sync : ggml

vulkan : add fp16 support for the conv_2d kernel (llama/14872)

vulkan: skip empty set_rows to avoid invalid API usage (llama/14860)

HIP: Enable Matrix cores for MMQ Kernels, Enable stream-K for CDNA 3 (llama/14624)

CANN: Implement GLU ops (llama/14884)

musa: fix build warnings (unused variable) (llama/14869)

ggml-cpu : disable GGML_NNPA by default due to instability (llama/14880)

metal: SSM_SCAN performance (llama/14743)

opencl: add fused `rms_norm_mul` (llama/14841)

ggml : remove invalid portPos specifiers from dot files (llama/14838)

rpc : check for null buffers in get/set/copy tensor endpoints (llama/14868)

sched : fix multiple evaluations of the same graph with pipeline parallelism (llama/14855)

musa: upgrade musa sdk to rc4.2.0 (llama/14498)

cmake : Indent ggml-config.cmake (ggml/1310)

sycl: fixed semantics of block offset calculation (llama/14814)

metal : fix fusion across different encoders (llama/14849)

sycl: fix undefined variable in work group size check (llama/14843)

CUDA: fix overflow in FA, tune performance (llama/14840)

CUDA: fix compilation with GGML_CUDA_F16 (llama/14837)

CUDA: fix quantized KV cache + multiple sequences (llama/14822)

ggml: fix loongarch quantize_row_q8_1 error (llama/14827)

CANN: weight format to NZ for Ascend310P3 (llama/14407)

CUDA: add fused rms norm (llama/14800)

vulkan: fix rms_norm_mul to handle broadcasting dim0 (llama/14817)

cuda : implement bf16 cpy ops and enable bf16 cont (llama/14763)

opencl: remove unreachable `return` (llama/14806)

cuda: remove linking to cublasLt (llama/14790)

opencl: fix `im2col` when `KW!=KH` (llama/14803)

opencl: add conv2d kernel (llama/14403)

sycl: Fix im2col (llama/14797)

kleidiai: add support for get_rows (llama/14676)

vulkan/cuda: Fix im2col when KW!=KH (llama/14789)

ggml: adds CONV_2D op and direct GEMM Vulkan implementation (llama/14316)

vulkan: Add logging for bf16 features to ggml_vk_print_gpu_info (#13274) (llama/14707)

Vulkan: Fix fprintf format-security warning (llama/14770)

cmake : fix usage issues (ggml/1257)

ggml-cpu : remove stdlib include from repack.cpp (ggml/1276)

Support static xcframework packaging in build-xcframework.sh (#3322)

examples : add note about WHISPER_WASM_SINGLE_FILE [no ci] (#3332)

ci : add paths to build.yml (#3333)

musa: upgrade musa sdk to rc4.2.0 (#3324)

server : hide language probabilities option behind flag (#3328)

go: fix Mac OS X builds (#3310)

sync : ggml

metal : fuse add, mul + add tests (llama/14596)

cuda : Fix Gemma3n not executed as CUDA_GRAPH on NVGPUs (llama/14741)

CUDA: set_rows + cpy.cu refactor (llama/14712)

use max work group size for device to replace the magic number (llama/14732)

ggml: Add initial WebGPU backend (llama/14521)

llama : add high-throughput mode (llama/14363)