(Slightly) Improve Python and Cython benchmarks

Makefile

@@ -27,7 +27,7 @@ clean:
 		*/gc.bin \
 		*/gc \
 		*/gc.o \
-		cython/gc.c \
+		cython*/gc.c \
 		java/gc.class \
 		nim/nimcache \
 		rust*/target \
@@ -72,6 +72,7 @@ report.md: ada.time ada.version \
 	crystal.001.csp.time crystal.version \
 	crystal.002.peek.time crystal.version \
 	cython.time cython.version \
+	cython.001.fopen.time cython.version \
 	d.time d.version \
 	fpc.time fpc.version \
 	go.time go.version \
@@ -85,7 +86,9 @@ report.md: ada.time ada.version \
 	node.time node.version \
 	perl.time perl.version \
 	pypy.time pypy.version \
+	pypy.001.binary.time pypy.version \
 	python.time python.version \
+	python.001.binary.time python.version \
 	rust.time rust.version \
 	rust.001.time rust.version \
 	rust.002.bitshift.time rust.version \
@@ -186,6 +189,10 @@ cython/gc.bin: cython/gc.pyx
 	cython --embed $< \
 		&& gcc -I/usr/include/python2.7 -O3 -o $@ cython/gc.c -lpython2.7
 
+cython%/gc.bin: cython%/gc.pyx
+	cython --embed $< \
+		&& gcc -I/usr/include/python2.7 -O3 -o $@ $(patsubst %.pyx,%.c,$<) -lpython2.7
+
 # D
 %.bin: %.d
 	ldc2 -O5 -boundscheck=off -release -of=$@ $<
@@ -240,12 +247,18 @@ perl/gc.bin: perl/gc.pl
 python/gc.bin: python/gc.py
 	cp $< $@
 
+python%/gc.bin: python%/gc.py
+	cp $< $@
+
 # Pypy
 # We need to copy the pypy script to the canonical path to simplify the e.g.
 # the cleaning rule
 pypy/gc.bin: pypy/gc.py
 	cp $< $@
 
+pypy%/gc.bin: pypy%/gc.py
+	cp $< $@
+
 # Rust
 rust/gc.bin: rust/src/main.rs rust/Cargo.toml
 	RUSTFLAGS="-C target-cpu=native" cargo +nightly build --release --manifest-path $(word 2,$^) -Z unstable-options --out-dir $(shell dirname $@) \

cython.001.fopen/gc.pyx

@@ -0,0 +1,45 @@
+# cython: boundscheck=False, wraparound=False, initializedcheck=False, cdivision=True
+
+from libc.stdio cimport fopen, fgets, fclose, FILE
+from libc.stdint cimport uint64_t
+
+
+cdef uint64_t value[256];
+
+for i in range(256):
+    value[i] = 0;
+value[ord('A')] = value[ord('T')] = 1;
+value[ord('G')] = value[ord('C')] = 1ull << 32;
+
+
+cpdef void main():
+    cdef FILE*    f
+    cdef double   at
+    cdef double   gc
+    cdef char     line[4096]
+    cdef uint64_t totals     = 0
+    cdef size_t   i          = 0
+
+    f = fopen("chry_multiplied.fa","r")
+    if f == NULL:
+        raise OSError("Can't open input file")
+
+    with nogil:
+
+        while fgets(line, sizeof(line), f):
+            if line[0] == '>':
+                continue
+            while line[i] != 0:
+                totals += value[line[i]]
+                i += 1
+            i = 0
+
+        fclose(f)
+        at = totals & 0xFFFFFFFFull;
+        gc = totals >> 32;
+
+    print( gc / (at + gc) )
+
+
+if __name__ == '__main__':
+    main()

pypy.001.binary/gc.py

@@ -0,0 +1,22 @@
+#!/usr/bin/env pypy
+import string
+
+def main():
+    file = open("chry_multiplied.fa","rb")
+    a = 0
+    t = 0
+    g = 0
+    c = 0
+    for line in file:
+        if line[0] != b">" and len(line) != line.count(b"N") + 1:
+            g += line.count(b"G")
+            c += line.count(b"C")
+            a += line.count(b"A")
+            t += line.count(b"T")
+    totalBaseCount = a + t + c + g
+    gcCount = g + c
+    gcFraction = float(gcCount) / totalBaseCount
+    print( gcFraction * 100 )
+
+if __name__ == '__main__':
+    main()

python.001.binary/gc.py

@@ -0,0 +1,25 @@
+#!/usr/bin/env python
+import string
+
+A,T,G,C,N = map(ord, "ATGCN")
+
+def main():
+    file = open("chry_multiplied.fa","rb")
+    a = 0
+    t = 0
+    g = 0
+    c = 0
+    for line in file:
+        if line[0] != 62 and len(line) != line.count(N) + 1:
+            g += line.count(G)
+            c += line.count(C)
+            a += line.count(A)
+            t += line.count(T)
+    totalBaseCount = a + t + c + g
+    gcCount = g + c
+
+    gcFraction = float(gcCount) / totalBaseCount
+    print( gcFraction * 100 )
+
+if __name__ == '__main__':
+    main()