Philipp Moritz

[Examples] Ray parallel webcrawling example

1

## Why are these changes needed? ## Related issue number ## Checks - [ ] I've run `scripts/format.sh` to lint the changes in this PR. - [ ] I've included...

Add support for pulling docker images on Apple Silicon

8

## PR Checklist Please check if your PR fulfills the following requirements: - [ ] Tests for the changes have been added (for bug fixes / features) - [ ]...

Improve Ray client error message when exception can't be unpickled

1

## Why are these changes needed? This can for example happen if there is a version mismatch between libraries used. We should still try to surface the error message in...

Pushing individual layers of docker images

2

# 🚀 feature request ### Relevant Rules This is about the `container_push` rule. ### Description There is a TODO item in https://github.com/bazelbuild/rules_docker/blob/2b35b2dd56f0be6cc6b8df957332a31435f6b3ce/container/push.bzl#L41 to push by later. What would be needed...

Set BAD_REQUEST error code for 400 error instead of UNKNOWN

1

*Issue #, if available:* This fixes https://github.com/aws/aws-sdk-cpp/issues/1843 More crucially, it removes the current inconsistency between boto3 and the AWS CLI on the one hand and the C++ SDK on the...

## Why are these changes needed? ## Related issue number ## Checks - [ ] I've signed off every commit(by using the -s flag, i.e., `git commit -s`) in this...

[Core] Only instantiate gcs channels on driver

1

## Why are these changes needed? After https://github.com/ray-project/ray/pull/36293 is merged we can now fix this. This fixes https://github.com/ray-project/ray/issues/36001 -- I can repro both the hanging and the fact that this...

Memory leak when using CUDA Graph with torch.distributed.all_reduce (vLLM default config)

4

Running the following on the latest vLLM master ``` python -m vllm.entrypoints.openai.api_server --model mistralai/Mixtral-8x7B-Instruct-v0.1 --tensor-parallel-size 8 --max-num-batched-tokens 32768 --max-num-seqs 192 ``` and then ``` git clone https://github.com/ray-project/llmperf cd llmperf export...

Sparse matrix vector product benchmark (in comparison with spark and dask)

3

The numbers on the 3Mx3M test matrix from https://snap.stanford.edu/data/com-Orkut.html look like this: scipy.sparse single threaded: 1.2s halo (1 node, 4 workers): 600ms halo (2 nodes, 4 workers each): 430ms dask...

Steps to reproduce: 1. Install https://gym.openai.com/ 2. Execute the following code in the shell: ``` python from gym.envs import make @ray.remote([], []) def f(): env = make("Pong-ram-v0") return f() ```...