Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

UnsatisfiedLinkError: no jniopenblas_nolapack #10

Open
brad-miller opened this issue Feb 17, 2022 · 4 comments
Open

UnsatisfiedLinkError: no jniopenblas_nolapack #10

brad-miller opened this issue Feb 17, 2022 · 4 comments
Assignees
Labels
enhancement New feature or request help wanted Extra attention is needed

Comments

@brad-miller
Copy link

Hi

I'm trying to test deepvision library 0.7.0 with processing 4.05b
on Mac 1M Pro, MacOS 12.2.1

I think javacpp and the dependencies, need to be updated to 1.5.7 to include macOSX-arm64
I'm testing Apple Silicon GPU performance against CUDA (RTX3090)

note I'm using javaFX render (as the openGL issues are on going with 4.04b (Java 17.0.2)and above

A library used by this sketch relies on native code that is not available.
UnsatisfiedLinkError: no jniopenblas_nolapack in java.library.path:

Kind regards

@cansik cansik added the enhancement New feature or request label Feb 17, 2022
@cansik cansik self-assigned this Feb 17, 2022
cansik added a commit that referenced this issue Feb 17, 2022
cansik added a commit that referenced this issue Feb 17, 2022
@cansik
Copy link
Owner

cansik commented Feb 17, 2022

@brad-miller I have updated the dependencies to 1.5.7 and added macosx-arm64 as a platform. I also added a CI build / release workflow to automatically upload the pre-built libraries. Would be great if you could test it after it has been built (~30min). I do not own a M1 Mac so I depend on your help there.

But just a question in general: Does de OpenCV DNN module even support apple silicon GPU?

@cansik cansik added the help wanted Extra attention is needed label Feb 17, 2022
@cansik
Copy link
Owner

cansik commented Feb 17, 2022

The release is now ready to download: https://github.com/cansik/deep-vision-processing/releases/tag/v0.8.0-alpha

@cansik
Copy link
Owner

cansik commented Feb 17, 2022

@brad-miller I am enabling now OpenCL support if it is possible. But it is possible to disable it and stick to the CPU with the following line:

DeepVision vision = new DeepVision(this);
vision.setUseDefaultBackend(true);

Version v0.8.1 is still building.

@brad-miller
Copy link
Author

Hi re testing

MacOS 12.2.1
Apple M1 Max

Processing_4.05b
Deepvision lib 0.8.1

Processing video library using GStreamer 1.20.0
Details here:
https://github.com/processing/processing-video/issues/194#issuecomment-1034972288

javaFX render

Test example: MidasDepthEstimationWebcam
Tested example file: MidasDepthEstimationWebcam

With both internal videocam and realsense435
Using FaceTime HD Camera
Enabling OpenCL DNN MODULE does not effect frameRate

FrameRate: 2fps

Output:
Processing video library using GStreamer 1.20.0
creating model...
loading model...
DNN OpenCL backend enabled
Available cameras:
[0] "Intel(R) RealSense(TM) Depth Camera 435 with RGB Module Depth"
[1] "FaceTime HD Camera"

Windows 10 (21H2)
Processing_4.05b
Deepvision lib 0.7.0
With CUDA enabled
Videocam input

Processing video library using GStreamer 1.16.2
FrameRate : 33fps

Re question in general: Does de OpenCV DNN module even support apple silicon GPU?
Coming in version 5.0 release however there is preliminary support for iOS in OpenCV 4.5.5 build

https://github.com/opencv/opencv/issues/21576

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request help wanted Extra attention is needed
Projects
None yet
Development

No branches or pull requests

2 participants