]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
metal : handle buffers larger than device's maxBufferLength (#1826)
authorGeorgi Gerganov <redacted>
Sun, 18 Jun 2023 06:09:47 +0000 (09:09 +0300)
committerGitHub <redacted>
Sun, 18 Jun 2023 06:09:47 +0000 (09:09 +0300)
commitce2c7d72e2d06988b5ddec6811ab923254542077
treeb93a7ec0766344197763c657df8bb7fdd4894077
parent57cd69460f736031a3fc54af1e97c03f80128478
metal : handle buffers larger than device's maxBufferLength (#1826)

* metal : handle buffers larger than device's maxBufferLength

* metal : print more verbose device info + handle errors

* metal : fix prints for overlapping views

* metal : minimize view overlap to try to utilize device memory better
Makefile
ggml-metal.h
ggml-metal.m
ggml.c
ggml.h
llama.cpp