Articles on agentic inference, AgentX results, chip performance, and ML infrastructure.
New to the terminology? Browse the AI inference glossary.
Where AMD’s vendor engine wins on performance per dollar, and what E2E Normalized Interactivity actually measures
At this operating point, free AMD silicon would still not close the gap
What four years of hardware and a 4-bit format buy on a long-context agentic workload
GatedDeltaNet, a 262k native context, and no AMD competition at all on the same engine
Day 0 Inference Performance, InferenceX, 100x performance improvement in 26 Days, Cost per Million Tokens, Huawei 950DT Inference Trace Analysis
Both SKUs run SGLang EAGLE MTP; the Blackwell generation lifts perf/$ by ~1.2x at the peak and the NVIDIA GLM-5-NVFP4 checkpoint on FlashInfer TRT-LLM sparse MLA stacks another ~2.4–3.0x on 8K/1K
The amd/deepseek_v4 side branch shipped TileLang attention indexer, Triton sparse MLA, fused RoPE/Hadamard, FlyDSL MoE, and FP4 weights across 31 performance optimizations PRs — lifting first-light 20 tok/s/GPU at 2.4 tok/s/user into 2,256 tok/s/GPU at 9.4 tok/s/user on 8K/1K, with both throughput and interactivity climbing together
14 weeks after GLM-5 launched, AMD landed both MTP and non-MTP SGLang FP8 recipes on MI355X — fused MLA + FP8 KV cache via TileLang flips the single-node FP8 cost curve in AMD favor across most of the performance Pareto
From v0.5.8 (Feb) → v0.5.10rc0 (Apr) → v0.5.12 (May), three AITER kernel landings on MI355X plus a TP=8 → TP=2/TP=4 retune push Qwen3.5 8k/1k peak from 1.3k to 6.4k tok/s/GPU and extend the curve out to 75 tok/s/user
Piecewise CUDA graphs for DeepSeek V3, a unified event loop, and JIT kernels push 8k/1k throughput from 508 to 907 tok/s/GPU on the same 16 GPU B200 pool
