I have a correct cuda project, which could run on my another machine for sure.
But for my personal desktop, after debugging this project for several times with small change (change n from 16 to 48, both work for my another machine), I can not run it correctly, the debug information is :
NVAPI: bLhThunkInit: failed assert: lhThunk.pfnGetDisplayConfigBufferSizes
NVAPI: bLhThunkInit: failed assert: lhThunk.pfnQueryDisplayConfig
NVAPI: bLhThunkInit: failed assert: lhThunk.pfnSetDisplayConfig
NVAPI: bLhThunkInit: failed assert: lhThunk.pfnDisplayConfigGetDeviceInfo
First-chance exception at 0x770d42eb in mytestProj.exe: Microsoft C++ exception: cudaError_enum at memory location 0x0012faf4…
First-chance exception at 0x770d42eb in mytestProj.exe: Microsoft C++ exception: cudaError_enum at memory location 0x0012faf4…
First-chance exception at 0x770d42eb in mytestProj.exe: Microsoft C++ exception: cudaError_enum at memory location 0x0012f924…
First-chance exception at 0x770d42eb in mytestProj.exe: Microsoft C++ exception: cudaError_enum at memory location 0x0012f924…
after about 30 mins, I try one more time it can run with no errors again. but after several runs, it can not run again. so weird… :wacko:
I use Galaxy GTS 250 512MB and VS2008 for my desktop.
In fact, I run the volkov’s matrix decomposition code. It returns these kind of messages. By checking step by step, I found it fails at cuda memory allocation function.
My home desktop is using vista32bit. I dont know whether it is due to the os or hardware problem.