fighterhit
fighterhit
`com.github.fridujo.rabbitmq.mock.exchange.FixDelayExchange` doesn't exist but the test import it in `src/test/java/com/github/fridujo/rabbitmq/mock/spring/SpringIntegrationTest.java`
I got a similar problem when I create a pod like [issue 18](https://github.com/tkestack/gpu-manager/issues/18). Please help analyze. ``` Warning UnexpectedAdmissionError 16m kubelet, ai-1080ti-62 Update plugin resources failed due to rpc error:...
**What is your proposal**: AI 任务一般都会用到 GPU,这类资源相对 CPU/内存来说往往更加昂贵,目前实现混部的开源方案有很多,基本是通过内核 cgroup 特性针对传统 CPU 任务来做,真正实现 GPU 混部的方案却很少。因此,实现支持 QoS based GPU 共享/虚拟化,即在支持 **故障隔离、显存隔离、算力隔离** 的同时 **兼顾 GPU 应用性能**,对混部来说是一种差异化竞争,意义重大。 **Why is this needed**: 目前业界已有的一些 GPU 共享/虚拟化方案:...
Why use `node.Status.Capacity` instead of `node.Status.Allocatable`? Doesn't the number of Allocatable determine whether it is schedulable? In this case, I feel that there will be a problem when `node.Status.Allocatable` is...
Hi @mYmNeo , I sometimes find that if a GPU pod is created while some GPU pods are being deleted or terminating, the `UnexpectedAdmissionError` will appear a little more frequently....
Hope to support logging in with a certificate in ssh
There is an 800MB object on one node. The actor on this node first calls `ref=ray.put(obj)`, and then passes the returned object reference to the actor on another node. The...
What is the performance difference between plasma and python multiprocessing shared_memory? If I want higher performance can I replace multiprocessing shared_memory with plasma?
Both the official json library and UltraJSON support these two functions, and I think orjson should also support them. I found a related issue #329, but it seems that these...
Unexpected Behavior: ModelInferRequest Fields Overwritten with Incorrect Values in Triton C++ Client
I use the `tritonserver:22.12-py3-sdk` NGC image and test it with the following code under `/workspace/install`, but the output seems like the fields are all overwritten by the last set value....