Commits · 101f1481f9c4e3e108d30ce2f8715ee89288992b · OpenDAS / vllm_cscc · GitLab

01 Apr, 2025 1 commit
- Add option to use DeepGemm contiguous grouped gemm kernel for fused MoE operations. (#13932) · e59ca942
  bnellnm authored Apr 01, 2025
```
Signed-off-by: Bill Nell <bnell@redhat.com>
```
  e59ca942
29 Mar, 2025 2 commits
- [FEAT] [ROCm] Add AITER int8 scaled gemm kernel (#15433) · 4965ec42
  TJian authored Mar 29, 2025
```
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
```
  4965ec42
- [TPU][V1][Bugfix] Fix w8a8 recompiilation with GSM8K (#15714) · da461f3c
  Nicolò Lucchesi authored Mar 29, 2025
```
Signed-off-by: NickLucche <nlucches@redhat.com>
```
  da461f3c
28 Mar, 2025 2 commits
- [Quantization][FP8] Adding support for fp8 gemm layer input in fp8 (#14578) · 4d0ec372
  Gregory Shtrasberg authored Mar 27, 2025
```
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
```
  4d0ec372
- [Quantization][V1] BitsAndBytes support V1 (#15611) · 726efc6a
  Jee Jee Li authored Mar 28, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  726efc6a
27 Mar, 2025 3 commits

[Quantization] Fp8 Channelwise Dynamic Per Token GroupedGEMM (#15587) · 43ed4143

Robert Shaw authored Mar 27, 2025


Signed-off-by: ElizaWszola <eliza@neuralmagic.com>
Signed-off-by: ElizaWszola <ewszola@redhat.com>
Signed-off-by: rshaw@neuralmagic.com <robertgshaw2@gmail.com>
Co-authored-by: ElizaWszola <eliza@neuralmagic.com>
Co-authored-by: Lucas Wilkinson <wilkinson.lucas@gmail.com>
Co-authored-by: ElizaWszola <ewszola@redhat.com>

43ed4143

[TPU] Avoid Triton Import (#15589) · e1e0fd75
Robert Shaw authored Mar 27, 2025
```
Signed-off-by: rshaw@neuralmagic.com <robertgshaw2@gmail.com>
```
e1e0fd75

[Kernel] CUTLASS grouped gemm fp8 MoE kernel (#13972) · 9239bf71

ElizaWszola authored Mar 27, 2025


Signed-off-by: ElizaWszola <eliza@neuralmagic.com>
Signed-off-by: ElizaWszola <ewszola@redhat.com>
Co-authored-by: Lucas Wilkinson <wilkinson.lucas@gmail.com>

9239bf71

26 Mar, 2025 1 commit

[FEAT][ROCm] Integrate Fused MoE Kernels from AITER (#14967) · 5ebf6674

vllmellm authored Mar 26, 2025


Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
Co-authored-by: tjtanaa <tunjian.tan@embeddedllm.com>

5ebf6674

25 Mar, 2025 1 commit
- Revert "Fix non-contiguous input passed to Marlin kernel (#15319)" (#15398) · b5269db9
  Tyler Michael Smith authored Mar 24, 2025
  
  b5269db9
24 Mar, 2025 3 commits
- [ROCm][Kernel] MoE weights padding (#14454) · f533b583
  Gregory Shtrasberg authored Mar 24, 2025
```
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
Signed-off-by: charlifu <charlifu@amd.com>
Co-authored-by: charlifu <charlifu@amd.com>
```
  f533b583
- [Misc] Upgrade BNB version (#15183) · 3892e58a
  Jee Jee Li authored Mar 24, 2025
  
  3892e58a
- Fix non-contiguous input passed to Marlin kernel (#15319) · d20e2611
  Qubitium-ModelCloud authored Mar 24, 2025
  
