LM/vLLM - Revision history

Tacoball: /* vLLM and oneccl mapping */

2026-10-26T09:49:25Z

vLLM and oneccl mapping

2026-10-26T09:41:05Z

vLLM and oneccl mapping

2026-10-26T09:40:01Z

vLLM and oneccl mapping

← Older revision		Revision as of 09:40, 26 October 2026
Line 200:		Line 200:

	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">
			# pip
	sudo docker exec vllm-xpu-test pip show oneccl		sudo docker exec vllm-xpu-test pip show oneccl
	sudo docker exec vllm-xpu-test ~~pip show oneccl~~
			# native
			sudo docker exec vllm-xpu-test bash -lc '
			grep -aoE "oneCCL[^[:cntrl:]]{0,100}\|CCL[_ -]?version[^[:cntrl:]]{0,100}\|202[0-9]\.[0-9]+(\.[0-9]+)?" \
			/opt/venv/lib/libccl.so.2 \| sort -u \| head -n 50
			'
	</syntaxhighlight>		</syntaxhighlight>

2026-10-26T09:30:21Z

vLLM and oneccl mapping

← Older revision		Revision as of 09:30, 26 October 2026
Line 198:		Line 198:
	\| 0.30.0 \|\| 2022.0.0 \|\|		\| 0.30.0 \|\| 2022.0.0 \|\|
	\|}		\|}

			<syntaxhighlight lang="bash">
			sudo docker exec vllm-xpu-test pip show oneccl
			sudo docker exec vllm-xpu-test pip show oneccl
			</syntaxhighlight>

	== TODO ==		== TODO ==

2026-10-26T09:27:36Z

CCL Optimization

2026-10-14T02:17:28Z

k8v4

← Older revision		Revision as of 02:17, 14 October 2026
Line 158:		Line 158:
	source /opt/intel/oneapi/setvars.sh		source /opt/intel/oneapi/setvars.sh

			# 疑似 AI 誤導
	# 3. 安裝 vLLM 0.30（XPU 版本）		# 3. 安裝 vLLM 0.30（XPU 版本）
	pip install vllm==0.30.0		pip install vllm==0.30.0
Line 165:		Line 166:
	cd k8v4-xpu		cd k8v4-xpu

			# 疑似 AI 誤導
	# 5. 編譯 KV Cache 核心庫（libxe2_kv.so）		# 5. 編譯 KV Cache 核心庫（libxe2_kv.so）
	# 專案包含 C/C++ 核心，需本機編譯		# 專案包含 C/C++ 核心，需本機編譯

2026-10-13T10:55:51Z

TODO

← Older revision		Revision as of 10:55, 13 October 2026
Line 192:		Line 192:

	* [https://hub.docker.com/r/vllm/vllm-openai-xpu/tags Official Images for XPU]		* [https://hub.docker.com/r/vllm/vllm-openai-xpu/tags Official Images for XPU]
			* [https://docs.vllm.ai/en/stable/getting_started/installation/gpu/#create-a-new-python-environment Install using Python]
	* [https://docs.vllm.ai/en/stable/cli/serve/ vllm serve arguments]		* [https://docs.vllm.ai/en/stable/cli/serve/ vllm serve arguments]
	* [https://huggingface.co/docs/huggingface_hub/package_reference/environment_variables HF_* Environments]		* [https://huggingface.co/docs/huggingface_hub/package_reference/environment_variables HF_* Environments]

2026-10-13T10:25:51Z

CCL Optimization

@@ Line 146: / Line 146: @@
 vllm-xpu-qwen38-prod  | 2026:09:30-07:47:41:  117:[0] |CCL_WARN| pidfd is not supported, fallbacks to drmfd exchange mode
 </pre>
 == CCL Optimization ==

2026-09-30T19:04:22Z

CCL Optimization

2026-09-30T19:04:12Z

TODO

← Older revision		Revision as of 19:04, 30 September 2026
Line 146:		Line 146:
	vllm-xpu-qwen38-prod \| 2026:09:30-07:47:41: 117:[0] \|CCL_WARN\| pidfd is not supported, fallbacks to drmfd exchange mode		vllm-xpu-qwen38-prod \| 2026:09:30-07:47:41: 117:[0] \|CCL_WARN\| pidfd is not supported, fallbacks to drmfd exchange mode
	</pre>		</pre>

			== CCL Optimization ==


	== TODO ==		== TODO ==