CUTLASS 2.2 (#96)

Adds support for NVIDIA Ampere Architecture features. CUDA 11 Toolkit recommended.
This commit is contained in:
Andrew Kerr
2020-06-08 16:17:35 -07:00
committed by GitHub
parent e33d90b361
commit 86931fef85
584 changed files with 51080 additions and 3373 deletions
+2 -1
View File
@@ -216,6 +216,7 @@ participating warps - since each warp now owns a partial sum (since they compute
The following additional resources describe design and implementation details of GEMMs
targeting NVIDIA GPUs.
- [Developing CUDA Kernels to Push Tensor Cores to the Absolute Limit on NVIDIA A100.](https://www.nvidia.com/en-us/gtc) (SR 21745)
- [CUTLASS: Fast Linear Algebra in CUDA C++](https://devblogs.nvidia.com/cutlass-linear-algebra-cuda/)
- [CUTLASS: SOFTWARE PRIMITIVES FOR DENSE LINEAR ALGEBRA AT ALL LEVELS AND SCALES WITHIN CUDA](https://on-demand-gtc.gputechconf.com/gtcnew/sessionview.php?sessionName=s8854-cutlass%3a+software+primitives+for+dense+linear+algebra+at+all+levels+and+scales+within+cuda)
- [Programming Tensor Cores: NATIVE VOLTA TENSOR CORES WITH CUTLASS](https://developer.download.nvidia.com/video/gputechconf/gtc/2019/presentation/s9593-cutensor-high-performance-tensor-operations-in-cuda-v2.pdf)
@@ -224,7 +225,7 @@ targeting NVIDIA GPUs.
# Copyright
Copyright (c) 2017-2019, NVIDIA CORPORATION. All rights reserved.
Copyright (c) 2017-2020, NVIDIA CORPORATION. All rights reserved.
```
Redistribution and use in source and binary forms, with or without modification, are permitted