gemma.cpp/ops
Jan Wassenberg c29e9752c7 Refactor/cleanup, remove even_odd
* New compression/shared.h, remove sfp.h
* Remove unused DistortionStats b_l1_
* Move exact arithmetic functions into fp_arith
* Remove even_odd optimization for MatVec (mostly unused)
* use BF16 typedef more widely
* Add kMaxSFP constant

PiperOrigin-RevId: 670996386
2024-09-04 09:25:13 -07:00
..
dot-inl.h Experiment with compensated dot product. 2024-08-27 12:05:35 -07:00
dot_test.cc Experiment with compensated dot product. 2024-08-27 12:05:35 -07:00
fp_arith-inl.h Refactor/cleanup, remove even_odd 2024-09-04 09:25:13 -07:00
gemma_matvec_test.cc Refactor/cleanup, remove even_odd 2024-09-04 09:25:13 -07:00
matmul-inl.h Major MatMul update, 1.9-2.3x speedup on Zen4 via bf16 mul 2024-08-16 07:52:20 -07:00
matmul.h Major MatMul update, 1.9-2.3x speedup on Zen4 via bf16 mul 2024-08-16 07:52:20 -07:00
matmul_test.cc Major MatMul update, 1.9-2.3x speedup on Zen4 via bf16 mul 2024-08-16 07:52:20 -07:00
matvec-inl.h Refactor/cleanup, remove even_odd 2024-09-04 09:25:13 -07:00
ops-inl.h Avoid duplication of RMSNorm, support all activation/weight types 2024-08-28 01:26:55 -07:00
ops_test.cc Avoid duplication of RMSNorm, support all activation/weight types 2024-08-28 01:26:55 -07:00