squall/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
fb335f6a5f
cutlass/test/unit/gemm/device
History
Andrew Kerr fb335f6a5f
CUTLASS 2.0 (#62) 
CUTLASS 2.0

Substantially refactored for

- Better performance, particularly for native Turing Tensor Cores
- Robust and durable templates spanning the design space
- Encapsulated functionality embodying modern C++11 programming techniques
- Optimized containers and data types for efficient, generic, portable device code

Updates to:
- Quick start guide
- Documentation
- Utilities
- CUTLASS Profiler

Native Turing Tensor Cores
- Efficient GEMM kernels targeting Turing Tensor Cores
- Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands

Coverage of existing CUTLASS functionality:
- GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs
- Volta Tensor Cores through native mma.sync and through WMMA API
- Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions
- Batched GEMM operations
- Complex-valued GEMMs

Note: this commit and all that follow require a host compiler supporting C++11 or greater.
2019-11-19 16:55:34 -08:00
..
CMakeLists.txt CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_b1t_b1n_s32n_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_b1t_b1n_s32n_wmma_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_b1t_b1n_s32t_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_b1t_b1n_s32t_wmma_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f16n_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f16n_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f16t_tensor_op_f32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f16t_volta_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f16t_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f16t_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f32n_tensor_op_f32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f32n_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f32t_tensor_op_f32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f32t_volta_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16n_f32t_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f16n_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f16n_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f16t_tensor_op_f16_sm75_slicedk.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f16t_tensor_op_f16_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f16t_volta_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f16t_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f16t_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f32n_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f32t_tensor_op_f32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f32t_volta_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16n_f16t_f32t_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16n_singlestage_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16n_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16n_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16t_singlestage_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16t_tensor_op_f16_sm75_slicedk.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16t_tensor_op_f16_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16t_volta_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16t_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f16t_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f32n_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f32t_singlestage_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f32t_tensor_op_f32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f32t_volta_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16n_f32t_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f16n_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f16n_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f16t_wmma_tensor_op_f16_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f16t_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f32n_tensor_op_f32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f32n_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f32t_tensor_op_f32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f32t_volta_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_f16t_f16t_f32t_wmma_tensor_op_f32_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s4n_s4t_s4n_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s4t_s4n_s32n_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s4t_s4n_s32n_wmma_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s4t_s4n_s32t_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s4t_s4n_s32t_wmma_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8n_s8t_s8n_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8t_s8n_s8n_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8t_s8n_s8n_wmma_tensor_op_s32_sm72.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8t_s8n_s8t_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8t_s8n_s8t_wmma_tensor_op_s32_sm72.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8t_s8n_s32n_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8t_s8n_s32n_wmma_tensor_op_s32_sm72.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8t_s8n_s32t_tensor_op_s32_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_s8t_s8n_s32t_wmma_tensor_op_s32_sm72.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_splitk_simt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_splitk_tensor_op_sm70.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_splitk_tensor_op_sm75.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
gemm_u8t_u8n_s32t_wmma_tensor_op_s32_sm72.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_cgemm_nn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_cgemm_nt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_cgemm_tn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_cgemm_tt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_dgemm_nn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_dgemm_nt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_dgemm_tn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_dgemm_tt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_hgemm_nn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_hgemm_nt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_hgemm_tn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_hgemm_tt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_igemm_nn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_igemm_nt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_igemm_tn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_igemm_tt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_int8_igemm_sm61_perf.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_int8_igemm_sm61_sliced_k.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_int8_igemm_sm61.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_sgemm_nn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_sgemm_nt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_sgemm_tn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_sgemm_tt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_sm50.py CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_zgemm_nn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_zgemm_nt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_zgemm_tn_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
simt_zgemm_tt_sm50.cu CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
testbed_complex.h CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
testbed_interleaved.h CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
testbed_splitk.h CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00
testbed.h CUTLASS 2.0 (#62) 2019-11-19 16:55:34 -08:00