Get the latest tech news

FP8 is ~100 tflops faster when the kernel name has "cutlass" in it

None

Get the Android app

Or read this on Hacker News

Read more on:

Photo of fp8

fp8

Photo of ~100

~100

Photo of cutlass

cutlass

Related news:

LLVM/Clang 20.1 Released With AMX-AVX512, AMX-FP8, AVX10.2, AMD GFX950 & Much More

LLVM 20 Feature Development Wraps Up With AMX-AVX512, AMX-FP8, AVX10.2 & AMD GFX950

« Pick up the Amazon Smart Plug on sale for $13 before Prime Day ends

Amazon Switch 2 restock: what to do if you get an invite to buy »