Minimal GPU requirements for demo

[FedorChe]

Hello,

What are the minimal requirements for the GPU to run the demo:
deep_image_analogy.exe ../models/ ../demo/content.png ../demo/style.png ../demo/output/ 0 0.5 2 0
?

I have nVIdia GTX 970 and Windows 8.1, and it crashes unfortunately somewhere in CUDA with the log:

[libprotobuf WARNING ..\src\google\protobuf\io\coded_stream.cc:537] Reading dang
erously large protocol message. If the message turns out to be larger than 2147
483647 bytes, parsing will be halted for security reasons. To increase the limi
t (or to disable these warnings), see CodedInputStream::SetTotalBytesLimit() in
google/protobuf/io/coded_stream.h.
[libprotobuf WARNING ..\src\google\protobuf\io\coded_stream.cc:78] The total num
ber of bytes read was 574671192
[libprotobuf WARNING ..\src\google\protobuf\io\coded_stream.cc:537] Reading dang
erously large protocol message. If the message turns out to be larger than 2147
483647 bytes, parsing will be halted for security reasons. To increase the limi
t (or to disable these warnings), see CodedInputStream::SetTotalBytesLimit() in
google/protobuf/io/coded_stream.h.
[libprotobuf WARNING ..\src\google\protobuf\io\coded_stream.cc:78] The total num
ber of bytes read was 574671192
Finding nearest neighbor field using PatchMatch Algorithm at layer:conv5_1.
Finding nearest neighbor field using PatchMatch Algorithm at layer:conv4_1.
Finding nearest neighbor field using PatchMatch Algorithm at layer:conv3_1.
Finding nearest neighbor field using PatchMatch Algorithm at layer:conv2_1.
Finding nearest neighbor field using PatchMatch Algorithm at layer:conv1_1.
Saving flow result.
Finished finding ann. Time : 37.588
WARNING: Logging before InitGoogleLogging() is written to STDERR
F0827 15:12:08.460532 6884 syncedmem.hpp:31] Check failed: error == cudaSuccess
(30 vs. 0) unknown error

This is both for exe-from-git, and for build by myself.

In Debug build I see the error:
Finding nearest neighbor field using PatchMatch Algorithm at layer:conv5_1.
F0827 15:41:30.334332 4332 pooling_layer.cu:212] Check failed: error == cudaSuc
cess (7 vs. 0) too many resources requested for launch

Best regards,
Fedor

[frostbitten]

Hi there, I had the same issue and it seems others here have too (active and closed tickets). For me and many the solution is to disable Timeout Detection & Recovery (TDR) on your system. See: https://docs.nvidia.com/gameworks/content/developertools/desktop/timeout_detection_recovery.htm