github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #7591] Cant compile ollma 0.4.1 on arm Jetson agx orin #51351

New Issue
Closed
opened 2026-04-28 19:38:26 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-28 19:38:26 -05:00
Owner
Copy Link

Originally created by @rebotnix on GitHub (Nov 9, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7591

What is the issue?

I have some issues to compile latest ollma on an ARM nvidia jetson plattform cuda version is 11.2 with jetpack 5.1.2.

ggml-quants.c: In function ‘ggml_vec_dot_q4_0_q8_0’:
ggml-quants.c:4023:52: warning: implicit declaration of function ‘vmmlaq_s32’; did you mean ‘vmlaq_s32’? [-Wimplicit-function-declaration]
4023 | sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)),
| ^~~~~~~~~~
| vmlaq_s32
ggml-quants.c:4023:52: error: incompatible type for argument 1 of ‘vcvtq_f32_s32’
4023 | sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)),
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| |
| int
4024 | l1, r1)), l2, r2)), l3, r3))), scale);
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ggml-cpu-impl.h:164,
from ggml-quants.c:32:
/usr/lib/gcc/aarch64-linux-gnu/9/include/arm_neon.h:15168:26: note: expected ‘int32x4_t’ but argument is of type ‘int’
15168 | vcvtq_f32_s32 (int32x4_t __a)
| ~~~~~~~~~~^~~
ggml-quants.c: In function ‘ggml_vec_dot_q4_1_q8_1’:
ggml-quants.c:4603:52: error: incompatible type for argument 1 of ‘vcvtq_f32_s32’
4603 | sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)),
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| |
| int
4604 | l1, r1)), l2, r2)), l3, r3))), scale);
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ggml-cpu-impl.h:164,
from ggml-quants.c:32:
/usr/lib/gcc/aarch64-linux-gnu/9/include/arm_neon.h:15168:26: note: expected ‘int32x4_t’ but argument is of type ‘int’
15168 | vcvtq_f32_s32 (int32x4_t __a)
| ~~~~~~~~~~^~~
ggml-quants.c: In function ‘ggml_vec_dot_q8_0_q8_0’:
ggml-quants.c:5607:52: error: incompatible type for argument 1 of ‘vcvtq_f32_s32’
5607 | sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)),
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| |
| int
5608 | l1, r1)), l2, r2)), l3, r3))), scale);
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from ggml-cpu-impl.h:164,
from ggml-quants.c:32:
/usr/lib/gcc/aarch64-linux-gnu/9/include/arm_neon.h:15168:26: note: expected ‘int32x4_t’ but argument is of type ‘int’
15168 | vcvtq_f32_s32 (int32x4_t __a)
| ~~~~~~~~~~^~~
make[2]: *** [make/gpu.make:85: /opt/ssd/ollama/ollama/llama/build/linux-arm64/runners/cuda_v11/ollama_llama_server] Error 1
make[2]: Leaving directory '/opt/ssd/ollama/ollama/llama'
make[1]: *** [Makefile:41: cuda_v11] Error 2
make[1]: Leaving directory '/opt/ssd/ollama/ollama/llama'
make: *** [Makefile:4: all] Error 2

Did anyone has found an fix for this?

