# Dense Model Benchmark:
# Cannot find config for target=llvm, workload=('dense_nopack.x86', ('TENSOR', (1, 768), 'float32'), ('TENSOR', (2, 768), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=llvm, workload=('dense_nopack.x86', ('TENSOR', (1, 768), 'float32'), ('TENSOR', (768, 768), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=llvm, workload=('dense_nopack.x86', ('TENSOR', (128, 3072), 'float32'), ('TENSOR', (768, 3072), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=llvm, workload=('dense_nopack.x86', ('TENSOR', (128, 768), 'float32'), ('TENSOR', (3072, 768), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=llvm, workload=('dense_nopack.x86', ('TENSOR', (128, 768), 'float32'), ('TENSOR', (768, 768), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=llvm, workload=('batch_matmul.x86', ('TENSOR', (12, 128, 128), 'float32'), ('TENSOR', (12, 64, 128), 'float32')). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=llvm, workload=('batch_matmul.x86', ('TENSOR', (12, 128, 64), 'float32'), ('TENSOR', (12, 128, 64), 'float32')). A fallback configuration is used, which may bring great performance regression.
# Runtime: 165.26 ms (12.83 ms)
# Block Sparse Model with 1x1 blocks:
# Runtime: 67.75 ms (8.83 ms)
# Here is the output of this script on a GPU (GTX 1070) with the target "cuda -libs=cublas".
#
# Dense Model Benchmark:
# Cannot find config for target=cuda -keys=cuda,gpu -libs=cublas -max_num_threads=1024 -thread_warp_size=32, workload=('dense_cublas.cuda', ('TENSOR', (1, 768), 'float32'), ('TENSOR', (2, 768), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=cuda -keys=cuda,gpu -libs=cublas -max_num_threads=1024 -thread_warp_size=32, workload=('dense_cublas.cuda', ('TENSOR', (1, 768), 'float32'), ('TENSOR', (768, 768), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=cuda -keys=cuda,gpu -libs=cublas -max_num_threads=1024 -thread_warp_size=32, workload=('dense_cublas.cuda', ('TENSOR', (128, 3072), 'float32'), ('TENSOR', (768, 3072), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=cuda -keys=cuda,gpu -libs=cublas -max_num_threads=1024 -thread_warp_size=32, workload=('dense_cublas.cuda', ('TENSOR', (128, 768), 'float32'), ('TENSOR', (3072, 768), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=cuda -keys=cuda,gpu -libs=cublas -max_num_threads=1024 -thread_warp_size=32, workload=('dense_cublas.cuda', ('TENSOR', (128, 768), 'float32'), ('TENSOR', (768, 768), 'float32'), None, 'float32'). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=cuda -keys=cuda,gpu -libs=cublas -max_num_threads=1024 -thread_warp_size=32, workload=('batch_matmul_cublas.cuda', ('TENSOR', (12, 128, 128), 'float32'), ('TENSOR', (12, 64, 128), 'float32'), (12, 128, 64)). A fallback configuration is used, which may bring great performance regression.
# Cannot find config for target=cuda -keys=cuda,gpu -libs=cublas -max_num_threads=1024 -thread_warp_size=32, workload=('batch_matmul_cublas.cuda', ('TENSOR', (12, 128, 64), 'float32'), ('TENSOR', (12, 128, 64), 'float32'), (12, 128, 128)). A fallback configuration is used, which may bring great performance regression.
# Runtime: 10.64 ms (0.29 ms)
# Block Sparse Model with 1x1 blocks:
# Runtime: 6.46 ms (0.05 ms)
- 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.