Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Intrigued by the single-thread main memory bandwidth being a multiple of what you get from a single SKX. We also see this with Graviton 2. The latency is not terrible, either. How would this much available bandwidth change your choices when optimizing your algorithms?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: