NVIDIA GPU engineer explains CUDA fat binaries: PTX is a virtual ISA, no final recompilation
matiasgoldberg · x · 2026-11-01
- Matias Goldberg (NVIDIA GPU engineer) explains CUDA compatibility: fat binaries ship code for every targeted sm version, so there is no final recompilation step. The "virtual sm" concept only hints that a given sm100 binary originated from, say, sm90 PTX — PTX being a virtual ISA usable for retargeting.
- He notes ISA-level guarantees apply only to output, not to specific timings (not at atomic-op level or which core unit handles the work), drawing a parallel to Spectre on x86.
Related event: GPU Experts Debate Whether Driver Versions Break CUDA Atomic Determinism(9 posts)→
More from Infra
- Local MiniMax H3 video gen on RTX 5060 8GB: 27 tests charted, gen time scales steeply with resolution — rguerraf · 2026-11-01
- China's Halloween cosplay peak: someone dressed up as Intel's -28.58% stock crash chart — basedjensen · 2026-11-01
- Is llama.cpp the new Ollama? Reddit debates the fragmented local LLM engine scene — fintip · 2026-11-01
- Researcher teases talk synthesizing a decade of large-scale training systems — _arohan_ · 2026-11-01
- vLLM cold starts on RunPod cut to 14s: DMTCP snapshots where CRIU fails — kbesplv · 2026-11-01
- Meme: writing an LLVM backend for a fully undocumented GPU architecture to ditch closed-source nvcc — mike64_t · 2026-11-01