Dennis96 comments

Results 13 comments of


                                            Dennis96

Use a fraction gpu resource, fail to get response from manager

same with me. did u solve it?

[need help] Failed to create a deployment requesting 0.x GPU cores

@seanchen022 i had the same problem like u days ago. when i change the images/codes of gpu-manager newer, it solved. maybe the branch master, commit: f0669de works. @paragor by the...

gpu-manager run failed

i had save problem too. so i made this file(/var/lib/kubelet/device-plugins/kubelet_internal_checkpoint) before gpu-manager start; it works now

ValueError: Shapes (18,) and (255,) are incompatible

same with u bro, so do u solve it these days?

get "invalid device context" or "segmentation fault" problems in aarch64 machine

@mYmNeo @hzliangbin please help

Pod ignores limits.

in my case, if process running exceeds the limits, is will return CUDA Out of Memory, but vcuda-core won't

plndr-cp-lock": context deadline exceeded

save with me. i tried version 0.4.2 and 0.5.0, both end up restarting periodically

seems not working on newest A100 80G

save with u. A100 80G did not be detected by gpu-manager, but A100 SXM 40G works good

在kubeedge中部署device-plugin插件使用自定义resource，容器无法启动：no devices found allocated in local cache for pod xxx, container xxx, resource tencent.com/vcuda-core

how's now? can u solve it?

在kubeedge中部署device-plugin插件使用自定义resource，容器无法启动：no devices found allocated in local cache for pod xxx, container xxx, resource tencent.com/vcuda-core

i fixed it 1. before callPreStartContainerIfNeeded, u should allocate ur extend-resource. For some reason pre-allocate in there did not work. 2. the cgroup fs path in edgenode is also different,...