aiqbenchInformații despre AI & calcul cuantic

...

Română · beta. Traducerea a fost făcută automat și este încă în curs de revizuire. Citește pagina în engleză

perspectiva aiqbenchCipuri & infrastructură

Astera Labs is addressing the critical bottleneck of KV-cache management in agentic AI by introducing fabric-attached memory controllers that offload context data directly to GPU fabrics. This architectural shift aims to improve memory utilization and reduce latency for high-bandwidth inference workloads without requiring immediate increases in raw memory capacity. The development highlights a strategic pivot toward specialized memory hierarchies as standard DDR5 supply constraints intensify for large-scale AI training and inference clusters.

Surse care acoperăHPCwire, Latent Space, AWS, SiliconANGLE

Ce este aceastaLectura unui model asupra articolului și a acoperirii conexe, scrisă pentru aiqbench. Nu sunt cuvintele publicației.

Astera Labs Expands Leo Memory Controllers for Agentic AI and Cloud Workloads

HPCwire15 septembrie 20265 min de cititArticolul și perspectiva sunt în engleză.

SAN JOSE, Calif., Sept. 15, 2026 — Astera Labs, Inc., a leader in semiconductor-based connectivity solutions for rack-scale AI infrastructure, today announced three new memory connectivity solutions—Leo X-Series and the Leo 2 E & P-Series—extending its hyperscale-proven Leo Smart Memory Controller family across direct fabric-attached GPU memory and CPU-attached expansion and pooled/shared memory architectures. “Agentic […] The post Astera Labs Expands Leo Memory Controllers for Agentic AI and Cloud Workloads appeared first on HPCwire.

SAN JOSE, Calif., Sept. 15, 2026 — Astera Labs, Inc., a leader in semiconductor-based connectivity solutions for rack-scale AI infrastructure, today announced three new memory connectivity solutions—Leo X-Series and the Leo 2 E & P-Series—extending its hyperscale-proven Leo Smart Memory Controller family across direct fabric-attached GPU memory and CPU-attached expansion and pooled/shared memory architectures. “Agentic AI adoption is outpacing the industry’s ability to provision memory for it, and that gap is widening every quarter,” said Matt Kimball, vice president and principal analyst at Moor Insights & Strategy. “That…

Extras afișat conform termenilor RSS ai publicației · 88 cuvinte · textul integral și figurile sunt pe site-ul publicației

Citește pe HPCwire ↗Întreabă despre acest articol