Thomas Huber

Thanks. Seems to work fine... but just found out that 12GB VRAM of the B580 is not enough for E2B .

But E4B runs fine on the B60. Thanks a lot for your help!

Hi.
I've already pinpointed the error:
My host has i7-12700k with iGPU, one ARC B60 and one ARC B580.

I get the error when I set export ONEAPI_DEVICE_SELECTOR="level_zero:0" to use the B60.
When I unset ONEAPI_DEVICE_SELECTOR vllm serves fine.

What would be the correct way to select the B60 or B580 GPU ?

sycl-ls
[level_zero:gpu][level_zero:0] Intel(R) oneAPI Unified Runtime over Level-Zero V2, Intel(R) Arc(TM) Pro B60 Graphics 20.1.0 [1.14.36300+8]
[level_zero:gpu][level_zero:1] Intel(R) oneAPI Unified Runtime over Level-Zero V2, Intel(R) Arc(TM) B580 Graphics 20.1.0 [1.14.36300+8]
[level_zero:gpu][level_zero:2] Intel(R) oneAPI Unified Runtime over Level-Zero V2, Intel(R) UHD Graphics 770 12.2.0 [1.14.36300+8]
[opencl:cpu][opencl:0] Intel(R) OpenCL, 12th Gen Intel(R) Core(TM) i7-12700K OpenCL 3.0 (Build 0) [2026.20.3.0.19_160000.xmain-hotfix]
[opencl:gpu][opencl:1] Intel(R) OpenCL Graphics, Intel(R) Arc(TM) Pro B60 Graphics OpenCL 3.0 NEO  [25.48.36300.8]
[opencl:gpu][opencl:2] Intel(R) OpenCL Graphics, Intel(R) Arc(TM) B580 Graphics OpenCL 3.0 NEO  [25.48.36300.8]
[opencl:gpu][opencl:3] Intel(R) OpenCL Graphics, Intel(R) UHD Graphics 770 OpenCL 3.0 NEO  [25.48.36300.8]

I using Ubuntu 24.04.4 LTS with most recent intel drivers etc.

I can not get E2B or E4B running on my Arc B60.
sycl-ls shows the GPU inside the Docker container but running vllm serve throws
terminate called after throwing an instance of 'sycl::_V1::exception'
I tried with the exact 66e86f1 version and with main