Duplicate: balanced-ternary K5 beats FP8 by 3.2× on L4
Code_MasterCody · reddit · 2026-10-30
Same author duplicate of the previous post (this one includes full repo and video links): TKintegers balanced-ternary K5 achieves 3.21× prefill and 3.10× decode throughput vs TorchAO FP8 on NVIDIA L4 with Qwen2.5-0.5B-Instruct.
Repo: https://github.com/CodeMasterCody3D/tkintegers
Related event: Balanced Ternary K5 Weight Scheme Beats FP8 Inference Throughput 3x(3 posts)→
More from Infra
- OpenRouter's new router: ~20% faster TTFT, ~60% fewer rate-limit errors for open-weight models — AccBalanced · 2026-10-30
- Amazon discloses over $77B in Q3 capex on AI infrastructure, far above Wall Street expectations — Polymarket · 2026-10-30
- Google Cloud revenue doubles to $30.7B but Alphabet stock drops 11% on search fears — fortune · 2026-10-30
- AWS revenue jumps 39.3% to $46B, fastest growth in five years; Amazon projects 40%+ through 2028 — Beth_Kindig · 2026-10-30
- Compute is ASI-era land, but only cap tables decide who gets a claim — krishnan · 2026-10-30
- PC downturn frees wafers for server CPUs, not accelerators, argues Bajarin — BenBajarin · 2026-10-30