What happened
Greg Brockman says OpenAI models achieved roughly an additional year of progress on an AMD chip in less than five months, with some results included in a design recently taped out. The quoted OpenAI post reports up to 14% PPA improvement on flagship GPU IP and says gains on Jalapeño, its first custom inference chip, were verified on real silicon.
Why it matters
OpenAI says its models reached roughly an additional year of progress on an AMD chip in less than five months, with some results included in a design recently taped out; the cited results also include gains verified on real silicon.
roughly an additional year of progress on @AMD's chip achieved in less than 5 months using our models, with some of these results included in a design that's recently been taped out — unprecedented results at the frontier of chip design.
View the quoted post on XWorking with @AMD, our models identified up to 14% PPA improvement on flagship GPU IP in less than five months—building on designs AMD’s engineers had already spent years optimizing.
The same approach helped improve Jalapeño, our first custom inference chip, with gains verified on real silicon.
New evaluations also show our models getting better at certain chip design tasks.
We’re sharing these results and how we’re working with @Synopsys to bring our model capabilities and their trusted EDA design and validation tools to more chip designers.
The story so far
OpenAI’s models have accelerated chip design timelines, achieving roughly an additional year of development progress on an AMD chip in under five months. Greg Brockman stated that results from these AI-assisted optimizations, which include up to 14% PPA (power, performance, area) improvements on flagship GPU intellectual property, have already been incorporated into a design that was recently taped out. OpenAI has also verified performance gains for its custom Jalapeño inference chip using real silicon. Additionally, the company reported that its models have successfully optimized circuits used in Jalapeño silicon and improved RISC-V core designs.
