The Museum

Raw Benchmark Data and Glossary

Dataset Browser

Language Filter:

N	Language	Kernel	Avg Time (s)	GFLOPS
10	c	naive	0.000001	N/A
10	c	tiled	0.000001	N/A
10	c	simd	0.000001	N/A
20	c	naive	0.000006	N/A
20	c	tiled	0.000005	N/A
20	c	simd	0.000001	N/A
30	c	naive	0.000017	N/A
30	c	tiled	0.000016	N/A
30	c	simd	0.000004	N/A
40	c	naive	0.000039	N/A
40	c	tiled	0.000037	N/A
40	c	simd	0.000006	N/A
50	c	naive	0.000076	N/A
50	c	tiled	0.000068	N/A
50	c	simd	0.000012	N/A
60	c	naive	0.000134	N/A
60	c	tiled	0.000119	N/A
60	c	simd	0.000022	N/A
70	c	naive	0.000234	N/A
70	c	tiled	0.000196	N/A
70	c	simd	0.000040	N/A
80	c	naive	0.000361	N/A
80	c	tiled	0.000292	N/A
80	c	simd	0.000045	N/A
90	c	naive	0.000529	N/A
90	c	tiled	0.000406	N/A
90	c	simd	0.000063	N/A
100	c	naive	0.000735	N/A
100	c	tiled	0.000572	N/A
100	c	simd	0.000086	N/A
200	c	naive	0.006440	N/A
200	c	tiled	0.004566	N/A
200	c	simd	0.000732	N/A
300	c	naive	0.022219	N/A
300	c	tiled	0.015264	N/A
300	c	simd	0.002410	N/A
400	c	naive	0.053575	N/A
400	c	tiled	0.036242	N/A
400	c	simd	0.006210	N/A
500	c	naive	0.104916	N/A
500	c	tiled	0.070800	N/A
500	c	simd	0.011402	N/A
600	c	naive	0.183185	N/A
600	c	tiled	0.122488	N/A
600	c	simd	0.021581	N/A
700	c	naive	0.291707	N/A
700	c	tiled	0.194633	N/A
700	c	simd	0.032773	N/A
800	c	naive	0.496373	N/A
800	c	tiled	0.290745	N/A
800	c	simd	0.050909	N/A
900	c	naive	0.676366	N/A
900	c	tiled	0.415413	N/A
900	c	simd	0.068217	N/A
1000	c	naive	0.934128	N/A
1000	c	tiled	0.569557	N/A
1000	c	simd	0.097907	N/A
2000	c	naive	9.735280	N/A
2000	c	tiled	4.590740	N/A
2000	c	simd	0.874592	N/A
10	cpp	naive	0.000001	N/A
10	cpp	tiled	0.000000	N/A
10	cpp	simd	0.000001	N/A
20	cpp	naive	0.000003	N/A
20	cpp	tiled	0.000003	N/A
20	cpp	simd	0.000003	N/A
30	cpp	naive	0.000010	N/A
30	cpp	tiled	0.000009	N/A
30	cpp	simd	0.000008	N/A
40	cpp	naive	0.000022	N/A
40	cpp	tiled	0.000020	N/A
40	cpp	simd	0.000010	N/A
50	cpp	naive	0.000046	N/A
50	cpp	tiled	0.000037	N/A
50	cpp	simd	0.000019	N/A
60	cpp	naive	0.000088	N/A
60	cpp	tiled	0.000064	N/A
60	cpp	simd	0.000032	N/A
70	cpp	naive	0.000147	N/A
70	cpp	tiled	0.000108	N/A
70	cpp	simd	0.000055	N/A
80	cpp	naive	0.000237	N/A
80	cpp	tiled	0.000161	N/A
80	cpp	simd	0.000062	N/A
90	cpp	naive	0.000367	N/A
90	cpp	tiled	0.000218	N/A
90	cpp	simd	0.000096	N/A
100	cpp	naive	0.000499	N/A
100	cpp	tiled	0.000307	N/A
100	cpp	simd	0.000131	N/A
200	cpp	naive	0.005669	N/A
200	cpp	tiled	0.002490	N/A
200	cpp	simd	0.000967	N/A
300	cpp	naive	0.020533	N/A
300	cpp	tiled	0.008257	N/A
300	cpp	simd	0.003722	N/A
400	cpp	naive	0.050361	N/A
