunk_summary_test/summary.csv
results/chunk_summary_test/*_machine.json
# active_list prototype raw CSVs; per-W summaries tracked.
results/active_list_test/*.csv
!results/active_list_test/*_summary.csv
results/active_list_test/*_machine.json
# predicate_scan prototype raw CSVs; per-W summaries tracked.
results/predicate_scan_test/*.csv
!results/predicate_scan_test/*_summary.csv
results/predicate_scan_test/*_machine.json
# distance_field prototype raw CSVs; per-W summaries tracked.
results/distance_field_test/*.csv
!results/distance_field_test/*_summary.csv
results/distance_field_test/*_machine.json
# hierarchical_radius prototype raw CSVs; summaries tracked.
results/hierarchical_radius_test/*.csv
!results/hierarchical_radius_test/*_summary.csv
results/hierarchical_radius_test/*_machine.json
# Genre re-scoring v2 raw + intermediate CSVs. The aggregated
# full_summary.csv + the scores_v2*/scores+breakdown CSVs are tracked.
results/genre_score_v2/*.csv
results/genre_score_v2/*_machine.json
!results/genre_score_v2/scores_*/scores.csv
!results/genre_score_v2/scores_*/score_breakdown.csv
!results/genre_score_v2/full_summary.csv
# Genre re-scoring v3 — same shape as v2, fresh bench, adds
# stencil_strategy knob.
results/genre_score_v3/*.csv
results/genre_score_v3/*_machine.json
!results/genre_score_v3/scores_*/scores.csv
!results/genre_score_v3/scores_*/score_breakdown.csv
!results/genre_score_v3/full_summary.csv
# W=8192 "truly enormous" probe; summary tracked.
results/w8192_test/*.csv
!results/w8192_test/*_summary.csv
results/w8192_test/*_machine.json
# parallel_chunk_update probe; summary tracked.
results/parallel_test/*.csv
!results/parallel_test/*_summary.csv
results/parallel_test/*_machine.json
# Hex prototype port raw CSVs; summary tracked.
results/hex_prototypes_test/*.csv
!results/hex_prototypes_test/*_summary.csv
results/hex_prototypes_test/*_machine.json
# chunk_query prototype raw CSVs; summary tracked.
results/chunk_query_test/*.csv
!results/chunk_query_test/*_summary.csv
results/chunk_query_test/*_machine.json
# Hex-native distance_field / hierarchical_radius raw CSVs; summary tracked.
results/hex_native_test/*.csv
!results/hex_native_test/*_summary.csv
results/hex_native_test/*_machine.json
# Sierpiński curve probe raw CSVs; summary tracked.
results/sierpinski_test/*.csv
!results/sierpinski_test/*_summary.csv
results/sierpinski_test/*_machine.json
# Full-block curve raw CSVs; summary tracked.
results/full_block_test/*.csv
!results/full_block_test/*_summary.csv
results/full_block_test/*_machine.json
# curve_order_scan raw CSVs; summary tracked.
results/curve_order_scan_test/*.csv
!results/curve_order_scan_test/*_summary.csv
results/curve_order_scan_test/*_machine.json
# neighbor_indices_8 optimization probe raw CSVs; summaries tracked.
results/neighbor8_test/*.csv
!results/neighbor8_test/*_summary.csv
results/neighbor8_test/*_machine.json
# neighbor_indices_4 probe; summaries tracked.
results/neighbor4_test/*.csv
!results/neighbor4_test/*_summary.csv
results/neighbor4_test/*_machine.json
# stencil_ring2_fast comparison; summary tracked.
results/stencil_ring2_fast_test/*.csv
!results/stencil_ring2_fast_test/*_summary.csv
results/stencil_ring2_fast_test/*_machine.json
# Genre-rescoring raw + intermediate CSVs (the 79 MB full_dev.csv +
# the 235 KB full_summary.csv + the small dev.csv). Aggregated
# scores.csv + score_breakdown.csv files in scores_*/ are tracked
# because they're the headline data RESULTS.md cites. machine.json
# fingerprints are per-host and not portable.
results/region_genre/*.csv
results/region_genre/*_machine.json
!results/region_genre/scores_*/scores.csv
!results/region_genre/scores_*/score_breakdown.csv
# Re-include aggregated summary CSVs that share the W<N> prefix
# (world_sweep_full produces W<N>_<grid>_summary.csv +
# W<N>_combined_summary.csv from world_size_sweep.py). The raw
# per-iteration W<N>_<grid>.csv files remain ignored.
!results/**/W[0-9]*_summary.csv
!results/**/W[0-9]*_combined_summary.csv
results/**/summary_before*.csv
results/**/summary_after*.csv
results/**/*_machine.json
results/**/machine.json
# Top-level per-build fingerprints (run_local.sh / run_fastmath_compare.sh
# regenerate these). `**` above only matches nested paths, so name them here.
results/*_machine.json
results/results.csv
# Profiling / sanitizer / pgo run outputs (transient per-host).
results/perf/
results/sanitizer/
results/pgo/
# Per-run GCP artifacts under results/gcp/<run-id>/ are large (raw CSVs
# + build/test logs + source snapshot tarball), per-run, and reproducible
# by re-running scripts/run_gcp_bench.sh. The aggregated headline data
# should be promoted into results/halo_final/ or a labeled subdir if
# kept.
results/gcp/
# Per-workload plot PNGs are reproducible from the tracked summary CSVs
# via analysis/plot.py — don't bloat the repo with them. Heatmap
# overview PNGs (the all-layouts × all-workloads view) are tracked
# because they're the headline visualization referenced in README.
results/**/plots/**/*.png
!results/**/plots/**/heatmap_*.png
# CTest scratch
Testing/
# Editor / OS
.DS_Store
*.swp
*~
.idea/
.vscode/