https://errorvault.dev/tags/compute-workloads/ Recent content in Compute-Workloads on ErrorVault — Developer Error Code Dictionary Hugo en-us Fri, 05 Jun 2026 03:08:04 +0800 https://errorvault.dev/openclaw/openclaw-clw-gpu-crash-gpu-memory-segmentation-fault/ Fri, 05 Jun 2026 03:08:04 +0800 https://errorvault.dev/openclaw/openclaw-clw-gpu-crash-gpu-memory-segmentation-fault/ <h2 id="1-symptoms">1. Symptoms</h2> <p>The <code>clw-gpu-crash</code> error occurs when an OpenClaw compute workload encounters a critical failure at the GPU level. This manifests as an abrupt termination of the GPU computation process, often leaving the device in an undefined state.</p> <h3 id="observable-symptoms">Observable Symptoms</h3> <p>The most common symptoms reported by developers include:</p> <ul> <li><strong>Sudden process termination</strong>: The OpenClaw worker process exits with a non-zero exit code immediately after launching GPU kernels.</li> <li><strong>Device becomes unresponsive</strong>: After the crash, subsequent GPU operations return <code>CL_DEVICE_NOT_AVAILABLE</code> or similar errors until the device is reset.</li> <li><strong>dmesg kernel errors</strong>: On Linux systems, the kernel ring buffer may contain entries indicating GPU memory access violations:</li> </ul> <div class="highlight"><pre tabindex="0" style="color:#f8f8f2;background-color:#282a36;-moz-tab-size:4;-o-tab-size:4;tab-size:4;-webkit-text-size-adjust:none;"><code class="language-fallback" data-lang="fallback"><span style="display:flex;"><span>[ 123.456789] NVRM: Xid (PCI:0000:01:00): GPU Crash, reason: GF100 </span></span><span style="display:flex;"><span>[ 123.456890] NVRM: Xid (PCI:0000:01:00): GPU memory access violation at address 0x12345678 </span></span><span style="display:flex;"><span>[ 123.456891] NVRM: Xid (PCI:0000:01:00): - GPU 0000:01:00.0: GPU has fallen off the bus </span></span></code></pre></div><ul> <li><strong>Error log output</strong>: The OpenClaw runtime emits the following error message:</li> </ul> <div class="highlight"><pre tabindex="0" style="color:#f8f8f2;background-color:#282a36;-moz-tab-size:4;-o-tab-size:4;tab-size:4;-webkit-text-size-adjust:none;"><code class="language-gdscript3" data-lang="gdscript3"><span style="display:flex;"><span>[ERROR] OpenClaw Worker: clw<span style="color:#ff79c6">-</span>gpu<span style="color:#ff79c6">-</span>crash detected </span></span><span style="display:flex;"><span>[ERROR] Device: NVIDIA Tesla T4 (ID: <span style="color:#bd93f9">0</span>) </span></span><span style="display:flex;"><span>[ERROR] Workload: matrix_multiply_v2<span style="color:#ff79c6">.</span>clw </span></span><span style="display:flex;"><span>[ERROR] Crash type: GPU_MEMORY_SEGFAULT </span></span><span style="display:flex;"><span>[ERROR] Context dump saved to: <span style="color:#ff79c6">/</span><span style="color:#ff79c6">var</span><span style="color:#ff79c6">/</span><span style="color:#8be9fd;font-style:italic">log</span><span style="color:#ff79c6">/</span>openclaw<span style="color:#ff79c6">/</span>crash_20241230_143255<span style="color:#ff79c6">.</span>dmp </span></span></code></pre></div><ul> <li><strong>Partial results</strong>: In some cases, the GPU may have completed a portion of the workload before crashing, leaving partial output in device memory.</li> <li><strong>Timeout behavior</strong>: If Watchdog timers are enabled, the system may report a kernel execution timeout before the crash is officially detected.</li> </ul> <h3 id="secondary-symptoms">Secondary Symptoms</h3> <p>After a <code>clw-gpu-crash</code>, you may observe:</p>