intel / intel/compute-runtime

[GSD-13006] UR_RESULT_ERROR_DEVICE_LOST and OUT_OF_DEVICE_MEMORY on dual Intel Arc B70 Pro (Battlemage) with Level Zero 1.15.38308

Open
#944 14 comments 0 reactions 0 assignees View on GitHub
OS: Linux Type: Bug
Dominant language
C++
Stars
1.4k
Forks
300
PR merge metrics
No merged PRs in 30d

Description

## Hardware
- **GPUs:** 2x Intel Arc B70 Pro (Battlemage G31), 32 GB each (64 GB total)
- **CPU:** Intel 12th Gen (Alder Lake)
- **OS:** Ubuntu 26.04 LTS, kernel 7.0.0-27

## Issue
Using the Level Zero API (via llama.cpp SYCL backend), the following errors occur:

1. **UR_RESULT_ERROR_DEVICE_LOST (error 20):** When attempting to allocate and initialize buffers on dual GPUs, the device connection is lost. This happens during memset operations after successful device enumeration. Even a single GPU can trigger this when larger allocations are made.

2. **UR_RESULT_ERROR_OUT_OF_DEVICE_MEMORY (error 39):** Occurs when loading models that should fit within available VRAM (64 GB total across 2 GPUs). The system reports the model at ~33 GB, but fails on allocation.

## Stack Trace
The crash originates in ggml_backend_sycl_buffer_clear during memset via a SYCL stream.

## Workaround
The Vulkan backend (Mesa 26.0.3) works correctly with dual GPUs on the same hardware. The issue appears specific to the Level Zero / SYCL path.

## Willing to help
Happy to provide additional logs, run diagnostic tools, or test patches.

Contributor guide

Open the contributing guide

Research direction

Start with the reported crash in ggml_backend_sycl_buffer_clear and the Level Zero/SYCL allocation and memset path, using the supplied hardware, stack trace, and diagnostics. Compare dual- and single-GPU behavior with the working Vulkan backend and gather the requested logs or diagnostic tool output. Done means model loading and buffer initialization no longer produce DEVICE_LOST or OUT_OF_DEVICE_MEMORY on the reported setup.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp, linux, ubuntu
Domain
backend, hpc, operating-systems
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.