GPU compute is deterministic; what varies is which kernel path gets taken
felipesuch · x · 2026-10-30
Pushing back on the common claim that GPUs are non-deterministic, the developer argues that given deterministic inputs, results are deterministic. What looks non-deterministic to users is usually which kernel/path the code takes, which depends on variables outside your control like batch size and GPU SKU — i.e., determinism is not the same as batch/shape invariance.
Related event: GPU Compute Is Deterministic; the Real Culprit Is Kernel Path Selection(2 posts)→
More from Infra
- GPU break-even throughput vs open-weight token prices up 25-83% since May — bookwormengr · 2026-10-30
- Quantum startups raised $4.9B in VC but sold only $1.9B — a field guide maps all four markets — antavedissian · 2026-10-30
- DeepSeek V4.1 on a single RTX 5090: 4000+ tok/s prefill, 7-24 tok/s decode via NVMe streaming — Front_Eagle739 · 2026-10-30
- Cohere to invest over C$4 billion in Canada, creating 1,500+ jobs by 2031 — cohere · 2026-10-30
- River API Now Trains LLMs With Any Custom PyTorch Loss, Compiled and Run at Scale on Its GPUs — ibab · 2026-10-30
- Your AI agent would cost 260% more without prompt caching, Arena data shows — arena · 2026-10-30