README_HIP.md 1.5 KB
Newer Older
zhuwenwen's avatar
zhuwenwen committed
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
26
27
28
29
30
31
32
# FastFold

## 安装
安装FastFold,你需要
+ Python 3.8 或者 3.9.

### 使用pip安装
fastfold whl包下载目录:[https://cancon.hpccube.com:65024/4/main/fastfold/dtk23.04](https://cancon.hpccube.com:65024/4/main/fastfold/dtk23.04)
根据对应的pytorch版本和python版本,下载对应fastfold的whl包

```shell
pip install fastfold* (下载的fastfold的whl包)
```

### 使用源码安装

#### 编译环境准备
pytorch whl包下载目录:[https://cancon.hpccube.com:65024/4/main/pytorch/dtk23.04](https://cancon.hpccube.com:65024/4/main/pytorch/dtk23.04)
根据python版本,下载对应pytorch的whl包
```shell
pip install torch* (下载的torch的whl包)
```

```shell
pip install setuptools=59.5.0 wheel
```

#### 编译安装
```shell
git clone -b dtk-23.04_fastfold0.2.1 https://developer.hpccube.com/codes/aicomponent/fastfold
cd fastfold
export FASTFOLD_BUILD_VERSION=abix.dtkxxx
zhuwenwen's avatar
zhuwenwen committed
33
export CPLUS_INCLUDE_PATH=$ROCM_PATH/include:$ROCM_PATH/cuda/targets/x86_64-linux/include:$ROCM_PATH/hiprand/include:$ROCM_PATH/rocrand/include:$CPLUS_INCLUDE_PATH
zhuwenwen's avatar
zhuwenwen committed
34
35
36
37
38
39
40
python setup.py bdist_wheel 
pip install dist/fastfold*
```

## Note
+ 若使用 pip install 下载安装过慢,可添加源:-i https://pypi.tuna.tsinghua.edu.cn/simple/
+ FASTFOLD_BUILD_VERSION为编译的版本号设置,版本号为0.2.1+gitxxx.abix.dtkxxx 
zhuwenwen's avatar
zhuwenwen committed
41
42
gitxxx:为代码自动获取;abi0:使用devtools的gcc编译;abi1:使用非devtools的gcc编译; dtkxxx为dtk的版本号:例如:dtk2304
+ ROCM_PATH为dtk的路径,默认为/opt/dtkxxx