Commit Graph
7 Commits
Author SHA1 Message Date
Akarshan Biswas 225088ea76 sycl: Improve mul_mat_id memory efficiency and add BF16 fast path (#22119)
* sycl: size mul_mat_id staging buffers by routed rows

Previously src1_contiguous/dst_contiguous in ggml_sycl_mul_mat_id were
sized to ggml_nelements(src1/dst), which over-allocates when ne12 > 1
and can fail with UR_RESULT_ERROR_OUT_OF_HOST_MEMORY on Level Zero for
MoE models (notably with --cpu-moe). Size them by the actual number of
routed rows (ids->ne[1] * n_ids) instead.

* sycl: add bf16 mul_mat fast path via DNNL

When src0 is BF16 (commonly the case for lm_head / output.weight), the
existing f16 path is skipped because bf16 isn't covered, and the f32
fallback dequantizes the entire src0 slab to f32 in a single pool alloc
(row_diff*ne00 floats). For large-vocab models this can reach several
GB and fail with UR_RESULT_ERROR_OUT_OF_HOST_MEMORY on Level Zero.

Add a bf16xbf16 -> f32 DNNL matmul fast path that uses the bf16 storage
in place and only materializes a small src1 bf16 conversion buffer. bf16
matmul accumulates in f32, so it's correct even when the op requests
GGML_PREC_F32 (as lm_head does).

- gemm.hpp: map bfloat16 to dnnl::memory::data_type::bf16.
- convert.{hpp,cpp}: expose ggml_get_to_bf16_sycl for f32/f16/bf16 -> bf16.
- ggml-sycl.cpp: take the bf16 path early in ggml_sycl_op_mul_mat_sycl
  when DNNL and GGML_SYCL_HAS_BF16 are both available.
2026-04-22 20:32:56 +08:00
Anton Mitkov 65a3ebb0aa sycl: Batched mulmat rework for oneDNN dispatch (#14617) 2025-07-14 10:37:35 +01:00
Anton Mitkov ed52f3668e sycl: Remove not needed copy f16->f32 for dnnl mul mat (#14125) 2025-06-12 15:15:11 +02:00
Łukasz Ślusarczyk 9c404ed54c sycl: use oneDNN for matrices multiplication (#12972) 2025-05-15 16:53:41 +02:00
Svetlozar Georgiev 9ffcc9e374 sycl: cleanup oneDNN related code (#12097) 2025-03-21 10:15:56 +08:00
Akarshan BiswasandAbhilash Majumder 83ed24a97b SYCL: Reduce most of the compiler warnings (#10748)
* Try to reduce some unused and typecast warnings

* Reduce compiler warnings step 2

* add a newline at the end of the file

* Initialize nreduce as size_t

* [SYCL] Remove pragma directives from mmq.cpp

* SYCL: mmq add condition to prevent blocks_per_tile_x_row variable from becoming 0

* SYCL softmax: Initialize nreduce as size_t

* ggml-sycl.cpp: fix some trailing whitespaces

* SYCL: remove the unused variables instead of commenting it out

* SYCL poo2d kernel: set NAN for invalid pooling op

* SYCL gemm.hpp: remove pragma directives

* SYCL gemm.hpp: use const cast to properly support dnnl::memory

* SYCL: wkv6 remove a comment

* SYCL: clean comments step 2

* SYCL: clean comments and variables step 3

* SYCL: Use GGML_UNUSED for unused variables

* SYCL: remove extra empty lines and a comment

* Remove TODO

* cleanup spaces

* add a stdout for unsupported op

* use sycl printf over fprintf

* remove prints for CI

* SYCL ggml-sycl: pool2D use sycl::nan and remove if-else block

---------

Co-authored-by: Abhilash Majumder <30946547+abhilash1910@users.noreply.github.com>
2024-12-13 12:12:15 +05:30
luoyu-intel 1731d4238f [SYCL] Add oneDNN primitive support (#9091)
* add onednn

* add sycl_f16

* add dnnl stream

* add engine map

* use dnnl for intel only

* use fp16fp16fp16

* update doc
2024-08-22 12:50:10 +08:00