Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

COMPANY · ENTITY #3714

NVIDIA A100

Related event timeline, sources and context from the news index.

EVENT TIMELINE

3

RESEARCH · 1 SOURCE · arXiv cs.AI

Echo: error‑compensated 4‑bit all-reduce for tensor‑parallel LLM training and inference

The arXiv paper introduces Echo, a 4‑bit all‑reduce protocol for tensor-parallel LLM training and inference that moves 4‑bit data but sums in high precision and feeds per-call rounding errors forward so they cancel rather than accumulate. The authors prove the accumulated error is depth‑independent and report that Echo is 2.09× faster than 16‑bit NCCL all‑reduce on 1 GB messages on A100s, speeds time‑to‑first‑token and training step latency by up to ~1.16×/1.09× on idle interconnects (and larger gains under background traffic), keeps average accuracy within 0.4 points on Qwen3‑8B, and yields <0.9% higher final loss after 10B‑token pretraining — with no FP8/FP4 hardware required.

7.0

RESEARCH · 1 SOURCE · arXiv cs.AI

FlashSinkhorn 2: block-sparse GPU solver for large entropic optimal transport

FlashSinkhorn 2 (FS2) is a two-stage GPU solver for squared‑Euclidean entropic optimal transport on low-dimensional point clouds that combines a coarse centroid stage with a block‑sparse fine stage (Morton-ordered blocks, screening, and fused tensor-core execution). The authors report FS2 reaching target marginal residuals on benchmark problems and solving discrete EOT between two 1.34×10^8‑particle measures from a cosmological N‑body simulation on a single NVIDIA A100 to an all-particle marginal residual below 0.01 in under 2.5 hours; an open-source implementation is available at the linked GitHub repository.

7.0

RESEARCH · 1 SOURCE · arXiv cs.AI

Inference Engineering Pareto Atlas (arXiv:2609.17863v1) maps cost/quality/latency trade-offs

The paper releases a Pareto atlas that maps cost, quality, and latency trade-offs for LLM inference by measuring 54 configurations of Qwen2.5-7B-Instruct on vLLM 0.12 across L4, A100, and H100 GPUs and calibrating a simulator that reproduces anchored measurements with <1.5% drift. Key findings: 18 of 36 calibrated configurations reach the Pareto frontier; combinations of optimizations outperform many single methods; AWQ 4bit cuts per-token latency to 0.34× but loses ~5.9% GSM8K accuracy; FP8 weights retain ~99.4% accuracy at 0.61–0.65× latency; a naive FP8 KV cache preserves throughput but answers none of 200 GSM8K questions correctly; H100 is best for tight latency while A100 is best for throughput and cost (~$0.106 per million tokens).

7.0