4 ms·
Its graph execution is still full of busyloops, e.g.: https://github.com/ggerganov/llama.cpp/blob/44f906e8537fcec965e312d621c80556d6aa9bec/ggml.c#L14575 https:
by yukIttEft 3y ago
Its graph execution is still full of busyloops, e.g.:
https://github.com/ggerganov/llama.cpp/blob/44f906e8537fcec965e312d621c80556d6aa9bec/ggml.c#L14575 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 3y agoSomeone ought to be along with a PR eventually.
- boywitharupee 3y agois graph execution used for training only or inference also?
- LoganDark 3y agoInference. It's a big bottleneck for RWKV.cpp, second only to the matrix multiplies.
- make3 3y agodoes tbb work with apple Silicon?
- yukIttEft 3y agoI guess https://formulae.brew.sh/formula/tbb https://formulae.brew.sh/formula/tbb
- mhh__ 3y agoIt's not a very good library IMO.
- dindresto 3y agoggml or Intel TBB?