Search results

deepseek sglang h200 experimental

Status: Open.

#870 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Mar 5, 2026

Profile test

Status: Open.

#863 In SemiAnalysisAI/InferenceX;

· Oseltamivir opened on Mar 4, 2026

vllm 0.16 single node mi300 kimi k2.5 vllm tp8

AMD

p1

Status: Open.

Feature

#859 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Mar 3, 2026

·  

kimi k2.5

track & frontend visualize the # of perf related flags

dev velocity

p1

Status: Open.

Feature

#858 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Mar 3, 2026

vllm 0.16 single node mi325 kimi k2.5 vllm tp8

AMD

Status: Open.

Feature

#856 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Mar 3, 2026

·  

kimi k2.5

GLM5 H200 FP8 SGL single node

Status: Open.

#844 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Mar 2, 2026

h100-cr_0, h100-cr_1 runner clean up storage

dev velocity

p1

Status: Open.

Feature

#835 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Mar 1, 2026

[experimental] gptoss h100 nvidia-smi power

Status: Open.

#828 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Mar 1, 2026

mi355 fp4 minimax vllm single node

AMD

Status: Open.

#826 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Mar 1, 2026

b200 local NVMe caching to speed up server start time

NVIDIA

p1

Status: Open.

Bug

#822 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Feb 27, 2026

test claude profile compare

Status: Open.

#819 In SemiAnalysisAI/InferenceX;

· functionstackx opened on Feb 26, 2026

Add step to run sweep workflows that automatically compares perf to latest official data

dev velocity

Status: Open.

Feature

#818 In SemiAnalysisAI/InferenceX;

· cquil11 opened on Feb 26, 2026