silenceli

创建 sleep.yaml 拦不住， webhook deployment 的 pod 会报这个错误： ``` 2019/10/25 09:33:01 http: TLS handshake error from 10.244.0.0:10131: remote error: tls: bad certificate ```

I meet the some problem. I don't known how to solve

> Not sure it is K8s doing it or something in the package. When watch resource like validatingwebhookconfiguration, it return from the watch.Next with EOF every 40 min , but...

1. ssh `` 2. `cp /etc/kubernetes/manifests/kube-scheduler.yaml /etc/kubernetes/` 3. modify `/etc/kubernetes/kube-scheduler.yaml` as the doc metions 4. `cp /etc/kubernetes/kube-scheduler.yaml /etc/kubernetes/manifests/` then, k8s will kill old scheduler pod, and start an new scheduler...

我之前也遇到过这个问题，目前的版本似乎只支持 cuda-12.6，其他的版本可能 hook 的头文件略有不同，将导致问题。 I have encountered this problem before. The current version seems to only support cuda-12.6. Other versions may have slightly different hook header files, which will cause...

I see that cudnn, cublas, and cudart libs are all hijacked in the code. According to my understanding, these requests will eventually reach libcuda.so. If we only hijack all interfaces...

@Fruneng 我也有这方面的需求，我在考虑如何将 Scuda 与 HAMi-core 进行集成，使得 gpu 具备池化的能力。 I also have needs in this regard. I am thinking about how to integrate Scuda with HAMi-core so that the gpu has...

> [@silenceli](https://github.com/silenceli) 太好了 我们可以讨论一下如何实现 可以加个微信聊一聊 :-) ，微信名：silenceli_1988