enable ci test: upstream ci for XPU #4

DiweiSun · 2025-04-09T09:40:56Z

Motivation

Modifications

Checklist

Format your code according to the Code Formatting with Pre-Commit.
Add unit tests as outlined in the Running Unit Tests.
Update documentation / docstrings / example tutorials as needed, according to Writing Documentation.
Provide throughput / latency benchmark results and accuracy evaluation results as needed, according to Benchmark and Profiling and Accuracy Results.
For reviewers: If you haven't made any contributions to this PR and are only assisting with merging the main branch, please remove yourself as a co-author when merging the PR.
Please feel free to join our Slack channel at https://slack.sglang.ai to discuss your PR.

DiweiSun · 2025-04-09T10:14:30Z

Test Results for PR #4

test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_eager_allreduce PASSED [ 50%]
test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_graph_allreduce PASSED [100%]

============================== 2 passed in 3.90s ===============================
============================================
Final test result: success
============================================
All tests completed successfully
ci_sglang_cpu

Full log available in Jenkins build: https://cje-fm-owrp-prod03.devtools.intel.com/stag-aia-fwkval/job/sglang_postci_regular/71/

DiweiSun · 2025-04-17T03:32:08Z

Test Results for PR #4 ✅

Build Status: ✅ SUCCESS

Benchmark Results

Metric	Value
Total Latency	1.320 s
Median Latency	0.03939 s
Median Throughput	50.78 tokens/s

Unit Test Results

============================= test session starts ==============================
platform linux -- Python 3.12.9, pytest-8.3.5, pluggy-1.5.0 -- /root/miniforge3/bin/python3.12
cachedir: .pytest_cache
rootdir: /root/sglang
plugins: anyio-4.9.0
collecting ... collected 2 items

test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_eager_allreduce PASSED [ 50%]
test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_graph_allreduce PASSED [100%]

============================== 2 passed in 3.85s ===============================

Summary: 2 passed, 0 failed

✅ All tests completed successfully

ci_sglang_cpu

```

[View Full Build Log](https://cje-fm-owrp-prod03.devtools.intel.com/stag-aia-fwkval/job/sglang_postci_regular/91/)

DiweiSun · 2025-04-22T03:05:49Z

Test Results for PR #4 ✅

Build Status: SUCCESS

Benchmark Results

Metric	Value
Total Latency	N/A s
Total Throughput	N/A tokens/s
Median Latency	N/A s
Median Throughput	N/A tokens/s

Unit Test Results

Passed: 0
Failed: 0

View Full Build Log

DiweiSun · 2025-04-22T03:11:48Z

## Test Results for PR #4 ✅
**Build Status:** SUCCESS

### Benchmark Results
| Metric | Value |
|--------|-------|
| Total Latency | N/A s |
| Total Throughput | N/A tokens/s |
| Median Latency | N/A s |
| Median Throughput | N/A tokens/s |

### Unit Test Results
**Passed:** 6  
**Failed:** 2

```text
+ echo '| Median Latency | 0.04068 s |'

echo '| Median Throughput | 49.17 tokens/s |'
echo ''
echo '## Unit Test Results'
echo ''
docker exec -w /root/sglang ci_sglang_cpu pytest -v test/srt/test_custom_allreduce.py
tee /home/pytorch/workspace/sglang_postci_regular/sglang/pytest.log
============================= test session starts ==============================
platform linux -- Python 3.12.9, pytest-8.3.5, pluggy-1.5.0 -- /root/miniforge3/bin/python3.12
cachedir: .pytest_cache
rootdir: /root/sglang
plugins: anyio-4.9.0
collecting ... collected 2 items

test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_eager_allreduce PASSED [ 50%]
test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_graph_allreduce PASSED [100%]

============================== 2 passed in 3.67s ===============================
++ grep -c PASSED /home/pytorch/workspace/sglang_postci_regular/sglang/pytest.log

PASSED_TESTS=2
++ grep -c FAILED /home/pytorch/workspace/sglang_postci_regular/sglang/pytest.log
++ true
FAILED_TESTS=0
echo '```text'
cat /home/pytorch/workspace/sglang_postci_regular/sglang/pytest.log
echo '```'
echo ''
echo 'Summary: 2 passed, 0 failed'
echo ''
echo ---
'[' success = success ']'
echo '✅ All tests completed successfully'
echo ============================================
============================================
cat /home/pytorch/workspace/sglang_postci_regular/sglang/test_results.md

