-
Notifications
You must be signed in to change notification settings - Fork 282
Pull requests: SemiAnalysisAI/InferenceX
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
ci: use Fable 5.1 for all Claude Code invocations
#2805
opened Sep 2, 2026 by
cquil11
Collaborator
Loading…
perf(agentx): refresh Kimi-K3 MI355X LMCache curve / 刷新 Kimi-K3 MI355X LMCache 曲线
agentx
AgentX benchmarks, recipes, and infrastructure
AMD
full-sweep-enabled
#2804
opened Sep 2, 2026 by
hyukjlee
Collaborator
Loading…
4 of 9 tasks
perf(agentx): add LMCache offload arm to Kimi-K3 B300 at conc 56/70
full-sweep-fail-fast
#2802
opened Sep 1, 2026 by
sammshen
Collaborator
Loading…
2 tasks
[AMD][MI35X] Enable fused allreduce+RMSNorm for Qwen3.5 MXFP4 MI355X agentic-mtp
full-sweep-enabled
#2801
opened Sep 1, 2026 by
ChangLiu0709
Collaborator
Loading…
[AMD][DSV4] Add a DP-attention arm and refresh the MI355X SGLang AgentX key
full-sweep-enabled
#2800
opened Sep 1, 2026 by
karverma-amd
Collaborator
Loading…
5 of 6 tasks
Validate vLLM P/D cache-source metrics on GB300
full-sweep-enabled
#2797
opened Sep 1, 2026 by
cquil11
Collaborator
Loading…
Validate bounded vLLM cache token sources on H100
full-sweep-enabled
#2796
opened Sep 1, 2026 by
cquil11
Collaborator
Loading…
[AMD] [AGENTX] Kimi Perf Tuning
agentx
AgentX benchmarks, recipes, and infrastructure
AMD
#2795
opened Sep 1, 2026 by
ajith-sirra-amd
Collaborator
Loading…
[AMD][MI35X] Serve Qwen3.5 MXFP4-AttnFP8-V2 on the MI355X SGLang
full-sweep-fail-fast
#2793
opened Sep 1, 2026 by
yichiche
Collaborator
Loading…
[Klaud Cold] [AMD] Bump DSV4 MI355X vLLM 8k/1k to the 2026-09-01 nightly / [Klaud Cold] [AMD] 将 DSV4 MI355X vLLM 8k/1k 更新至 2026-09-01 nightly
AMD
full-sweep-enabled
#2792
opened Sep 1, 2026 by
jiacao-amd
Collaborator
Loading…
feat(glm52-agentx): add compact B300 Dynamo+TRT-LLM recipes / 新增紧凑布局的 B300 Dynamo+TRT-LLM 配方
full-sweep-enabled
#2789
opened Aug 31, 2026 by
Ankur-singh
Collaborator
Loading…
5 of 9 tasks
CollectiveX: wire-basis bandwidth, nccl-ep LL hold, routed HT window, AsyncLL, DSv4-Pro workload
#2786
opened Aug 30, 2026 by
Oseltamivir
Collaborator
Loading…
[AMD][Power] fix: wait for AMD telemetry to cover the benchmark window end before stopping the monitor / 修复 AMD 遥测在基准窗口结束前被截断导致功耗校验失败的问题
#2767
opened Aug 28, 2026 by
edwingao28
Collaborator
Loading…
Align CI business priority across node counts and model families
#2765
opened Aug 27, 2026 by
cquil11
Collaborator
Loading…
[Klaud Cold] qwen3.8next-fp8-mi355x-sglang-agentic: Qwen3.8-Flash-Next FP8 SGLang AgentX on MI355X / MI355X 上 Qwen3.8-Flash-Next FP8 SGLang AgentX 配方
full-sweep-fail-fast
#2754
opened Aug 27, 2026 by
functionstackx
Collaborator
Loading…
[Klaud Cold] qwen3.8next-fp4-b200-sglang-agentic-mtp: day-zero Qwen3.8-Flash-Next NVFP4 SGLang AgentX on B200 / B200 上 Qwen3.8-Flash-Next NVFP4 SGLang AgentX 首发配方
full-sweep-fail-fast
#2751
opened Aug 27, 2026 by
functionstackx
Collaborator
Loading…
[CI] Import the DCGM exporter via enroot registry syntax / 用 enroot registry 语法导入 DCGM exporter
#2735
opened Aug 26, 2026 by
edwingao28
Collaborator
Loading…
[Klaud Cold] minimaxm3-fp4-mi355x-atom-agentic-mtp: day-zero MiniMax-M3 MXFP4 ATOM AgentX recipe on MI355X / MI355X 上 MiniMax-M3 MXFP4 ATOM AgentX 首发配方
full-sweep-fail-fast
#2733
opened Aug 26, 2026 by
functionstackx
Collaborator
Loading…
test: validate upstream-native vLLM Router topologies
#2731
opened Aug 25, 2026 by
cquil11
Collaborator
Loading…
[WIP][AMD][AgentX] Add Qwen3.8 FP8 MI355X two-node vLLM
agentx-fast
Run AgentX throughput with 1 warmup request per lane and a 20-minute profile; not reusable
sweep-enabled
#2724
opened Aug 25, 2026 by
haic0
Collaborator
Loading…
6 of 7 tasks
Previous Next
ProTip!
Follow long discussions with comments:>50.