Bug#1094164: slurm-wlm-rsmi-plugin: rsmi uses librocm_smi64.so, only provided by librocm-smi-dev but not librocm-smi64-1

2025-01-25 Thread Yiyang Wu
Package: slurm-wlm-rsmi-plugin Version: 22.05.8-4+deb12u2 Severity: normal Tags: upstream X-Debbugs-Cc: Benda Xu , xgreenlandfor...@gmail.com slurm-wlm-rsmi-plugin depends on librocm-smi64-1 which provides librocm_smi64.so.1 but not librocm_smi64.so. However, slurm uses dlopen to load librocm_smi6

Bug#1090072: linux-image-amd64: Enable P2P and HMM feature for AMDGPU

2025-01-06 Thread Yiyang Wu
;d say multi-GPU setup is still only common in workstation and servers. I guess there is a trend people put more GPU devices into server for accelerated computing. 2. Enabling CONFIG_HSA_AMD_SVM CONFIG_HSA_AMD_SVM is useful in general compute even on single GPU system: > On Mon, Dec 16, 2

Bug#1090072: linux-image-amd64: Enable P2P and HMM feature for AMDGPU

2024-12-15 Thread Yiyang Wu
of packages linux-image-amd64 depends on: pn linux-image-6.11.10-amd64 pn linux-image-6.12.3-amd64 pn linux-image-6.5.0-5-amd64 linux-image-amd64 recommends no packages. linux-image-amd64 suggests no packages. >From 468dc1822dcb3898203e773176ee414e6c534238 Mon Sep 17 00:00:00 2001

Bug#1090031: slurmctld: Cannot log to /var/log/slurm/slurmctld.log after logrotate due to reconfig failure by wrong pid ownership

2024-12-15 Thread Yiyang Wu
Package: slurmctld Version: 22.05.8-4+deb12u2 Severity: normal Tags: patch X-Debbugs-Cc: Benda Xu , xgreenlandfor...@gmail.com To reproduce on a non-systemd machine which uses /etc/init.d/slurmctld sudo logrotate -fv /etc/logrotate.d/slurmctld Result contains the following error messages. runni