LLM/Intel Pro Arc B70/2026-06-20 - Revision history

Tacoball: /* Test Models */

2026-07-13T06:18:19Z

Test Models

← Older revision		Revision as of 06:18, 13 July 2026
Line 121:		Line 121:
	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">
	# test model		# test model
			./llama-server -hf google/gemma-4-2B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9004 # 90 t/s
	./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9004 # 90 t/s		./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9004 # 90 t/s
	./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9012 # 47 t/s		./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9012 # 47 t/s

2026-07-02T06:09:59Z

Install dependencies

← Older revision		Revision as of 06:09, 2 July 2026
Line 64:		Line 64:

	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">
	~~# setup apt source for OpenVINO~~
	~~# TODO~~

	# install oneAPI		# install oneAPI
	sudo apt install intel-oneapi-mkl-devel		sudo apt install intel-oneapi-mkl-devel

2026-06-22T15:06:54Z

Test SYCL(oneAPI) version

← Older revision		Revision as of 15:06, 22 June 2026
Line 60:		Line 60:

	== Test SYCL(oneAPI) version ==		== Test SYCL(oneAPI) version ==

			=== Install dependencies ===

	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">
Line 70:		Line 72:
	source /opt/intel/oneapi/setvars.sh		source /opt/intel/oneapi/setvars.sh
	</syntaxhighlight>		</syntaxhighlight>

			=== Link shared library ===

	<pre>		<pre>
Line 85:		Line 89:
	:: oneAPI environment initialized ::		:: oneAPI environment initialized ::
	</pre>		</pre>

			=== Build ===

	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">
Line 113:		Line 119:
	SYCL0: Intel(R) Arc(TM) Pro B70 Graphics (32656 MiB, 32128 MiB free)		SYCL0: Intel(R) Arc(TM) Pro B70 Graphics (32656 MiB, 32128 MiB free)
	</pre>		</pre>

			=== Test Models ===

	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">

2026-06-22T15:04:37Z

Test Vulkan version

2026-06-20T18:06:21Z

Test SYCL(oneAPI) version

← Older revision		Revision as of 18:06, 20 June 2026
Line 116:		Line 116:
	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">
	# test model		# test model
	./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0 # 90 t/s		./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9004 # 90 t/s
	./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 # 47 t/s		./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9012 # 47 t/s
	./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0 # 52 t/s		./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9026 # 52 t/s
	./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0 # 23 t/s		./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0 --port 9031 # 23 t/s
	</syntaxhighlight>		</syntaxhighlight>

2026-06-20T17:34:27Z

Test SYCL(oneAPI) version

← Older revision		Revision as of 17:34, 20 June 2026
Line 119:		Line 119:
	./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 # 47 t/s		./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 # 47 t/s
	./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0 # 52 t/s		./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0 # 52 t/s
	./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0		./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0 # 23 t/s
	</syntaxhighlight>		</syntaxhighlight>

2026-06-20T17:23:12Z

Test SYCL(oneAPI) version

← Older revision		Revision as of 17:23, 20 June 2026
Line 118:		Line 118:
	./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0 # 90 t/s		./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0 # 90 t/s
	./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 # 47 t/s		./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 # 47 t/s
	./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0		./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0 # 52 t/s
	./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0		./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0
	</syntaxhighlight>		</syntaxhighlight>

2026-06-20T17:10:37Z

Test SYCL(oneAPI) version

← Older revision		Revision as of 17:10, 20 June 2026
Line 116:		Line 116:
	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">
	# test model		# test model
	./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0		./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0 # 90 t/s
	./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0		./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0 # 47 t/s
	./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0		./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0
	./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0		./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0

2026-06-20T17:06:36Z

Test SYCL(oneAPI) version

← Older revision		Revision as of 17:06, 20 June 2026
Line 116:		Line 116:
	<syntaxhighlight lang="bash">		<syntaxhighlight lang="bash">
	# test model		# test model
	./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf -host 0.0.0.0 ~~-port 9999~~		./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf --host 0.0.0.0
	./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf -host 0.0.0.0 ~~-port 9999~~		./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf --host 0.0.0.0
	./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf -host 0.0.0.0 ~~-port 9999~~		./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf --host 0.0.0.0
	./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf -host 0.0.0.0 ~~-port 9999~~		./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf --host 0.0.0.0
	</syntaxhighlight>		</syntaxhighlight>

2026-06-20T17:04:27Z

Test SYCL(oneAPI) version

← Older revision		Revision as of 17:04, 20 June 2026
Line 117:		Line 117:
	# test model		# test model
	./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf -host 0.0.0.0 -port 9999		./llama-server -hf google/gemma-4-E4B-it-qat-q4_0-gguf -host 0.0.0.0 -port 9999
			./llama-server -hf google/gemma-4-12B-it-qat-q4_0-gguf -host 0.0.0.0 -port 9999
			./llama-server -hf google/gemma-4-26B-A4B-it-qat-q4_0-gguf -host 0.0.0.0 -port 9999
			./llama-server -hf google/gemma-4-31B-it-qat-q4_0-gguf -host 0.0.0.0 -port 9999
	</syntaxhighlight>		</syntaxhighlight>