PANews|11月 18, 2026 23:04
[OpenAI: Its Model Improves AMD Flagship GPU Design Metrics by Up to 14%] According to OpenAI's official website, its model, developed in collaboration with AMD, achieved up to 14% improvements in power consumption, performance, and area (PPA) on the latter's flagship GPU IP relative to a fully optimized benchmark. Further improvements were found by rewriting the circuit description code, representing approximately a year's worth of progress in less than five months. OpenAI stated that AMD has incorporated some of the changes into designs recently submitted for manufacturing, and the remaining changes have been written into production design documents for future versions. AMD also verified that designs generated by this model delivered single-digit percentage improvements in key modules of its next-generation flagship CPU. The same methodology was also used in OpenAI's first self-developed inference chip, Jalapeño, and the improvements have been validated on actual chips.
Share To
Timeline
HotFlash
APP
X
Telegram
CopyLink