ROCm / MIVisionX

MIVisionX toolkit is a set of comprehensive computer vision and machine intelligence libraries, utilities, and applications bundled into a single toolkit. AMD MIVisionX also delivers a highly optimized open-source implementation of the Khronos OpenVX™ and OpenVX™ Extensions.
https://rocm.docs.amd.com/projects/MIVisionX/en/latest/
MIT License
186 stars 74 forks source link

Neural Net - Can This framework be used with OpenCL on Windows? #410

Open eladmaimoni opened 3 years ago

eladmaimoni commented 3 years ago

I am very confused by the various libraries offered by AMD.

So basically: If I want to run inference on Windows with OpenCL Interoperability. What are my possibilities? what path should I take?

Thanks

kiritigowda commented 3 years ago

@elad8a for windows currently we offer WinML for inference path. In WinML you hand off the ONNX model to underlying libraries and don't really have control over the model optimizations, but you can choose where the model is run by using one of the options below

LearningModelDeviceKind::Default
LearningModelDeviceKind::Cpu,
LearningModelDeviceKind::DirectX,
LearningModelDeviceKind::DirectXHighPerformance,
LearningModelDeviceKind::DirectXMinPower
eladmaimoni commented 3 years ago

@kiritigowda like I said, I need OpenCL Interoperability. WinML / DirectML are based on DirectML which is based on DirectX 12.

AMD does not offer any interoperability between OpenCL & DirectX12 so I can't use this for my needs.

Is there a reason why MIOpen which is OpenCL based is not exposed to Windows Clients?

ppanchad-amd commented 2 weeks ago

@eladmaimoni Apologies for the lack of response. Do you still need assistance with this ticket? Thanks!