Test Results

Docker Build Results

✅ Docker build succeeded

Benchmark Results

Metric	Value
Total Latency	1.341 s
Total Throughput	95.43 tokens/s
Median Latency	0.04068 s
Median Throughput	49.17 tokens/s

Unit Test Results

============================= test session starts ==============================
platform linux -- Python 3.12.9, pytest-8.3.5, pluggy-1.5.0 -- /root/miniforge3/bin/python3.12
cachedir: .pytest_cache
rootdir: /root/sglang
plugins: anyio-4.9.0
coll
    ```
    
    [View Full Build Log](https://cje-fm-owrp-prod03.devtools.intel.com/stag-aia-fwkval/job/sglang_postci_regular/99/)

DiweiSun · 2025-04-22T03:37:47Z

## Test Results for PR #4 ✅
**Build Status:** SUCCESS

### Benchmark Results
| Metric | Value |
|--------|-------|
| Total Latency | N/A s |
| Total Throughput | N/A tokens/s |
| Median Latency | N/A s |
| Median Throughput | N/A tokens/s |

### Unit Test Results
**Passed:** 6  
**Failed:** 2

```text
+ echo '| Median Latency | 0.03873 s |'

echo '| Median Throughput | 51.64 tokens/s |'
echo ''
echo '## Unit Test Results'
echo ''
docker exec -w /root/sglang ci_sglang_cpu pytest -v test/srt/test_custom_allreduce.py
tee /home/pytorch/workspace/sglang_postci_regular/sglang/pytest.log
============================= test session starts ==============================
platform linux -- Python 3.12.9, pytest-8.3.5, pluggy-1.5.0 -- /root/miniforge3/bin/python3.12
cachedir: .pytest_cache
rootdir: /root/sglang
plugins: anyio-4.9.0
collecting ... collected 2 items

test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_eager_allreduce PASSED [ 50%]
test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_graph_allreduce PASSED [100%]

============================== 2 passed in 3.82s ===============================
++ grep -c PASSED /home/pytorch/workspace/sglang_postci_regular/sglang/pytest.log

PASSED_TESTS=2
++ grep -c FAILED /home/pytorch/workspace/sglang_postci_regular/sglang/pytest.log
++ true
FAILED_TESTS=0
echo '```text'
cat /home/pytorch/workspace/sglang_postci_regular/sglang/pytest.log
echo '```'
echo ''
echo 'Summary: 2 passed, 0 failed'
echo ''
echo ---
'[' success = success ']'
echo '✅ All tests completed successfully'
echo ============================================
============================================
cat /home/pytorch/workspace/sglang_postci_regular/sglang/test_results.md

Test Results

Docker Build Results

✅ Docker build succeeded

Benchmark Results

Metric	Value
Total Latency	1.282 s
Total Throughput	99.84 tokens/s
Median Latency	0.03873 s
Median Throughput	51.64 tokens/s

Unit Test Results

============================= test session starts ==============================
platform linux -- Python 3.12.9, pytest-8.3.5, pluggy-1.5.0 -- /root/miniforge3/bin/python3.12
cachedir: .pytest_cache
rootdir: /root/sglang
plugins: anyio-4.9.0
coll
    ```
    
    [View Full Build Log](https://cje-fm-owrp-prod03.devtools.intel.com/stag-aia-fwkval/job/sglang_postci_regular/101/)

DiweiSun · 2025-04-22T10:01:22Z

Test Results

Docker Build Results

✅ Docker build succeeded

Benchmark Results

Metric	Value
Total Latency	1.373 s
Total Throughput	93.22 tokens/s
Median Latency	0.04128 s
Median Throughput	48.45 tokens/s

Unit Test Results

============================= test session starts ==============================
platform linux -- Python 3.12.9, pytest-8.3.5, pluggy-1.5.0 -- /root/miniforge3/bin/python3.12
cachedir: .pytest_cache
rootdir: /root/sglang
plugins: anyio-4.9.0
collecting ... collected 2 items

test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_eager_allreduce PASSED [ 50%]
test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_graph_allreduce PASSED [100%]

============================== 2 passed in 3.83s ===============================

Summary: 2 passed, 0 failed

✅ All tests completed successfully

DiweiSun · 2025-04-22T10:01:24Z

Test Results

Docker Build Results

✅ Docker build succeeded

