| .. |
|
allocator.cc
|
Warning fix (unused var), Windows build fix (missing member variable)
|
2025-10-21 10:17:34 -07:00 |
|
allocator.h
|
Warning fix (unused var), Windows build fix (missing member variable)
|
2025-10-21 10:17:34 -07:00 |
|
args.h
|
Abort if args are unrecognized, refactor argument passing
|
2025-12-15 03:18:45 -08:00 |
|
basics.cc
|
Replace mt19937 with new generator to enable parallel sampling
|
2025-09-04 23:49:10 -07:00 |
|
basics.h
|
Change (old) attention behavior to disallow wraparound, enforced via assertion.
|
2025-11-04 11:52:40 -08:00 |
|
basics_test.cc
|
Replace mt19937 with new generator to enable parallel sampling
|
2025-09-04 23:49:10 -07:00 |
|
mat.cc
|
Add 8-bit integer quantization (I8Stream) to Gemma.cpp.
|
2025-10-15 09:25:20 -07:00 |
|
mat.h
|
Rewrote flash attention to use BF16, transpose k and v, rewrote the task distribution, increase parallelism on decode, and use double the registers for the core of flash attention.
|
2026-02-13 01:58:48 -08:00 |
|
test_util.h
|
Rewrote flash attention to use BF16, transpose k and v, rewrote the task distribution, increase parallelism on decode, and use double the registers for the core of flash attention.
|
2026-02-13 01:58:48 -08:00 |
|
threading.cc
|
Major cleanup of profiler zones, add Caller annotation for all pool.Run
|
2025-10-23 01:54:24 -07:00 |
|
threading.h
|
1.01x speedup: improved autotune
|
2025-10-27 05:35:31 -07:00 |
|
threading_context.cc
|
Add tensor stats and output
|
2025-12-11 22:52:46 -08:00 |
|
threading_context.h
|
Abort if args are unrecognized, refactor argument passing
|
2025-12-15 03:18:45 -08:00 |
|
threading_test.cc
|
Change (old) attention behavior to disallow wraparound, enforced via assertion.
|
2025-11-04 11:52:40 -08:00 |
|
topology.cc
|
Avoid warning when OS affinity limits us to the second socket
|
2025-12-08 07:10:43 -08:00 |
|
topology.h
|
Avoid warning when OS affinity limits us to the second socket
|
2025-12-08 07:10:43 -08:00 |
|
zones.cc
|
Rewrote flash attention to use BF16, transpose k and v, rewrote the task distribution, increase parallelism on decode, and use double the registers for the core of flash attention.
|
2026-02-13 01:58:48 -08:00 |
|
zones.h
|
Rewrote flash attention to use BF16, transpose k and v, rewrote the task distribution, increase parallelism on decode, and use double the registers for the core of flash attention.
|
2026-02-13 01:58:48 -08:00 |
7b55d41f46