Add constant_cycles_kernel benchmark + header split and Makefile

Benchmark supports:
=> Looping on the GPU for a specified number of iterations
=> Looping infinitely on the GPU
=> Logging the sum time (ns and ms) for all iterations to complete

Also splits `testbench.h` out of `copy_testbench.h` for use in other
benchmarks.

1 files changed, 14 insertions, 0 deletions

diff --git a/testbench.h b/testbench.h
new file mode 100644
index 0000000..28d7f77
--- /dev/null
+++ b/testbench.h
@@ -0,0 +1,14 @@
+/* Copyright 2021-2023 Joshua Bakita
+ * Header for miscellaneous experimental helper functions.
+ */
+
+#define SAFE(x) \
+        if ((err = (cudaError_t)(x)) != 0) { \
+                printf("CUDA error %d! %s\n", err, cudaGetErrorString(err)); \
+                printf("Suspect line: %s\n", #x); \
+                exit(1); \
+        }
+
+#define s2ns(s) ((s)*1000l*1000l*1000l)
+#define ns2ms(s) ((s)/(1000l*1000l))
+#define timediff(start, end) ((s2ns((end).tv_sec) + (end).tv_nsec) - (s2ns((start).tv_sec) + (start).tv_nsec))