]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
sycl: quantize and reorder the input to q8_1 when reorder is enabled (llama/13826)
authorAtharva Dubey <redacted>
Mon, 2 Jun 2025 09:12:20 +0000 (10:12 +0100)
committerGeorgi Gerganov <redacted>
Tue, 10 Jun 2025 06:59:25 +0000 (09:59 +0300)
commitdc0b52e681b8ac2ed2d5bb70709b823ae43868ab
treefd1d5c165af4834986a67f5c8cfc615b28e8fbea
parentdbbbe0c2172fe6f691a3ecccce20d6fb4efac2a7
sycl: quantize and reorder the input to q8_1 when reorder is enabled (llama/13826)

* [WIP]: fuse q8 quantization and reorder

* wip2: fuse q8 quantization and reorder

* working q8 reorder commit

* restored common.hpp

* remove debug prints

* remove unnecessary headers and remove trailing whitespace

* Update src/ggml-sycl/ggml-sycl.cpp

Co-authored-by: Alberto Cabrera Pérez <redacted>
---------

Co-authored-by: Alberto Cabrera Pérez <redacted>
src/ggml-sycl/ggml-sycl.cpp
src/ggml-sycl/mmvq.cpp
src/ggml-sycl/vecdotq.hpp