Does it work on AMD Radeon 780M?

[Giovo17]

Are there any tests on AMD integrated graphics, such as the Radeon 780M?

[Mar2ck]

From the Readme:

Integrated GPUs (as tested with Radeon 680M) work in a limited way. Some rarely used GPU operations (abort, printf, etc.) will hang or crash the application. Additionally, performance library support (cuBLAS, cuDNN, etc.) might be limited, rendering more complex applications inoperable.

[HysterLc]

i have tested it in stable diffusion,it works well

[Giovo17]

i have tested it in stable diffusion,it works well

@HysterLc Thank you so much for the feedback, could you please share your configuration and setup?

[HysterLc]

i have tested it in stable diffusion,it works well

@HysterLc Thank you so much for the feedback, could you please share your configuration and setup?

https://github.com/likelovewant/ROCmLibs-for-gfx1103-AMD780M-APU- the project is here.
there are more comprehensive information but i dont know if it can be accessed outside china
https://www.bilibili.com/read/cv32897766/?jump_opus=1

[JunHe001]

1. install AMD HIP SDK for Windows
   https://www.amd.com/en/developer/resources/rocm-hub/hip-sdk.html
2. download ZLUDA and unzip it to folder
   https://github.com/lshqqytiger/ZLUDA
3. download rocm gfx1103 AMD780M phoenix V3.7z and unzip it to folder, thanks HysterLc.
   https://github.com/likelovewant/ROCmLibs-for-gfx1103-AMD780M-APU-
4. add "???:\Program Files\AMD\ROCm\5.7\bin" to env path
5. add ZLUDA to path
6. copy content of rocm gfx1103 AMD780M phoenix V3 and overwrite to "???:\Program Files\AMD\ROCm\5.7\bin\rocblas.dll" and "???:\Program Files\AMD\ROCm\5.7\bin\rocblas\library"
7. My ADM 780M work with it

2024-03-30 15:49:23.703106: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cudart64_101.dll'; dlerror: cudart64_101.dll not found
2024-03-30 15:49:23.703188: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
D:\Program Files\Python\Python38\lib\site-packages\stable_baselines_init_.py:32: UserWarning: stable-baselines is in maintenance mode, please use Stable-Baselines3 (SB3) for an up-to-date version. You can find a migration guide in SB3 documentation.

... ...

D:\Program Files\Python\Python38\lib\site-packages\gym\logger.py:30: UserWarning: �[33mWARN: Box bound precision lowered by casting to float16�[0m
warnings.warn(colorize('%s: %s'%('WARN', msg % args), 'yellow'))

2024-03-30 15:49:25.380976: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
2024-03-30 15:49:25.391200: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x192a2ea3fd0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2024-03-30 15:49:25.391259: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version
2024-03-30 15:49:25.399572: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library nvcuda.dll
2024-03-30 15:49:25.423385: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1561] Found device 0 with properties:
pciBusID: 0000:c4:00.0 name: AMD Radeon 780M Graphics [ZLUDA] computeCapability: 8.8
coreClock: 2.7GHz coreCount: 6 deviceMemorySize: 28.71GiB deviceMemoryBandwidth: 0B/s
2024-03-30 15:49:25.423820: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cudart64_101.dll'; dlerror: cudart64_101.dll not found
2024-03-30 15:49:25.424125: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cublas64_10.dll'; dlerror: cublas64_10.dll not found
2024-03-30 15:49:25.424359: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cufft64_10.dll'; dlerror: cufft64_10.dll not found
2024-03-30 15:49:25.424634: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'curand64_10.dll'; dlerror: curand64_10.dll not found
2024-03-30 15:49:25.424922: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cusolver64_10.dll'; dlerror: cusolver64_10.dll not found
2024-03-30 15:49:25.425203: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cusparse64_10.dll'; dlerror: cusparse64_10.dll not found
2024-03-30 15:49:25.425501: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cudnn64_7.dll'; dlerror: cudnn64_7.dll not found
2024-03-30 15:49:25.425577: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1598] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.
Skipping registering GPU devices...
2024-03-30 15:49:25.425936: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1102] Device interconnect StreamExecutor with strength 1 edge matrix:
2024-03-30 15:49:25.426147: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1108] 0
2024-03-30 15:49:25.426438: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1121] 0: N
2024-03-30 15:49:25.431664: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x192a2ea5e60 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2024-03-30 15:49:25.431726: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): AMD Radeon 780M Graphics [ZLUDA], Compute Capability 8.8

[JunHe001]

copy and rename some dll file in ZLUDA folder could make application load part of dll success as below

2024-03-30 16:46:21.628376: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1561] Found device 0 with properties:
pciBusID: 0000:c4:00.0 name: AMD Radeon 780M Graphics [ZLUDA] computeCapability: 8.8
coreClock: 2.7GHz coreCount: 6 deviceMemorySize: 28.71GiB deviceMemoryBandwidth: 0B/s
2024-03-30 16:46:21.628900: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cudart64_101.dll'; dlerror: cudart64_101.dll not found
2024-03-30 16:46:21.628931: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cublas64_10.dll
2024-03-30 16:46:21.628956: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cufft64_10.dll
2024-03-30 16:46:21.629173: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'curand64_10.dll'; dlerror: curand64_10.dll not found
2024-03-30 16:46:21.629385: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cusolver64_10.dll'; dlerror: cusolver64_10.dll not found
2024-03-30 16:46:21.629411: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusparse64_10.dll
2024-03-30 16:46:21.629994: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cudnn64_7.dll'; dlerror: cudnn64_7.dll not found
2024-03-30 16:46:21.630026: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1598] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.
Skipping registering GPU devices...
2024-03-30 16:46:21.630073: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1102] Device interconnect StreamExecutor with strength 1 edge matrix:
2024-03-30 16:46:21.630094: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1108] 0
2024-03-30 16:46:21.630113: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1121] 0: N