The program I’m writing is based on the Tutorial example.
I am implementing Huygens Raytracing for Macroscopic Objects which means that I have a large Screen-Size -
currently 19101x4831.
I have two Buffers which have Screen Size:
output_buffer (uchar4 , 2) which is the regular Buffer which holds RGB data
result_buffer (float4 , 2) which holds results of the EM Calculation.
If I raytrace without accesing the result_buffer everything is fine.
If I do OptiX crashes. If I reduce Screensize everything starts working again.
'tutorial.exe' (Win32): Loaded 'C:\Windows\System32\cryptbase.dll'. Symbols loaded.
First-chance exception at 0x000007FEFCE5B3DD in tutorial.exe: Microsoft C++ exception: optix::shared::CudaError at memory location 0x000000000024D1A0.
First-chance exception at 0x000007FEFCE5B3DD in tutorial.exe: Microsoft C++ exception: optix::Exception at memory location 0x000000000024E030.
The thread 0x1120 has exited with code 2 (0x2).
The thread 0x2470 has exited with code 2 (0x2).
The thread 0x1824 has exited with code 2 (0x2).
The thread 0x2af4 has exited with code 2 (0x2).
The thread 0x2f3c has exited with code 2 (0x2).
The thread 0x1e14 has exited with code 2 (0x2).
The program '[10668] tutorial.exe' has exited with code 2 (0x2).
Due to this I suppose it’s a memory size problem. Is there a way to split the Result_Buffer into multiple buffers which I could “read out” separatley?
Or any other way?
If possible i might increase ScreenSize even further.
I’d highly appreciate any help!
Mike
System Configuration:
Optix Prime 3.8
Windows 7 Enterprise
Visual Studio 2012
CUDA 7.0
NVIDIA GeForce GTX 760