Niklas Wenzel and GitHub
971595d669
metal : add remaining fa-vec tunings for M2 Max ( #28458 )
2026-09-06 07:37:01 +02:00
Niklas Wenzel and GitHub
6a1a922d26
metal : fix memory leak in early return ( #28399 )
2026-09-05 12:19:47 +02:00
Niklas Wenzel and GitHub
8f83678fd8
metal : add remaining fa-vec tunings for M3 Max ( #28373 )
2026-09-04 11:46:31 +02:00
Niklas Wenzel and GitHub
03dbcc53e1
ci : check for missing autoreleasepools ( #27884 )
...
* ci: check for missing autoreleasepools
* ci : generalize graphics device name pattern
2026-09-02 20:54:48 +03:00
fe2120bc9d
metal : fix more leaks due to missing autoreleasepools ( #27883 )
...
* metal : fix more leaks due to missing autoreleasepools
* metal : rename variable
* metal : fix another missing pool warning
Co-authored-by: YiChen Lv <63285796+forforever73@users.noreply.github.com >
---------
Co-authored-by: YiChen Lv <63285796+forforever73@users.noreply.github.com >
2026-09-01 13:50:47 +02:00
Niklas Wenzel and GitHub
5d4a3be26d
metal : add fa-vec tunings for M1 ( #28078 )
2026-08-31 13:58:55 +02:00
Niklas Wenzel and GitHub
17252c769a
metal : add remaining fa-vec tunings for M4 Pro ( #27915 )
2026-08-29 14:50:13 +02:00
Niklas Wenzel and GitHub
915dc6d38c
metal : fix memory leaks due to missing autoreleasepools ( #27758 )
2026-08-27 12:53:08 +03:00
Niklas Wenzel and GitHub
a278dcef04
contrib : recommend waiting for CI before merging ( #27603 )
2026-08-23 15:56:47 +03:00
Niklas Wenzel and GitHub
9a286ac98d
docs: improve Windows build instructions ( #27381 )
2026-08-21 21:49:27 +03:00
3e7344670a
Revert "common: share thread pools when n_threads differ ( #27138 )" ( #27337 )
...
* Revert "common: share thread pools when `n_threads` differ (#27138 )"
This reverts commit 04b569142dmaxk@qti.qualcomm.com >
* common: add comment about inability to share threadpool
---------
Co-authored-by: Max Krasnyansky <maxk@qti.qualcomm.com >
2026-08-19 18:05:48 +03:00
Niklas Wenzel and GitHub
04b569142d
common: share thread pools when n_threads differ ( #27138 )
2026-08-18 16:23:43 +02:00
Niklas Wenzel and GitHub
a94d563ed8
common: apply CPU parameters across tools ( #27026 )
2026-08-13 20:40:59 +02:00
Niklas Wenzel and GitHub
2468576f24
requirements: use stable torch packages on s390x ( #26864 )
2026-08-11 21:58:53 +08:00
Niklas Wenzel and GitHub
34e9ee57f5
ui: set npm min-release-age to protect against supply-chain attacks ( #26711 )
...
* ui: set npm `min-release-age` to protect against supply-chain attacks
* ui: bump to 7 days
2026-08-07 14:53:51 +02:00
Niklas Wenzel and GitHub
d8d9887228
ci: abort if build requirements are missing ( #26368 )
...
1. Abort CI if build requirements are missing.
2. Add check to make sure Git LFS has been configured.
3. Add trailing newlines to log messages.
2026-08-07 07:50:48 +03:00
Niklas Wenzel and GitHub
360e1349f0
tests: re-enable MiniMax M3 in test-llama-archs ( #26633 )
2026-08-05 17:58:34 +02:00
Niklas Wenzel and GitHub
d52ec04a66
build : remove GGML_METAL_USE_BF16 from all build scripts ( #26604 )
2026-08-05 10:44:34 +02:00
Niklas Wenzel and GitHub
6b5224cfcc
ci: fix pre-built binaries no longer working on macOS 15 and below ( #26375 )
...
* ci: fix pre-built binaries no longer working on macOS 15 and below
* ci: add macOS deployment target to disabled KleidiAI build
2026-08-04 15:03:38 +02:00
d0bfb19812
metal: fix memory unwire if model is freed without any GPU operations ( #26082 )
...
* metal: fix memory leak if model is freed without any GPU operations
* metal: run dummy work only if residency sets are used
* metal: wrap function in #if defined
* metal: measure system-wide wired memory in test
* metal: always build regression test
Co-authored-by: YiChen Lv <63285796+forforever73@users.noreply.github.com >
---------
Co-authored-by: YiChen Lv <63285796+forforever73@users.noreply.github.com >
2026-07-30 11:11:27 +03:00
Niklas Wenzel and GitHub
e0833bf686
mtmd: use RAII for setting and resetting non-causal attention ( #25723 )
...
* mtmd: use RAII for setting and resetting non-causal attention
* mtmd: drop dependency on <optional>
* mtmd: shorten class and variable names
2026-07-22 18:10:03 +02:00
