llama.cpp
SYCL: Implement few same quantized type copy kernels
#13739

Merged

SYCL: Implement few same quantized type copy kernels #13739

qnixsynapse merged 6 commits into master from sycl/same_q_cpy

qnixsynapse marked this pull request as draft 1 year ago

github-actions added ggml

github-actions added SYCL

qnixsynapse marked this pull request as ready for review 1 year ago

Rbiessy commented on 2025-05-26

qnixsynapse marked this pull request as draft 1 year ago

qnixsynapse force pushed to 3cdc64b9 1 year ago

qnixsynapse force pushed from 3cdc64b9 to c8c22786 1 year ago

qnixsynapse marked this pull request as ready for review 1 year ago

qnixsynapse requested a review from

Alcpz 1 year ago

SYCL: Implement few same quantized type copy kernels

c26934dd

Use memcpy for copying contiguous tensors

608e8811

feat(sycl): add contiguous tensor copy support and device checks

faeb7f34

refactor: replace specific block copy functions with template

b36c550d

Exclude BF16 support for COPY tensors for now

b6db0056

qnixsynapse force pushed to b6db0056 1 year ago

Rbiessy commented on 2025-06-03

perf: adjust SYCL copy kernel block sizes for efficiency

4aa261af

Rbiessy approved these changes on 2025-06-06

qnixsynapse merged 228f34c9 into master 1 year ago

qnixsynapse deleted the sycl/same_q_cpy branch 1 year ago

Reviewers

Rbiessy

Alcpz

Assignees

No one assigned

Labels

ggml SYCL

Milestone

No milestone