leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 25 Packages Projects Releases Wiki Activity

releaase 2.11 (#703)

Browse Source
This commit is contained in:
Aditya Atluri
2022-11-19 09:02:15 -05:00
committed by GitHub
parent 3c90f6aea6
commit c975e2ccbb
329 changed files with 47332 additions and 10607 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tools/library/scripts/pycutlass/README.md
+2 -9
View File
@@ -85,18 +85,11 @@ You can run the PyCUTLASS on NGC PyTorch container.
```shell
docker run --gpus all -it --rm nvcr.io/nvidia/pytorch:22.09-py3
```
PyCUTLASS requires additional dependency Boost C++ library, which can be installed with
```bash
apt-get update
apt-get -y install libboost-all-dev
```
### Environment variables
PyCUTLASSS requires two environment variables:
* `CUTLASS_PATH`: the root directory of CUTLASS
* `CUDA_INSTALL_PATH`: the directory where cuda toolkit is installed
* `CUTLASS_PATH`: the root directory of CUTLASS. You can set this from the location at which you cloned CUTLASS via: `export CUTLASS_PATH=$(pwd)`.
* `CUDA_INSTALL_PATH`: the directory where cuda toolkit is installed. If running in bash with `nvcc` installed under a CUDA toolkit, you can set this to the location of your `nvcc` installation via: `export CUDA_INSTALL_PATH=$(which nvcc | awk -F'/bin/nvcc' '{print $1}')`
After setting these two environment variables, PyCUTLASS can be installed with
```shell