[GH-ISSUE #127] Will disaggregated vllm with MooncakeStoreConnector support TP and PP？ #66

New issue

Closed

opened 2026-10-29 21:08:39 +00:00 by bsakaguchi · 5 comments

bsakaguchi commented

2026-10-29 21:08:39 +00:00

Owner

Originally created by @c-guo16 on GitHub (Mar 5, 2025).
Original GitHub issue: https://github.com/kvcache-ai/Mooncake/issues/127

it seems like that in current version, it only support multiple standalone instance, each on 1 gpu card.

Originally created by @c-guo16 on GitHub (Mar 5, 2025). Original GitHub issue: https://github.com/kvcache-ai/Mooncake/issues/127 it seems like that in current version, it only support multiple standalone instance, each on 1 gpu card.

bsakaguchi

2026-10-29 21:08:39 +00:00

closed this issue
added the
stale

auto-closed
labels

bsakaguchi commented

2026-10-29 21:08:40 +00:00

Author

Owner

@ShangmingCai commented on GitHub (Mar 19, 2025):

TP is supported, but PP is not supported currently. We will work with the vLLM community to see how we make pp works.

@ShangmingCai commented on GitHub (Mar 19, 2025): TP is supported, but PP is not supported currently. We will work with the vLLM community to see how we make pp works.

bsakaguchi commented

2026-10-29 21:08:40 +00:00

Author

Owner

@frankie-ys commented on GitHub (Aug 29, 2025):

TP is supported, but PP is not supported currently. We will work with the vLLM community to see how we make pp works.

Hi, is there any plan to make pp works? It seems that still not work in v0.3.5.

@frankie-ys commented on GitHub (Aug 29, 2025): > TP is supported, but PP is not supported currently. We will work with the vLLM community to see how we make pp works. Hi, is there any plan to make pp works? It seems that still not work in v0.3.5.

bsakaguchi commented

2026-10-29 21:08:40 +00:00

Author

Owner

@ShangmingCai commented on GitHub (Aug 29, 2025):

TP is supported, but PP is not supported currently. We will work with the vLLM community to see how we make pp works.

Hi, is there any plan to make pp works? It seems that still not work in v0.3.5.

@frankie-ys ~~Prefill PP is supported on the main branch now. Or you can try sglang v0.5.1? You don't need to upgrade Mooncake though, v0.3.5 is fine. Related PR: https://github.com/sgl-project/sglang/pull/8846~~

~~Another thing is you need to make sure the model supports PP originally, even without PD disaggregation.~~

MooncakeStoreConnector doesn't support PP now, we still need some time.

@ShangmingCai commented on GitHub (Aug 29, 2025): > > TP is supported, but PP is not supported currently. We will work with the vLLM community to see how we make pp works. > > Hi, is there any plan to make pp works? It seems that still not work in v0.3.5. @frankie-ys ~Prefill PP is supported on the main branch now. Or you can try sglang v0.5.1? You don't need to upgrade Mooncake though, v0.3.5 is fine. Related PR: https://github.com/sgl-project/sglang/pull/8846~ ~Another thing is you need to make sure the model supports PP originally, even without PD disaggregation.~ MooncakeStoreConnector doesn't support PP now, we still need some time.

bsakaguchi commented

2026-10-29 21:08:40 +00:00

Author

Owner

@github-actions[bot] commented on GitHub (Jul 22, 2026):

This issue has had no activity for 90 days and will be closed in 7 days if there is no further activity. Please comment or react if it should stay open.

@github-actions[bot] commented on GitHub (Jul 22, 2026): This issue has had no activity for 90 days and will be closed in 7 days if there is no further activity. Please comment or react if it should stay open.

bsakaguchi commented

2026-10-29 21:08:40 +00:00

Author

Owner

@github-actions[bot] commented on GitHub (Jul 29, 2026):

Closing due to 3 months of inactivity. If this is still relevant, please comment and we can reopen.

@github-actions[bot] commented on GitHub (Jul 29, 2026): Closing due to 3 months of inactivity. If this is still relevant, please comment and we can reopen.