Skip to content

Commit c8c6f1a

Browse files
committed
update guide
1 parent 4f3c27b commit c8c6f1a

1 file changed

Lines changed: 20 additions & 0 deletions

File tree

docs/backend/SYCL.md

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -51,6 +51,12 @@ The packages for FP32 and FP16 would have different accuracy and performance on
5151

5252
## News
5353

54+
- 2026.04
55+
56+
- Optimize mul_mat by reorder feature for data type: Q4_K, Q5_K, Q_K, Q8_0.
57+
- Fused MoE.
58+
- Upgrate CI and built package for oneAPI 2025.3.3, support Ubuntu 24.04 built package.
59+
5460
- 2026.03
5561
- Support Flash-Attention: less memory usage, performance impact depends on LLM.
5662

@@ -349,6 +355,12 @@ Choose one of following methods to run.
349355
./examples/sycl/test.sh
350356
```
351357

358+
- Run llama-server:
359+
360+
```sh
361+
./examples/sycl/start-svr.sh -m PATH/MODEL_FILE
362+
```
363+
352364
2. Command line
353365
Launch inference
354366

@@ -637,10 +649,18 @@ Choose one of following methods to run.
637649

638650
1. Script
639651

652+
- Run test:
653+
640654
```
641655
examples\sycl\win-test.bat
642656
```
643657

658+
- Run llama-server:
659+
660+
```
661+
examples\sycl\win-start-svr.bat -m PATH\MODEL_FILE
662+
```
663+
644664
2. Command line
645665

646666
Launch inference

0 commit comments

Comments
 (0)