llama.cpp
Optimize RWKV6 Operator Naming and Implement Multi-core CPU/ SYCL Acceleration
#10133
Merged
Merged
Optimize RWKV6 Operator Naming and Implement Multi-core CPU/ SYCL Acceleration #10133
airMeng
merged
33 commits into
ggml-org:master
from
zhiyuan1i:master
zhiyuan1i
changed the title
rwkv6: rename to wkv6
f66c75a4
rwkv6: support avx2 avx512 armv8 armv9
b4254c55
rwkv6: update cuda file name
e198f7b9
rwkv6: rename params
3f75f121
wkv on sycl
2fc42b6a
sycl: add some ops
bee1cec7
sycl: Enhance OP support judgment
1c58096f
Merge branch 'ggerganov:master' into master
042c3e0f
wkv6: drop armv9 and tranfer to GGML style
811aa872
flake.lock: Update (#10146)
4d266310
metal : minor fixup in FA kernel (#10143)
b1896308
ggml : move CPU backend to a separate file (#10144)
89812b15
metal : fix minor string leaks (ggml/1004)
8050d021
cmake : make it possible linking ggml as external lib (ggml/1003)
eb5711c4
sync : ggml
153251f7
Merge branch 'ggerganov:master' into master
5f792141
fix: update changes to upstream
61c665b7
fix: add defualt
9ea34a78
Update ggml/src/ggml-sycl/outprod.cpp
8c7b4ec2
Update ggml/src/ggml-sycl/wkv6.cpp
bb0685fa
update the function to use appropriate types
81cb3012
fix define error
a878502f
Update ggml/src/ggml-cpu.c
b8160247
add appropriate asserts
72e44325
move element-wise functions outside
35a1a2df
Update ggml/src/ggml-sycl/concat.cpp
6a1e977e
put the declaration outside the loop
a749ba77
rewrite to be more inline with the common pattern for distributing th…
4693b461
use recommended way GGML_TENSOR_LOCALS
4574795c
Merge branch 'ggerganov:master' into master
acb1b9d2
remove some codes
e264c35f
NeoZhangJianyu
approved these changes
on 2024-11-05
update lint
623db3b0
Merge branch 'ggerganov:master' into master
98e070c1
ggerganov
approved these changes
on 2024-11-05
Reviewers
Assignees
No one assigned
Labels
documentation
testing
Nvidia GPU
ggml
SYCL
Milestone
No milestone
Login to write a write a comment.
Login via GitHub