Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Remove BIGDL_LLM_XMX_DISABLED in documentation #12445

Merged
merged 13 commits into from
Nov 27, 2024
+119 −1,443
Merged

Remove BIGDL_LLM_XMX_DISABLED in documentation #12445

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
e0c56c6
fix: remove BIGDL_LLM_XMX_DISABLED in mddocs
cranechu0131 Nov 26, 2024
e02b7a8
fix: remove set SYCL_CACHE_PERSISTENT=1 in example
cranechu0131 Nov 26, 2024
eafdbd7
fix: remove BIGDL_LLM_XMX_DISABLED in workflows
cranechu0131 Nov 26, 2024
e94bf11
fix: merge igpu and A-series Graphics
cranechu0131 Nov 26, 2024
ba4c1b6
fix: remove set BIGDL_LLM_XMX_DISABLED=1 in example
cranechu0131 Nov 26, 2024
737aa60
fix: remove BIGDL_LLM_XMX_DISABLED in workflows
cranechu0131 Nov 26, 2024
304f0fd
fix: merge igpu and A-series Graphics
cranechu0131 Nov 26, 2024
c799e63
Merge branch 'remove-BIGDL_LLM_XMX_DISABLED' of https://github.com/cr…
cranechu0131 Nov 26, 2024
63ed1e0
Merge branch 'main' into remove-BIGDL_LLM_XMX_DISABLED
cranechu0131 Nov 26, 2024
09679b6
Merge branch 'main' into remove-BIGDL_LLM_XMX_DISABLED
cranechu0131 Nov 27, 2024
fe9fc82
fix: textual adjustment
cranechu0131 Nov 27, 2024
e1948bf
fix: textual adjustment
cranechu0131 Nov 27, 2024
c1b8841
fix: textual adjustment
cranechu0131 Nov 27, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
34 changes: 0 additions & 34 deletions .github/workflows/llm_performance_tests.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -805,7 +805,6 @@ jobs:
call conda activate igpu-perf
if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -841,7 +840,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -875,7 +873,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -910,7 +907,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -947,7 +943,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1003,7 +998,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1039,7 +1033,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1073,7 +1066,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1108,7 +1100,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1145,7 +1136,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1200,7 +1190,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1236,7 +1225,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1270,7 +1258,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1305,7 +1292,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1342,7 +1328,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1397,7 +1382,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1435,7 +1419,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1474,7 +1457,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1522,7 +1504,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1572,7 +1553,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1642,7 +1622,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1680,7 +1659,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1717,7 +1695,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1756,7 +1733,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1814,7 +1790,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1852,7 +1827,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1888,7 +1862,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1925,7 +1898,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -1964,7 +1936,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -2021,7 +1992,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -2059,7 +2029,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -2095,7 +2064,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -2132,7 +2100,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down Expand Up @@ -2171,7 +2138,6 @@ jobs:

if "${{ matrix.platform }}"=="perf-mtl" (
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
)
if "${{ matrix.platform }}"=="perf-lnl" (
set SYCL_CACHE_PERSISTENT=1
Expand Down
4 changes: 0 additions & 4 deletions docs/mddocs/DockerGuides/docker_pytorch_inference_gpu.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -97,10 +97,6 @@ root@arda-arc12:/# sycl-ls
> # Reduce memory accesses by fusing SDP ops.
> # Recommended for use on Intel Data Center GPU Max Series.
> export ENABLE_SDP_FUSION=1
>
> # Disable XMX computation.
> # Recommended for use on integrated GPUs.
> export BIGDL_LLM_XMX_DISABLED=1
> ```


Expand Down
9 changes: 1 addition & 8 deletions docs/mddocs/Overview/install_gpu.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -150,13 +150,7 @@ call "C:\Program Files (x86)\Intel\oneAPI\setvars.bat"

Please also set the following environment variable if you would like to run LLMs on: -->

- For **Intel iGPU**:
```cmd
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
```

- For **Intel Arc™ A-Series Graphics**:
- For **Intel iGPU** and **Intel Arc™ A-Series Graphics**:
```cmd
set SYCL_CACHE_PERSISTENT=1
```
Expand Down Expand Up @@ -596,7 +590,6 @@ To use GPU acceleration on Linux, several environment variables are required or
source /opt/intel/oneapi/setvars.sh

export SYCL_CACHE_PERSISTENT=1
export BIGDL_LLM_XMX_DISABLED=1
```

> [!NOTE]
Expand Down
19 changes: 2 additions & 17 deletions docs/mddocs/Quickstart/benchmark_quickstart.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,34 +80,19 @@ Some parameters in the yaml file that you can configure:

Please refer to [here](../Overview/install_gpu.md#runtime-configuration) to configure oneAPI environment variables. Choose corresponding commands base on your device.

- For **Intel iGPU**:
cranechu0131 marked this conversation as resolved.
Show resolved Hide resolved
- For **Intel iGPU** and **Intel Arc™ A-Series Graphics**:

```bash
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1

python run.py
```

- For **Intel Arc™ A300-Series or Pro A60**:

```bash
set SYCL_CACHE_PERSISTENT=1
python run.py
```

- For **Other Intel dGPU Series**:

```bash
# e.g. Arc™ A770
python run.py
```

## Run on Linux

Please choose corresponding commands base on your device.

- For **Intel Arc™ A-Series and Intel Data Center GPU Flex**:
- For **Intel Arc™ A-Series** and **Intel Data Center GPU Flex**:

For Intel Arc™ A-Series Graphics and Intel Data Center GPU Flex Series, we recommend:

Expand Down
1 change: 0 additions & 1 deletion docs/mddocs/Quickstart/install_linux_gpu.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -422,7 +422,6 @@ To use GPU acceleration on Linux, several environment variables are required or
source /opt/intel/oneapi/setvars.sh

export SYCL_CACHE_PERSISTENT=1
export BIGDL_LLM_XMX_DISABLED=1
```

> [!NOTE]
Expand Down
1 change: 0 additions & 1 deletion docs/mddocs/Quickstart/install_linux_gpu.zh-CN.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -390,7 +390,6 @@ conda activate llm
source /opt/intel/oneapi/setvars.sh

export SYCL_CACHE_PERSISTENT=1
export BIGDL_LLM_XMX_DISABLED=1
```

> [!NOTE]
Expand Down
11 changes: 2 additions & 9 deletions docs/mddocs/Quickstart/install_windows_gpu.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,19 +109,12 @@ You can verify if `ipex-llm` is successfully installed following below steps.

- Set the following environment variables according to your device:

- For **Intel iGPU**:
- For **Intel iGPU** and **Intel Arc™ A770**:

```cmd
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
```

- For **Intel Arc™ A770**:

```cmd
set SYCL_CACHE_PERSISTENT=1
```


> [!TIP]
> For other Intel dGPU Series, please refer to [this guide](../Overview/install_gpu.md#runtime-configuration) for more details regarding runtime configuration.

Expand Down
9 changes: 1 addition & 8 deletions docs/mddocs/Quickstart/install_windows_gpu.zh-CN.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -108,14 +108,7 @@ conda activate llm

- 根据你的设备，设置以下环境参数：

- **Intel iGPU**:

```cmd
set SYCL_CACHE_PERSISTENT=1
set BIGDL_LLM_XMX_DISABLED=1
```

- **Intel Arc™ A770**:
- **Intel iGPU** and **Intel Arc™ A770**:

```cmd
set SYCL_CACHE_PERSISTENT=1
Expand Down
Loading