Benchmark Results

Metric	Value
Total Latency	1.373 s
Total Throughput	93.22 tokens/s
Median Latency	0.04128 s
Median Throughput	48.45 tokens/s

Unit Test Results

============================= test session starts ==============================
platform linux -- Python 3.12.9, pytest-8.3.5, pluggy-1.5.0 -- /root/miniforge3/bin/python3.12
cachedir: .pytest_cache
rootdir: /root/sglang
plugins: anyio-4.9.0
collecting ... collected 2 items

test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_eager_allreduce PASSED [ 50%]
test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_graph_allreduce PASSED [100%]

============================== 2 passed in 3.83s ===============================

Summary: 2 passed, 0 failed

✅ All tests completed successfully

DiweiSun · 2025-04-22T10:04:29Z

Test Results

Docker Build Results

✅ Docker build succeeded

Benchmark Results

Metric	Value
Total Latency	1.373 s
Total Throughput	93.22 tokens/s
Median Latency	0.04128 s
Median Throughput	48.45 tokens/s

Unit Test Results

============================= test session starts ==============================
platform linux -- Python 3.12.9, pytest-8.3.5, pluggy-1.5.0 -- /root/miniforge3/bin/python3.12
cachedir: .pytest_cache
rootdir: /root/sglang
plugins: anyio-4.9.0
collecting ... collected 2 items

test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_eager_allreduce PASSED [ 50%]
test/srt/test_custom_allreduce.py::TestCustomAllReduce::test_graph_allreduce PASSED [100%]

============================== 2 passed in 3.83s ===============================

Summary: 2 passed, 0 failed

✅ All tests completed successfully

…on (sgl-project#11161)

…project#11185)

Signed-off-by: Alex Chi Z <[email protected]>

…oject#11173)

…is enabled (sgl-project#10873) Co-authored-by: molocule <[email protected]>

Signed-off-by: Hao Lu <[email protected]>

…sgl-project#11192)

…oject#10892)

…ject#11452) Co-authored-by: yizhang2077 <[email protected]>

Co-authored-by: Lianmin Zheng <[email protected]>

Co-authored-by: Cheng Wan <[email protected]>

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

DiweiSun closed this Apr 9, 2025

DiweiSun reopened this Apr 9, 2025

DiweiSun changed the title ~~enable ci test~~ enable ci test: upstream ci for XPU Apr 10, 2025

This was referenced Apr 16, 2025

Enable CPU CI: upstream CI enabling with github workflow #3

Closed

Update README.md #1

Closed

DiweiSun force-pushed the main branch from 94e1912 to b3edfb2 Compare April 16, 2025 13:10

DiweiSun force-pushed the molly/xpu_ci branch from 48c9c73 to 8f30492 Compare April 16, 2025 13:15

DiweiSun force-pushed the main branch from b3edfb2 to 0b05058 Compare May 12, 2025 02:52

DiweiSun force-pushed the main branch from 0b05058 to 0d50309 Compare May 26, 2025 08:26

DiweiSun force-pushed the main branch from c3485f2 to 7891bac Compare July 18, 2025 05:36

DiweiSun force-pushed the molly/xpu_ci branch from 0cdc0e0 to 883144d Compare August 22, 2025 06:07

b8zhong and others added 11 commits October 3, 2025 11:30

[1/n] Enable DCA CUDA graph capture (sgl-project#9537)

a2faf89

[Fix] Update to v0.1.5.post4 and refine HIP attention backend selecti…

b00a0c7

…on (sgl-project#11161)

[CI]] Tee server logs to both file and stdout/stderr using PIPE (sgl-…

de89ef4

…project#11185)

fix: radix cache memory accounting (sgl-project#10637)

1a31229

Signed-off-by: Alex Chi Z <[email protected]>

Tiny add PD disaggregation + DP attention test (sgl-project#11167)

6794d21

[router] Steaming support for MCP Tool Calls in OpenAI Router (sgl-pr…

34151f1

…oject#11173)

[Feature] Option to save model weights to CPU when memory saver mode …

8c57490

…is enabled (sgl-project#10873) Co-authored-by: molocule <[email protected]>

Add --thinking-mode to run_eval (sgl-project#11189)

d6777a7

Signed-off-by: Hao Lu <[email protected]>

[hot-fix] Fix CI break which caused by adding thinking_mode in eval (…

04b86b3

