8 Commits

Author	SHA1	Message	Date
mqy	67bb367962	typos	2023-06-18 14:27:56 +08:00
mqy	06b00827a0	bulk refactoring task profile and related to run CL GPU offloading. ... * removed ggml_task_backend, infavour of ggml_task_profile.runner and newly added id and name. * extracted mul_mat blas codes into ggml_compute_forward_mul_mat_blas, thus align with CUDA/CL a bit more and make it easier to fix profile and run tune. * rewrote task profile and update/add some cuda/cl codes, finnaly made CL GPU offloading work. * misc minor fix/update to tune, the data format was changed.	2023-06-18 14:27:56 +08:00
mqy	6b83a3e16f	try make CL run w/o tunning, but -ngl stucks no output. had to add task runer and profile id, many changes, see the f codes	2023-06-18 14:27:56 +08:00
mqy	5342dc075f	tunning: support k_quants; disabled rope shapes (workaround); make cache thread safe; fixed shape comprison	2023-06-18 14:27:56 +08:00
mqy	21e9379707	tunning: add f16, todo: f32 failed with CL	2023-06-18 14:27:56 +08:00
mqy	7c05049f8b	tunning: check GPU offloading before loading model	2023-06-18 14:27:56 +08:00
mqy	48016f685c	bulk refactored task profile to support complete fallback; enable tune by default for ease of dev	2023-06-18 14:27:56 +08:00
mqy	213f133701	initial	2023-06-18 14:27:53 +08:00