TESLA LIVE 24.com · 24시간 반도체·SOXL 시세·속보 한국어 실시간

GPU용 32TB RAM: 마벨, 최대 16개 GPU 연결 및 CAPEX 절감 광자 기반 메모리 모듈 공개… 엔비디아도 방관하지 않아

TechRadar · 2026.10.24 06:35 · 원문 사이트
🇰🇷 한글 번역
GPU를 위한 32TB RAM: 마벨, 최대 16개의 GPU를 광으로 연결하고 자본 지출(CAPEX)을 대폭 절감하는 혁신적인 광자성 패브릭 메모리 모듈을 선보임 — 하지만 엔비디아도 가만히 있지 않을 것 테크라더
📄 원문 (English)
32 TB RAM for your GPUs: Marvell shows revolutionary Photonic Fabric Memory Module that uses light to link up to 16 GPUs and slash CAPEX — but Nvidia won't stick around doing nothing    TechRadar
▶ 칩 라이브 24 홈으로 — chiplive24.com

다른 반도체 뉴스

NVIDIA RTX Remix 1.6 Launches With DLSS 5, 6X Multi Frame Generation, Minecraft Java Edition Support, And More · WccftechNeoclouds secure record funds as Nvidia’s Huang says more investment required · Yahoo FinanceWhat Is Super Micro Computer (SMCI) Telling The Market With Its New AI Rack? · Yahoo FinanceMicron Technology Stock Price Forecast: CEO Reaffirms Tight Supply for Next Two Years, How Will HBM4 Mass Production and CXMT Controversy Impact Stock Price Expectations? · TradingKeyWhat Could Micron Technology (MU) Strike Risk And Supply Moves Mean For Investors? · Yahoo FinanceAMD Zen 6 Shows Up To 24% Higher Per-Clock Performance Than Zen 5, Promising A Big Leap For Ryzen Upgraders · WccftechPrediction: Nvidia Will Outperform Broadcom Over the Next 5 Years · Yahoo FinanceHow Baker Hughes is levering the Dell AI Factory with NVIDIA to Accelerate Centrifugal Compressor Simulations with Physics-Informed AI · Oil & Gas Journal