MIT researchers developed Attention Matching, a KV cache compaction technique that compresses LLM memory by 50x in seconds — ...

Penguin Solutions today announced its MemoryAI KV cache server, the industry's first production-ready KV cache server utilizing CXL memory technology.

This approach can be viewed as a memory plug-in for large models, providing a fresh perspective and direction for solving the ...

Nvidia's KV Cache Transform Coding (KVTC) compresses LLM key-value cache by 20x without model changes, cutting GPU memory costs and time-to-first-token by up to 8x for multi-turn AI applications.

Marvell Technology, Inc. (NASDAQ: MRVL), a leader in data infrastructure semiconductor solutions, today announced Marvell® ...

As AI workloads extend across nearly every technology sector, systems must move more data, use memory more efficiently, and respond more predictably than traditional design methodologies allow. These ...

Lightbits Labs Ltd. today is introducing a new architecture aimed at addressing one of the most stubborn bottlenecks in large-scale artificial intelligence inference: the growing mismatch between the ...

OriginAI inference solutions are designed leveraging Penguin Solutions 3.3+ billion hours of GPU runtime experience and more ...

AMD's Threadripper Pro 5000 series hit the diy market last year, and today Luke takes a look at the 5995WX, 5975WX, and ...

The latest Area-51 desktop from Alienware centers around AMD’s Ryzen 7 9800X3D, an 8-core processor with 104MB of total cache designed for gaming workloads. Paired with an RTX 5080 graphics card, 64GB ...

M5 Pro and M5 Max both use the same 18-core CPU die, but Pro uses a 20-core GPU die, and Max gets a 40-core GPU die. (Because the memory controller is also part of the GPU die, the Max chip still ...

Marvell introduces the Structera S 30260, a next‑generation CXL switch designed to enable rack‑level memory pooling.