add benchmark

.gitignore

@@ -8,3 +8,5 @@ test-kseq
 *.so
 /test-seqio-*
 test-cigar
+/seqkit-ben*
+/benchmark-*

Makefile

@@ -8,19 +8,22 @@ export seqioObj := $(seqioSource:.c=.o)
 export cigarSource := $(ROOT_DIR)/cigar.c
 export cigarObj := $(cigarSource:.c=.o)
 
-all: build-test libseqio.so
+all: build-test libseqio.so build-benchmark
 
 ^.o:^.c
 	$(cc) $(CFLAGS) $(LIBS) -c $<
 
 build-test:
 	$(MAKE) -C ./test
 
+build-benchmark:
+	$(MAKE) -C ./benchmark
+
 libseqio.so:
 	$(cc) $(CFLAGS) -fPIC -c -o seqio.o seqio.c
 	$(cc) -shared -fPIC -o libseqio.so seqio.o
 
 clean:
-	rm -f *.o test-seqio test-kseq libseqio.so test-seqio-* test-cigar
+	rm -f *.o test-seqio test-kseq libseqio.so test-seqio-* test-cigar benchmark-*
 
 .PHONY:clean

benchmark/Makefile

@@ -0,0 +1,9 @@
+all: $(ROOT_DIR)/benchmark-seqio $(ROOT_DIR)/benchmark-kseq
+
+$(ROOT_DIR)/benchmark-seqio: seqio.c $(seqioObj)
+	@$(cc) $(CFLAGS) -DREAD -o $(ROOT_DIR)/benchmark-seqio-read $^ $(LIBS)
+	@$(cc) $(CFLAGS) -DWRITE -o $(ROOT_DIR)/benchmark-seqio-write $^ $(LIBS)
+	@$(cc) $(CFLAGS) -DWRITE_GZ -o $(ROOT_DIR)/benchmark-seqio-write-gz $^ $(LIBS)
+
+$(ROOT_DIR)/benchmark-kseq: kseq.c kseq.h $(seqioObj)
+	@$(cc) $(CFLAGS) -o $@ $^ $(LIBS)

benchmark/kseq.c

@@ -0,0 +1,53 @@
+#include "kseq.h"
+#include <stdio.h>
+#include <time.h>
+#include <zlib.h>
+
+KSEQ_INIT(gzFile, gzread)
+
+inline static float
+timer(void)
+{
+  struct timespec ts;
+  if (clock_gettime(CLOCK_MONOTONIC, &ts) == -1) {
+    return -1;
+  }
+  return (float)ts.tv_sec + (float)ts.tv_nsec / 1000000000.0;
+}
+
+static inline void
+benchmark_kseq_read(char* filename)
+{
+  gzFile fp;
+  kseq_t* seq;
+  int l;
+  int nseq = 0;
+  size_t sum_length = 0;
+  if (filename == NULL) {
+    return;
+  }
+  fp = gzopen(filename, "r");
+  seq = kseq_init(fp);
+  while ((l = kseq_read(seq)) >= 0) {
+    sum_length += seq->seq.l;
+    nseq++;
+  }
+  kseq_destroy(seq);
+  gzclose(fp);
+  printf("Sum of sequence length: %lu\n", sum_length);
+  printf("Number of sequences: %d\n", nseq);
+}
+
+int
+main(int argc, char* argv[])
+{
+  if (argc < 2) {
+    fprintf(stderr, "Usage: %s <in.seq>\n", argv[0]);
+    return 1;
+  }
+  float start = timer();
+  benchmark_kseq_read(argv[1]);
+  float end = timer();
+  printf("Time: %f\n", end - start);
+  return 0;
+}