OS

Linux

GPU

Nvidia embedded AGX Orin 64GB

CPU

ARM

Ollama version

0.4.1

Originally created by @rebotnix on GitHub (Nov 9, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7591 ### What is the issue? I have some issues to compile latest ollma on an ARM nvidia jetson plattform cuda version is 11.2 with jetpack 5.1.2. ggml-quants.c: In function ‘ggml_vec_dot_q4_0_q8_0’: ggml-quants.c:4023:52: warning: implicit declaration of function ‘vmmlaq_s32’; did you mean ‘vmlaq_s32’? [-Wimplicit-function-declaration] 4023 | sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), | ^~~~~~~~~~ | vmlaq_s32 ggml-quants.c:4023:52: error: incompatible type for argument 1 of ‘vcvtq_f32_s32’ 4023 | sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | int 4024 | l1, r1)), l2, r2)), l3, r3))), scale); | ~~~~~~~~~~~~~~~~~~~~~~~~~~~ In file included from ggml-cpu-impl.h:164, from ggml-quants.c:32: /usr/lib/gcc/aarch64-linux-gnu/9/include/arm_neon.h:15168:26: note: expected ‘int32x4_t’ but argument is of type ‘int’ 15168 | vcvtq_f32_s32 (int32x4_t __a) | ~~~~~~~~~~^~~ ggml-quants.c: In function ‘ggml_vec_dot_q4_1_q8_1’: ggml-quants.c:4603:52: error: incompatible type for argument 1 of ‘vcvtq_f32_s32’ 4603 | sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | int 4604 | l1, r1)), l2, r2)), l3, r3))), scale); | ~~~~~~~~~~~~~~~~~~~~~~~~~~~ In file included from ggml-cpu-impl.h:164, from ggml-quants.c:32: /usr/lib/gcc/aarch64-linux-gnu/9/include/arm_neon.h:15168:26: note: expected ‘int32x4_t’ but argument is of type ‘int’ 15168 | vcvtq_f32_s32 (int32x4_t __a) | ~~~~~~~~~~^~~ ggml-quants.c: In function ‘ggml_vec_dot_q8_0_q8_0’: ggml-quants.c:5607:52: error: incompatible type for argument 1 of ‘vcvtq_f32_s32’ 5607 | sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | int 5608 | l1, r1)), l2, r2)), l3, r3))), scale); | ~~~~~~~~~~~~~~~~~~~~~~~~~~~ In file included from ggml-cpu-impl.h:164, from ggml-quants.c:32: /usr/lib/gcc/aarch64-linux-gnu/9/include/arm_neon.h:15168:26: note: expected ‘int32x4_t’ but argument is of type ‘int’ 15168 | vcvtq_f32_s32 (int32x4_t __a) | ~~~~~~~~~~^~~ make[2]: *** [make/gpu.make:85: /opt/ssd/ollama/ollama/llama/build/linux-arm64/runners/cuda_v11/ollama_llama_server] Error 1 make[2]: Leaving directory '/opt/ssd/ollama/ollama/llama' make[1]: *** [Makefile:41: cuda_v11] Error 2 make[1]: Leaving directory '/opt/ssd/ollama/ollama/llama' make: *** [Makefile:4: all] Error 2 Did anyone has found an fix for this? ### OS Linux ### GPU Nvidia embedded AGX Orin 64GB ### CPU ARM ### Ollama version 0.4.1
GiteaMirror added the buildbugnvidia labels 2026-04-28 19:38:27 -05:00
GiteaMirror commented 2026-04-28 19:38:33 -05:00
Author
Owner
Copy Link

@davidADSP commented on GitHub (Nov 10, 2024):

Perhaps you need to upgrade to jetpack 6?

<!-- gh-comment-id:2466665405 --> @davidADSP commented on GitHub (Nov 10, 2024): Perhaps you need to upgrade to jetpack 6?
GiteaMirror commented 2026-04-28 19:38:36 -05:00
Author
Owner
Copy Link

@rebotnix commented on GitHub (Nov 10, 2024):

I dont know yet, older version including 0.4.1 of ollma compiled. Have to check the ggml-quants.c files, it seem ARM neon related, not cuda. @davidADSP Will check today.

<!-- gh-comment-id:2466666780 --> @rebotnix commented on GitHub (Nov 10, 2024): I dont know yet, older version including 0.4.1 of ollma compiled. Have to check the ggml-quants.c files, it seem ARM neon related, not cuda. @davidADSP Will check today.
GiteaMirror commented 2026-04-28 19:38:44 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Nov 13, 2024):

Now that #7217 is merged, you should be able to build main as long as you rebase.

<!-- gh-comment-id:2471988433 --> @dhiltgen commented on GitHub (Nov 13, 2024): Now that #7217 is merged, you should be able to build main as long as you rebase.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug build nvidia
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#51351
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?