Its graph execution is still full of busyloops, e.g.: https://github.com/ggergan... | Hacker News

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

		yukIttEft on June 6, 2023 \| parent \| context \| favorite \| on: GGML – AI at the Edge Its graph execution is still full of busyloops, e.g.: https://github.com/ggerganov/llama.cpp/blob/44f906e8537fcec9... I wonder how much more efficient it would be when Taskflow lib was used instead, or even inteltbb.

moffkalast on June 6, 2023 | [–]

Someone ought to be along with a PR eventually.

mhh__ on June 6, 2023 | | [–]

It's not a very good library IMO.

dindresto on June 7, 2023 | | [–]

ggml or Intel TBB?

boywitharupee on June 6, 2023 | | [–]

is graph execution used for training only or inference also?

LoganDark on June 6, 2023 | | [–]

Inference. It's a big bottleneck for RWKV.cpp, second only to the matrix multiplies.

make3 on June 6, 2023 | [–]

does tbb work with apple Silicon?

yukIttEft on June 6, 2023 | [–]

I guess https://formulae.brew.sh/formula/tbb

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