AMD Acquires MEXT to Enhance Data Center Memory Efficiency
AMD has acquired MEXT, a startup that developed memory tiering technology to make NAND flash memory appear as DRAM to the operating system, addressing growing memory constraints in data centers and enabling operators to save on DRAM costs.
Key points
- AMD has acquired MEXT, a startup that developed memory tiering technology to address growing memory constraints in data centers.
- MEXT's technology moves infrequently accessed data from expensive DRAM to NAND storage, increasing usable memory available to applications.
- The acquisition aims to improve system efficiency, lower operating costs, and deploy large-scale workloads more quickly.
- MEXT's Predictive Memory Engine uses AI models to anticipate which data stored in flash will be needed next and proactively transfers it back into DRAM.
- The technology is transparent to applications, preserving performance levels and increasing memory utilization.
AMD's acquisition of MEXT is a strategic move to address the growing memory constraints in data centers. As AI models continue to expand and datasets grow larger, memory availability has become an increasingly important factor affecting overall system performance. MEXT's memory tiering technology enables operators to save money on DRAM by moving infrequently accessed data to NAND storage.
The Predictive Memory Engine, developed by MEXT, uses AI models to anticipate which data stored in flash will be needed next and proactively transfers it back into DRAM. This technology is transparent to applications, preserving performance levels and increasing memory utilization.
The acquisition is expected to improve system efficiency, lower operating costs, and deploy large-scale workloads more quickly. With MEXT's technology, data center operators can optimize their memory resources and improve overall system performance.
The acquisition is a significant move by AMD to enhance its offerings in the data center market. With MEXT's technology, AMD can provide its customers with more efficient and cost-effective solutions for their memory needs.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.