Skip to content

Rebase to Upstream #31

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
gmlwns2000 wants to merge 639 commits into from
Closed

Rebase to Upstream #31

gmlwns2000 wants to merge 639 commits into from

Conversation

@gmlwns2000
Copy link
Collaborator

@gmlwns2000 gmlwns2000 commented Sep 12, 2025

Rebase to Upstream

Fridge003 and others added 30 commits September 5, 2025 19:20
@Fridge003
Add lora_path argument to bench_multiturn.py (sgl-project#10092)
beac202
@xiezhq-hermann
[HiStorage] Remove delete and clear as necessary methods (sgl-project…
0b8c572 
…#10039)
@hzh0425
Modify ci workflow for auto-partitioning in 2-GPU backend tests (sgl-…
1a3d6f3 
…project#10029)
@zhyncs
Revert "[1/N][Bug] Fix w4afp8 MoE NaN issue (sgl-kernel) (sgl-project…
0e78c63 
…#9953)" (sgl-project#10097)
@sogalin @hubertlu-tw
Fix RMSNorm API CALL mismatch issue. (sgl-project#10032)
8d114f2 
Co-authored-by: Hubert Lu <Hubert.Lu@amd.com>
@shadowpa0327
fix double sparsity initialization (sgl-project#6905)
ad26f29
@DevashishLal-CB @BBuf
[Fix] illegal sync based on undefined behaviour (sgl-project#9620)
dbb1235 
Signed-off-by: Devashish Lal <devashish@rivosinc.com>
Co-authored-by: Xiaoyu Zhang <35585791+BBuf@users.noreply.github.com>
@ch-wan
[7/N] MoE Refactor: the implementation of new framework (sgl-project#…
3fa62da 
…9269)
@kaixih
[NVIDIA] Remove unused get_fused_moe_impl_class function (sgl-proje…
9a719b7 
…ct#9764)
@kaixih
[NVIDIA] disable chunked prefix cache when dp and blackwell is used (s…
90dfe3d 
…gl-project#9861)
@jinyangyuan-nvidia
perf: Avoid unnecessary data type conversions for DeepSeek-V3 on Blac…
012584e 
…kwell (sgl-project#9834)

Signed-off-by: Jinyang Yuan <154768711+jinyangyuan-nvidia@users.noreply.github.com>
@ch-wan @gemini-code-assist
[Fix] Compatibility between DP attention and pipeline parallelism (sg…
21af5c0 
…l-project#10100)

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
@ch-wan
Fix circular import (sgl-project#10107)
a5a0320
@hlu1
Disable kernel cutlass_mla_decode on SM103 (sgl-project#10058)
4c22ebe 
Signed-off-by: Hao Lu <14827759+hlu1@users.noreply.github.com>
@hlu1
Remove non-accelerated targets(100 and up) from cmake (sgl-project#10041
039cef7 
)

Signed-off-by: Hao Lu <14827759+hlu1@users.noreply.github.com>
@hlu1
[chore] Remove unused ep_moe cuda kernels (sgl-project#9956)
5f1eb20
@ShangmingCai
[CI] Refactor disaggregation tests (sgl-project#10068)
00974e4 
Signed-off-by: Shangming Cai <csmthu@gmail.com>
@key4ng
increase the rust e2e timeout (sgl-project#10116)
b3e7a2c
@key4ng
[router] Improve the router e2e tests (sgl-project#10102)
9eb50ec
@merrymercy @github-actions @hanming-lu
[Auto Sync] Update server_args.py (20250906) (sgl-project#10117)
f3b6760 
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Hanming Lu <69857889+hanming-lu@users.noreply.github.com>
@yuan-luo @BBuf
Optimize moe_sum_reduce_kernel (sgl-project#9477)
cb3918a 
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
Co-authored-by: Xiaoyu Zhang <35585791+BBuf@users.noreply.github.com>
@Oasis-Git @xiezhq-hermann
[Feature] LMCache Connector Integration (sgl-project#9741)
9a7ced4 
Signed-off-by: Oasis-Git <ayw.sirius19@gmail.com>
Signed-off-by: YuhanLiu11 <yliu738@wisc.edu>
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
@jianyingzhu
CUTLASS fp8 blockwise gemm support of sm120 (sgl-project#9969)
dd1e268
@HydraQYH
Optimize nvfp4 block scaled gemm kernel when M is small. (sgl-project…
85ed8e0 
…#10101)
@benbarsdell
Fix cuda graph mode in flashinfer attn backend (sgl-project#10056)
a12061d
@stmatengss @hzh0425
[HiCache] fix: check clear() method for storage backend (sgl-project#…
41628dc 
…10096)

Co-authored-by: hzh0425 <hzh0425@apache.org>
@miter6
add dataset_path for bench_one_batch_server.py (sgl-project#10113)
111b137 
Signed-off-by: linhuang <linhuang@ruijie.com.cn>
Co-authored-by: linhuang <linhuang@ruijie.com.cn>
@merrymercy @github-actions @jzhou-xai
[Auto Sync] Update parallel_state.py (20250907) (sgl-project#10126)
617aa2b 
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: jzhou-xai <jzhou@x.ai>
@DarkSharpness
[Minor] fix lint in main (sgl-project#10128)
0672468
@hnyls2002
[1/2] Refactor multi-tokenizer manager (sgl-project#10074)
e719bb0
@gmlwns2000 gmlwns2000 changed the base branch from main to deepauto/dev September 12, 2025 02:18
@gmlwns2000 gmlwns2000 closed this Oct 1, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.