How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip
What happened
On 25 August, OpenAI fully unveiled Jalapeño, the company’s debut AI accelerator chip. Jalapeño delivers up to 13.4 petaflops of 4-bit compute and accesses 232 gigabytes of the most advanced memory available, linking to it at a blazing 15.4 terabytes per second. Benchmarks cited by OpenAI show that Jalapeño can reduce end-to-end latency (the time between prompt to last token) by up to 3.6x when compared to Nvidia’s…
Summary assembled by rule from the sources below
Why it's spreading
Timeline
- First appeared on IEEE Spectrum AIIEEE Spectrum AI
- Discussion started on Hacker NewsHacker News
Sources
News
- How OpenAI Used Its Own LLMs to Design Its Jalapeño ChipIEEE Spectrum AI