-
Notifications
You must be signed in to change notification settings - Fork 306
Pull requests: SemiAnalysisAI/InferenceX
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
feat(agentx): retune GLM5.2 FP4 MI355X ATOM recipe on _0921 docker
agentx
AgentX benchmarks, recipes, and infrastructure
AMD
full-sweep-enabled
#3359
opened Sep 22, 2026 by
zhuyuhua-v
Collaborator
Loading…
[PowerX] drive GB200 AgentX power from resolved recipes / 根据实际配方接入 GB200 AgentX 功耗
#3358
opened Sep 22, 2026 by
edwingao28
Collaborator
Loading…
7 of 10 tasks
[PowerX] retain healthy window measurements in audit sidecars / 在审计文件中保留健康窗口测量
#3357
opened Sep 22, 2026 by
edwingao28
Collaborator
Loading…
5 of 10 tasks
Add DSV4 B200 Dynamo+SGLang AgentX configs with default optimizations / 新增采用默认优化的 DSV4 B200 Dynamo+SGLang AgentX 配置
full-sweep-enabled
#3356
opened Sep 22, 2026 by
nvpohanh
Collaborator
Loading…
[PowerX] report power sample gaps instead of voiding the window / 报告功耗采样间隙而非作废窗口
#3353
opened Sep 21, 2026 by
edwingao28
Collaborator
Loading…
5 of 10 tasks
[GB200][SGLang][AgentX] Improve DeepSeek V4.1 Flash performance
#3347
opened Sep 21, 2026 by
cquil11
Collaborator
Loading…
[B200][SGLang][AgentX] Improve DeepSeek V4.1 Flash performance
all-evals
Expand eval selection to every fixed-sequence config
full-sweep-enabled
#3346
opened Sep 21, 2026 by
cquil11
Collaborator
Loading…
[H100][SGLang][AgentX] Improve DeepSeek V4.1 Flash performance
all-evals
Expand eval selection to every fixed-sequence config
full-sweep-enabled
#3345
opened Sep 21, 2026 by
cquil11
Collaborator
Loading…
[GB300][SGLang][AgentX] Improve DeepSeek V4.1 Flash performance
all-evals
Expand eval selection to every fixed-sequence config
full-sweep-enabled
#3344
opened Sep 21, 2026 by
cquil11
Collaborator
Loading…
[MI355X][SGLang][AgentX] Improve DeepSeek V4.1 Flash performance
all-evals
Expand eval selection to every fixed-sequence config
engine-patch
Modifies inference engine or serving-stack code; apply patchwork CI priority
#3343
opened Sep 21, 2026 by
cquil11
Collaborator
Loading…
[B300][SGLang][AgentX] Improve DeepSeek V4.1 Flash performance
#3342
opened Sep 21, 2026 by
cquil11
Collaborator
Loading…
[H200][SGLang][AgentX] Improve DeepSeek V4.1 Flash performance
all-evals
Expand eval selection to every fixed-sequence config
full-sweep-enabled
#3341
opened Sep 21, 2026 by
cquil11
Collaborator
Loading…
[Klaud Cold] Add MI300X TP4 DeepSeek-V4.1-Flash vLLM AgentX arm with Engram host offload / 新增 Engram 主机卸载的 MI300X TP4 DeepSeek-V4.1-Flash vLLM AgentX 臂
full-sweep-enabled
#3336
opened Sep 21, 2026 by
functionstackx
Collaborator
Loading…
6 tasks
[Klaud Cold] Update dsv4-fp4-b200-sglang-agentic-hicache-mtp SGLang image to v0.5.20-cu130 / 将 dsv4-fp4-b200-sglang-agentic-hicache-mtp 的 SGLang 镜像更新至 v0.5.20-cu130
full-sweep-fail-fast
#3334
opened Sep 21, 2026 by
Klaud-Cold
Collaborator
Loading…
[TileRT] Add GLM-5.3 FP8 MI355X prefill/decode-disaggregated recipes (8k1k + AgentX) / 新增 GLM-5.3 FP8 MI355X 的 PD 分离配方（8k1k + AgentX）
engine-patch
Modifies inference engine or serving-stack code; apply patchwork CI priority
full-sweep-fail-fast
#3330
opened Sep 21, 2026 by
Oseltamivir
Collaborator
Loading…
7 of 11 tasks
[AMD] Update GLM-5.2 MI355X image, HiCache capacity, and sweep / 更新 GLM-5.2 MI355X 镜像、HiCache 容量与 sweep
agentx
AgentX benchmarks, recipes, and infrastructure
AMD
full-sweep-enabled
#3329
opened Sep 21, 2026 by
jiejingzhangamd
Collaborator
Loading…
12 of 14 tasks
config(dsv41flash): sweep MI355X at TP=2 and TP=4 through concurrency 128 / MI355X DSv4.1-Flash 在 TP=2 与 TP=4 下扫描至并发 128
full-sweep-enabled
#3326
opened Sep 21, 2026 by
Fangzhou-Ai
Collaborator
Loading…
5 of 7 tasks
Test H100 ModelScope loading from empty caches / 测试 H100 ModelScope 冷缓存加载
full-sweep-enabled
#3324
opened Sep 20, 2026 by
functionstackx
Collaborator
Loading…
Add H100 TensorRT-LLM ModelScope coverage / 添加 H100 TensorRT-LLM ModelScope 覆盖
full-sweep-fail-fast
#3323
opened Sep 20, 2026 by
functionstackx
Collaborator
Loading…
[PowerX] skip unreadable in-window power cells instead of voiding the point / 跳过窗口内读不出的功耗单元格而非作废整个点
full-sweep-enabled
priority
Preempt other runs on this sweep's runners; restore them at the end (org members only)
skip_queue
#3316
opened Sep 20, 2026 by
edwingao28
Collaborator
Loading…
6 of 10 tasks
[AgentX GB300] refresh the full Kimi-K3 curve so every point carries measured power / 重测完整曲线使每个点都带实测功耗
full-sweep-enabled
priority
Preempt other runs on this sweep's runners; restore them at the end (org members only)
skip_queue
#3315
opened Sep 20, 2026 by
edwingao28
Collaborator
Loading…
1 of 5 tasks
[AgentX H200] refresh the full Kimi-K3 curve so every point carries measured power / 重测完整曲线使每个点都带实测功耗
full-sweep-enabled
priority
Preempt other runs on this sweep's runners; restore them at the end (org members only)
skip_queue
#3309
opened Sep 20, 2026 by
edwingao28
Collaborator
•
Draft
1 of 5 tasks
[Klaud Cold] MI325X Qwen3.8-27B BF16 MTP, synthetic AL 2.51 / MI325X Qwen3.8-27B BF16 MTP，合成 AL 2.51
full-sweep-fail-fast
qwen3.8-27b
Qwen3.8-27B (bf16) vLLM TP1 DSpark 1k1k recipes and their eager variants
#3308
opened Sep 20, 2026 by
Oseltamivir
Collaborator
•
Draft
Previous Next
ProTip!
Filter pull requests by the default branch with base:main.