Install
Packaging & Memory
Chiplets, 3D IC, HBM stacks, DDR5/PCIe lanes, and bandwidth tradeoffs.
- 25 Tracked terms
- Last 30 days Feed window
What this topic collects on
An article joins this feed when it matches these terms. Each one is also a search of its own.
Related topics
Latest in Packaging & Memory
Kioxia Targets Hyperscale Data Centres With New High-Density QLC Solid-State Drives
1+ hour, 28+ min ago (575+ words) Kioxia Corporation has taken the wraps off its first E1.L form factor solid-state drive built on eighth-generation BiCS FLASH QLC (quad-level cell) flash memory, pitching the new LD4 Series at hyperscale data centre operators grappling with surging storage demands. The drives…...
AI Infrastructure Market to Hit $621.14 Billion by 2035 as AI Inference Demand Surges
29+ min ago (1503+ words) But the next phase of AI infrastructure development could look considerably different. This transition is creating a new infrastructure investment cycle. Unlike model training, which can be scheduled around large computational jobs, inference workloads must often respond to unpredictable demand…...
Kioxia to Demo GPU-Focused SSD Exceeding 20 Million IOPS at OCP Global Summit
3+ hour, 27+ min ago (1662+ words) Enter to search English articles · Esc to close · ↑↓ to move Translated from the Japanese original Kioxia plans to show the upgraded GP Series at the OCP Global Summit, to be held October 12–15 in San Jose, California. According to the company's…...
[News] NVIDIA’s NVHBM Reshapes Base Die Strategies: How Micron, Samsung Are Bringing TSMC Into the Fold
4+ hour, 35+ min ago (364+ words) NVIDIA’s NVHBM, announced in August, could mark a turning point for custom HBM. By moving the memory controller into the HBM base die, the architecture gives base dies a bigger role—and is already reshaping how memory makers source them....
Why Custom Silicon Matters in AI Data Centers
12+ hour, 3+ min ago (560+ words) That is why custom silicon has moved from a specialized option to an increasingly important part of data center architecture. Instead of forcing every workload through the same general-purpose architecture, designers can optimize silicon around the functions that matter most....
Kioxia Announces E1.L QLC-Based SSD Designed for Hyperscale Data Centers
15+ hour, 54+ min ago (168+ words) TechPowerUp Kioxia Announces E1.L QLC-Based SSD Designed for Hyperscale Data Centers - PCIe 5.0 (Gen 4 x4) 1, NVMe 2.0e specification compliant - Support for OCP Datacenter NVMe SSD Specification 2.6 - Single-port design for hyperscale and data center-class servers - Available in 15.36 TB2 and 30.72 TB capacity points today, with a…...
KIOXIA showcases next-generation AI Data Storage at 2026 OCP Global Summit
16+ hour, 19+ min ago (351+ words) 2026 OCP Global Summit will be held in San Jose next week and KIOXIA will be on hand to showcase its high-performance flash and storage solutions. At the 2026 OCP Global Summit in San Jose, KIOXIA will demonstrate flash storage for AI…...
The Hardware Plan Was Finally Done. Mostly.
15+ hour, 59+ min ago (1689+ words) Part 6 of my journey into building a local AI server: after weeks of changing GPUs, power supplies, cases and even the architecture of the machine itself, the parts list finally started to look permanent. At the end of the previous…...
10M Batch LLM Inference at $0 Cloud Cost: O(1) Memory Clamped Architecture
17+ hour, 6+ min ago (517+ words) High cloud API costs and Out-Of-Memory (OOM) failures in large-scale data pipelines are architectural defects, not hardware constraints. This report documents the performance of a 10,000,000-record batch LLM inference workload executed locally on an HP Z4 G4 workstation (Intel Xeon W-2155, 128GB ECC DDR4 RAM,…...
Kioxia LD4 E1.L QLC NVMe SSD Announced
23+ hour, 50+ min ago (230+ words) The Kioxia LD4 is a higher-capacity E1.L NVMe SSD designed to cram a lot of sorage while taking a minimal amount of faceplate area The post Kioxia LD4 E1.L QLC NVMe SSD Announced appeared first on ServeTheHome. Kioxia has introduced its first…...