Niklas Wenzel 915dc6d38c metal : fix memory leaks due to missing autoreleasepools (#27758)

2026-08-27 12:53:08 +03:00

..

kernels

ggml-metal: add chunked SSD MMA for Mamba-2 prefill optimization (#26647)

2026-08-26 11:57:07 +03:00

CMakeLists.txt

metal : per-device tuned (Q, NE) for flash-attn vec (#26570)

2026-08-24 19:22:27 +03:00

ggml-metal-common.cpp

models : optimize qwen3next graph (#19375)

2026-02-14 12:57:36 +02:00

ggml-metal-common.h

metal : refactor + optimize v2 (#15995)

2025-09-17 20:38:12 +03:00

ggml-metal-context.h

metal : support virtual devices (#18919)

2026-02-02 14:29:44 +02:00

ggml-metal-context.m

metal : fix memory leaks due to missing autoreleasepools (#27758)

2026-08-27 12:53:08 +03:00

ggml-metal-device.cpp

ggml-metal: add chunked SSD MMA for Mamba-2 prefill optimization (#26647)

2026-08-26 11:57:07 +03:00

ggml-metal-device.h

ggml-metal: add chunked SSD MMA for Mamba-2 prefill optimization (#26647)

2026-08-26 11:57:07 +03:00

ggml-metal-device.m

metal : fix memory leaks due to missing autoreleasepools (#27758)

2026-08-27 12:53:08 +03:00

ggml-metal-impl.h

ggml-metal: add chunked SSD MMA for Mamba-2 prefill optimization (#26647)

2026-08-26 11:57:07 +03:00

ggml-metal-ops.cpp

ggml-metal: add chunked SSD MMA for Mamba-2 prefill optimization (#26647)

2026-08-26 11:57:07 +03:00

ggml-metal-ops.h

metal : dequantize quantized KV to F16 before flash attention (#27390)

2026-08-20 13:43:59 +03:00

ggml-metal-tuning.cpp

metal : per-device tuned (Q, NE) for flash-attn vec (#26570)

2026-08-24 19:22:27 +03:00

ggml-metal-tuning.h

metal : per-device tuned (Q, NE) for flash-attn vec (#26570)

2026-08-24 19:22:27 +03:00

ggml-metal.cpp

metal : null-check buffer alloc to fix OOM crash (#25371)

2026-08-25 14:35:39 +03:00