CPUs are quietly becoming very well-balanced machines for BS 1 inference. The la... | Hacker News

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

pclmulqdq on June 1, 2025 | parent | context | favorite | on: Why DeepSeek is cheap at scale but expensive to ru...

CPUs are quietly becoming very well-balanced machines for BS 1 inference. The latest Intel Xeons should be at ~20 TPS.

Spooky23 on June 1, 2025 [–]

A base Mac Mini is ~20 :)

pclmulqdq on June 1, 2025 | [–]

Oh yeah, I did that math not assuming any quantization. I think if you can get a 3-4 bit quant working + int8 math, ~80 might be achievable.

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