  d20e2611
23 Mar, 2025 1 commit
- [Misc] Add tuned R1 w8a8 and MoE configs for NVIDIA L20 (#15322) · f90d34b4
  DefTruth authored Mar 23, 2025
```
Signed-off-by: DefTruth <qiustudent_r@163.com>
```
  f90d34b4
20 Mar, 2025 1 commit
- [Attention] Flash Attention 3 - fp8 (#14570) · a597a575
  Mickaël Seznec authored Mar 20, 2025
```
Signed-off-by: Mickael Seznec <mickael@mistral.ai>
```
  a597a575
18 Mar, 2025 2 commits
- [Bugfix] Fix broken CPU quantization due to triton import (#15038) · 179a619c
  Isotr0py authored Mar 18, 2025
```
Signed-off-by: Isotr0py <2037008807@qq.com>
```
  179a619c
- MI325 configs, fused_moe_kernel bugfix (#14987) · 8b793f7e
  ekuznetsov139 authored Mar 18, 2025
```
Signed-off-by: Eugene Kuznetsov <eugene.kuznetsov@amd.com>
```
  8b793f7e
16 Mar, 2025 1 commit
- [Kernel] Add more tuned configs (#14877) · 71c1e071
  Simon Mo authored Mar 15, 2025
```
Signed-off-by: simon-mo <simon.mo@hey.com>
```
  71c1e071
14 Mar, 2025 1 commit
- [Attention] MLA get rid of materialization (#14770) · 9532c498
  Lucas Wilkinson authored Mar 14, 2025
```
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
  9532c498
13 Mar, 2025 2 commits
- [Bugfix] Fix chunked prefill for GGUF (#14666) · 55211b01
  Szymon Ożóg authored Mar 13, 2025
```
Signed-off-by: SzymonOzog <szymon.ozog@aleph-alpha.com>
```
  55211b01
- [ROCm][FP8] Fix for adjustments needed only for fnuz (#14689) · a94a699c
  Gregory Shtrasberg authored Mar 12, 2025
```
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
```
  a94a699c
12 Mar, 2025 2 commits
- [Kernel] Add ModelOpt FP4 Checkpoint Support (#12520) · debd6bbf
  Pavani Majety authored Mar 11, 2025
```
Signed-off-by: Pavani Majety <pmajety@nvidia.com>
```
  debd6bbf
- [Kernel] GGUF MoE kernel (#14613) · e22ee1e7
  Szymon Ożóg authored Mar 12, 2025
```
Signed-off-by: SzymonOzog <szymon.ozog@aleph-alpha.com>
```
  e22ee1e7
11 Mar, 2025 1 commit
- dynamic distpatch of fp8 kernels (#14245) · a1c8f379
  Jeff Daily authored Mar 11, 2025
```
Signed-off-by: Jeff Daily <jeff.daily@amd.com>
```
  a1c8f379
10 Mar, 2025 1 commit

[Kernel] Add more dtype support for GGUF kernels (#14043) · 89cdaa83

Szymon Ożóg authored Mar 10, 2025


Signed-off-by: SzymonOzog <szymon.ozog@aleph-alpha.com>
Signed-off-by: SzymonOzog <szymon.ozog@gmail.com>

89cdaa83

07 Mar, 2025 2 commits
- [Kernel] optimize performance of gptq marlin kernel when n is small (#14138) · d0feea31
  Jinzhen Lin authored Mar 08, 2025
```
Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>
```
  d0feea31
- [FP8] Refactor apply_fp8_linear and apply_fp8_linear_generic into an object (#14390) · e1744502
  Luka Govedič authored Mar 07, 2025
```
Signed-off-by: luka <luka@neuralmagic.com>
```
  e1744502
05 Mar, 2025 1 commit

[Bugfix] Fix gptq_marlin for deepseek-v3 (#13750) · 0df25101

rainkert authored Mar 05, 2025


Signed-off-by: dangshunya <dangshunya@baichuan-inc.com>
Co-authored-by: dangshunya <dangshunya@baichuan-inc.com>

0df25101

01 Mar, 2025 1 commit
- [Misc][Kernel]: Add GPTQAllSpark Quantization (#12931) · 6a92ff93
  YajieWang authored Mar 01, 2025
  
  6a92ff93
28 Feb, 2025 2 commits
- [torch.compile] Fix RMSNorm + quant fusion in the non-cutlass-fp8 case, rename... · bd56c983
  Luka Govedič authored Feb 28, 2025
```
[torch.compile] Fix RMSNorm + quant fusion in the non-cutlass-fp8 case, rename RedundantReshapesPass to NoopEliminationPass (#10902)
Signed-off-by: luka <luka@neuralmagic.com>
```
  bd56c983
- [Bugfix] Fix MoeWNA16Method activation (#14024) · e0734387
  Jee Jee Li authored Feb 28, 2025
  
  e0734387
27 Feb, 2025 1 commit
- [Model] Deepseek GGUF support (#13167) · 7f0be2aa
  Szymon Ożóg authored Feb 27, 2025
  
  7f0be2aa
26 Feb, 2025 2 commits
- [BugFix] Make FP8 Linear compatible with torch.compile (#13918) · b382a7f2
  Woosuk Kwon authored Feb 26, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  b382a7f2
- [Model] Support Grok1 (#13795) · 07c43530
  Michael Goin authored Feb 25, 2025
```
Signed-off-by: mgoin <mgoin64@gmail.com>
```
  07c43530
25 Feb, 2025 3 commits
- Fix CompressedTensorsWNA16MoE with grouped scales (#13769) · 4d251ad0
  Michael Goin authored Feb 25, 2025
  
  4d251ad0
- [Bugfix][Quantization] Fix FP8 + EP (#13784) · 1e15aaef
  Tyler Michael Smith authored Feb 24, 2025
```
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
```
  1e15aaef
- [Misc] set single whitespace between log sentences (#13771) · 51010a18
  cjackal authored Feb 25, 2025
```
Signed-off-by: cjackal <44624812+cjackal@users.noreply.github.com>
```
  51010a18
24 Feb, 2025 1 commit
- Expert Parallelism (EP) Support for DeepSeek V2 (#12583) · 781096e3
  Jongseok Park authored Feb 24, 2025
  
  781096e3
22 Feb, 2025 1 commit
- [ROCm] Apply FP8 weights padding to values not divisible by 512 bytes on ROCm (#13231) · c904fddd
  Gregory Shtrasberg authored Feb 22, 2025
  
  c904fddd
21 Feb, 2025 1 commit

[Attention] MLA with chunked prefill (#12639) · 288cc6c2

Lucas Wilkinson authored Feb 21, 2025


Signed-off-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Patrick Horn <patrick.horn@gmail.com>
Co-authored-by: simon-mo <xmo@berkeley.edu>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>

288cc6c2