400	cpp	tiled	0.019467	N/A
400	cpp	simd	0.008130	N/A
500	cpp	naive	0.100307	N/A
500	cpp	tiled	0.037864	N/A
500	cpp	simd	0.016425	N/A
600	cpp	naive	0.175949	N/A
600	cpp	tiled	0.065575	N/A
600	cpp	simd	0.030589	N/A
700	cpp	naive	0.284257	N/A
700	cpp	tiled	0.104285	N/A
700	cpp	simd	0.044208	N/A
800	cpp	naive	0.429051	N/A
800	cpp	tiled	0.156286	N/A
800	cpp	simd	0.069136	N/A
900	cpp	naive	0.607257	N/A
900	cpp	tiled	0.222535	N/A
900	cpp	simd	0.092915	N/A
1000	cpp	naive	0.835162	N/A
1000	cpp	tiled	0.305381	N/A
1000	cpp	simd	0.134594	N/A
2000	cpp	naive	8.484150	N/A
2000	cpp	tiled	2.477330	N/A
2000	cpp	simd	1.125940	N/A
10	pytorch	aten	0.000002	N/A
10	numpy	openblas/mkl	0.000001	N/A
20	pytorch	aten	0.000002	N/A
20	numpy	openblas/mkl	0.000001	N/A
30	pytorch	aten	0.000003	N/A
30	numpy	openblas/mkl	0.000002	N/A
40	pytorch	aten	0.000003	N/A
40	numpy	openblas/mkl	0.000003	N/A
50	pytorch	aten	0.000005	N/A
50	numpy	openblas/mkl	0.000004	N/A
60	pytorch	aten	0.000005	N/A
60	numpy	openblas/mkl	0.000006	N/A
70	pytorch	aten	0.000009	N/A
70	numpy	openblas/mkl	0.000009	N/A
80	pytorch	aten	0.000010	N/A
80	numpy	openblas/mkl	0.000011	N/A
90	pytorch	aten	0.000018	N/A
90	numpy	openblas/mkl	0.000015	N/A
100	pytorch	aten	0.000017	N/A
100	numpy	openblas/mkl	0.000020	N/A
200	pytorch	aten	0.000128	N/A
200	numpy	openblas/mkl	0.000132	N/A
300	pytorch	aten	0.000412	N/A
300	numpy	openblas/mkl	0.000418	N/A
400	pytorch	aten	0.000981	N/A
400	numpy	openblas/mkl	0.000961	N/A
500	pytorch	aten	0.001851	N/A
500	numpy	openblas/mkl	0.001839	N/A
600	pytorch	aten	0.003142	N/A
600	numpy	openblas/mkl	0.003102	N/A
700	pytorch	aten	0.004952	N/A
700	numpy	openblas/mkl	0.005011	N/A
800	pytorch	aten	0.007352	N/A
800	numpy	openblas/mkl	0.007334	N/A
900	pytorch	aten	0.010359	N/A
900	numpy	openblas/mkl	0.010355	N/A
1000	pytorch	aten	0.014185	N/A
1000	numpy	openblas/mkl	0.014203	N/A
2000	pytorch	aten	0.115306	N/A
2000	numpy	openblas/mkl	0.112743	N/A

Glossary

AVX2: Advanced Vector Extensions 2. A 256-bit SIMD instruction set native to Intel and AMD processors.
Bump Allocator: A memory allocation strategy where memory is grabbed in one huge chunk, and subsequent allocations just move a pointer forward. No free lists, no OS jitter.
GFLOPS: Giga-Floating Point Operations Per Second. The primary metric for compute throughput.
L1 Cache: The absolute fastest, but smallest (32-64KB), memory bank located directly on the CPU core.
OpenMP: Open Multi-Processing. An API that supports multi-platform shared memory multiprocessing programming in C, C++, and Fortran.