…sgl-project#11192)

Tiny move files to utils folder (sgl-project#11166)

fdc4e1e

Fix CUDA illegal memory access issues in speculative decoding (sgl-pr…

6ae3f05

…oject#10892)

merrymercy and others added 30 commits October 12, 2025 07:45

Fix unit tests (sgl-project#11503)

5a6ec8f

Bugfix: Fix Type consistency for KV indices in SWARadixCache (sgl-pro…

f5b34a5

…ject#11452) Co-authored-by: yizhang2077 <[email protected]>

doc: add doc for adding new models into nightly-ci (sgl-project#11443)

9f2b457

Co-authored-by: Lianmin Zheng <[email protected]>

[CI] fix lint (sgl-project#11509)

2157d12

Deprecate global_server_args_dict (sgl-project#11331)

1083e7e

chore: remove flashinfer cleanup cache (sgl-project#11514)

05f015f

fix: revert temporarily remove b200 tests (sgl-project#11515)

0ecb426

[Fix] Improve longbench prompt and other logics (sgl-project#11474)

0aa65f9

Sync changes on io_struct.py and deterministic ops (sgl-project#11498)

2ac46e9

[lint] Fix the lint issue (sgl-project#11516)

6cd2969

Revert "Deprecate global_server_args_dict" (sgl-project#11520)

1bdd010

Improve dp attention port assignment scheme (sgl-project#5889)

a20e7df

Co-authored-by: Cheng Wan <[email protected]>

fix format issue

64b44c3

Update test/srt/run_suite.py

90fdd31

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

Update test/srt/test_intel_xpu_backend.py

c3eb182

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

Update test/srt/test_intel_xpu_backend.py

8b64cde

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>

fix xpu pipeline

1c71d0b

fix typo

70e5732

[router] openai router: support grok model (sgl-project#11511)

63e8435

Merge branch 'main' into molly/xpu_ci

542cb56

revert changes on w4afp8

aaa3071

Update w4afp8.py

d7d369d

refine format

0963bed

label TODO for embedding kernel

61036a4

create test/xpu for xpu specific cases

e0c40cf

fix typo

53c8e64

Update pr-test-xpu.yml

abe1cd3

fix xpu ci pipeline

ca45bd8

refine pipeline to utilize hf cache

5b625f7

format fix

53f9403

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

enable ci test: upstream ci for XPU #4

enable ci test: upstream ci for XPU #4

Uh oh!

DiweiSun commented Apr 9, 2025

Uh oh!

DiweiSun commented Apr 9, 2025

Uh oh!

DiweiSun commented Apr 17, 2025 •

edited

Loading

Uh oh!

DiweiSun commented Apr 22, 2025

Uh oh!

DiweiSun commented Apr 22, 2025

Uh oh!

DiweiSun commented Apr 22, 2025

Uh oh!

DiweiSun commented Apr 22, 2025

Uh oh!

DiweiSun commented Apr 22, 2025

Uh oh!

DiweiSun commented Apr 22, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

78 participants

enable ci test: upstream ci for XPU #4

Are you sure you want to change the base?

enable ci test: upstream ci for XPU #4

Uh oh!

Conversation

DiweiSun commented Apr 9, 2025

Motivation

Modifications

Checklist

Uh oh!

DiweiSun commented Apr 9, 2025

Test Results for PR #4

Uh oh!

DiweiSun commented Apr 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Test Results for PR #4 ✅

Build Status: ✅ SUCCESS

Benchmark Results

Unit Test Results

✅ All tests completed successfully

Uh oh!

DiweiSun commented Apr 22, 2025

Test Results for PR #4 ✅

Benchmark Results

Unit Test Results

Uh oh!

DiweiSun commented Apr 22, 2025

Test Results

Docker Build Results

Benchmark Results

Unit Test Results

Uh oh!

DiweiSun commented Apr 22, 2025

Test Results

Docker Build Results

Benchmark Results

Unit Test Results

Uh oh!

DiweiSun commented Apr 22, 2025

Test Results

Docker Build Results

Benchmark Results

Unit Test Results

Uh oh!

DiweiSun commented Apr 22, 2025

Test Results

Docker Build Results

Benchmark Results

Unit Test Results

Uh oh!

DiweiSun commented Apr 22, 2025

Test Results

Docker Build Results

Benchmark Results

Unit Test Results

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

78 participants

DiweiSun commented Apr 17, 2025 •

edited

Loading