llama.cpp
cuda : improve cuda pool efficiency using virtual memory
#4606
Merged
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Overview
Commits
22
Changes
View On
GitHub
cuda : improve cuda pool efficiency using virtual memory
#4606
slaren
merged 22 commits into
master
from
sl/cuda-virt-pool
cuda : improve cuda pool efficiency using virtual memory
0d77fbd7
JohannesGaessler
commented on 2023-12-22
fix mixtral
eb223dcd
fix cmake build
bd78dc9a
slaren
marked this pull request as ready for review
1 year ago
slaren
force pushed
1 year ago
check for vmm support, disable for hip
872408cf
slaren
force pushed
to
872408cf
1 year ago
ggerganov
approved these changes on 2023-12-23
JohannesGaessler
commented on 2023-12-23
fix hip build
9452d0d5
clarify granularity
20860dae
JohannesGaessler
approved these changes on 2023-12-23
move all caps to g_device_caps
4c0f300a
refactor error checking
545f23d0
slaren
force pushed
1 year ago
add cuda_pool_alloc, refactor most pool allocations
110b5055
slaren
force pushed
to
110b5055
1 year ago
slaren
requested a review
from
ggerganov
1 year ago
fix hip build
b7da1ba0
slaren
force pushed
to
b7da1ba0
1 year ago
CUBLAS_TF32_TENSOR_OP_MATH is not a macro
d8b06c21
ggerganov
added
high priority
ggerganov
added
need feedback
slaren
force pushed
1 year ago
slaren
force pushed
1 year ago
more hip crap
9f5ac6d2
slaren
force pushed
to
9f5ac6d2
1 year ago
llama : fix msvc warnings
5eb62622
ggml : fix msvc warnings
6fe9da0f
minor
d8883623
Merge remote-tracking branch 'origin/master' into sl/cuda-virt-pool
26e97b58
minor
ab6ad5e6
cuda : fallback to CPU on host buffer alloc fail
5acc9e50
JohannesGaessler
approved these changes on 2023-12-24
Update ggml-cuda.cu
b9c5a6e7
Update ggml-cuda.cu
3081c4e7
ensure allocations are always aligned
3ad45fc3
act_size -> actual_size
532cb9b9
ggerganov
approved these changes on 2023-12-24
slaren
merged
5bf3953d
into master
1 year ago
slaren
deleted the sl/cuda-virt-pool branch
1 year ago
SomeoneSerge
commented on 2023-12-25
Login to write a write a comment.
Login via GitHub
Reviewers
JohannesGaessler
ggerganov
SomeoneSerge
sorasoras
Assignees
No one assigned
Labels
high priority
need feedback
Milestone
No milestone
Login to write a write a comment.
Login via GitHub