The P-1 factoring CUDA program

[ET_]

Originally Posted by James Heinrich

Note that the "<gpu> fft.txt" and "<gpu> threads.txt" files are distinct from each other.

<gpu> fft.txt should look something like

Device              GeForce GTX 670
Compatibility       3.0
clockRate (MHz)     980
memClockRate (MHz)  3004

  fft    max exp  ms/iter
    4      85933   0.0697
   16     333803   0.1153
   32     657719   0.1306
   36     738083   0.1618
   48     978041   0.1635
... skip a whole bunch of fft lines ...
28800  511382147  76.5273
32768  580225813  79.6749

Whereas "<gpu> threads.txt" should be quite short (and more cryptic), mine looks like:

17496  256   64  512  45.9160
 3456  256  128   32   8.0790

I suspect it didn't make a "<gpu> threads.txt" file for you because it appears to have failed partway through the process:

Thanks James.

From what you said, I assume that there should be 2 distinct files: the first created by cufftbench 1.8192 1, the second by -cufftbench 4096 4096 4

I'll try to modify the r parameter of the second bench run and see if it suffices.

Luigi

[ET_]

Originally Posted by ET_

Thanks James.

From what you said, I assume that there should be 2 distinct files: the first created by cufftbench 1.8192 1, the second by -cufftbench 4096 4096 4

I'll try to modify the r parameter of the second bench run and see if it suffices.

Luigi

Sadly, I always get "CUDAPm1.cu(2163) : cufftSafeCall() CUFFT error 6: CUFFT_EXEC_FAILED" with r between 1 and 5 and Threads=128 or 256.

Hints?

Luigi

[owftheevil]

Does it always fail at the same place in the test?

Also try putting

cutilSafeThreadSync();

after the cufft call on line 2161 and after the square call on 2162. That will at least tell us what is failing.

[ET_]

Originally Posted by owftheevil

Does it always fail at the same place in the test?

Also try putting

cutilSafeThreadSync();

after the cufft call on line 2161 and after the square call on 2162. That will at least tell us what is failing.

Yes, it always fails at the same place.

Added the line in the 2 places you asked. A new result:

CUDAPm1.cu(2165) : cufftSafeCall() CUFFT error 6: CUFFT_EXEC_FAILED

Added a new sync after line 2165: same error.

Luigi

[owftheevil]

Sorry, I jumped too quickly on the safecall stuff. More is needed. Let me think a bit.

[ET_]

Originally Posted by owftheevil

Sorry, I jumped too quickly on the safecall stuff. More is needed. Let me think a bit.

No hurry. I'm actually playing with Threads=128 and the program is working: I just tried to squeeze some more juice from it.

I'll be quietly waiting for your thoughts, thank you.

Luigi

[owftheevil]

Could you try this little snippet after the square call on 2162?

cudaThreadSynchronize();
{
  cudaError_t error = cudaGetLastError();
  if(error != cudaSuccess)
  {
    printf("CUDA error: %s\n", cudaGetErrorString(error));
    exit(2);
  }
}

[ET_]

Originally Posted by owftheevil

Could you try this little snippet after the square call on 2162?

cudaThreadSynchronize();
{
  cudaError_t error = cudaGetLastError();
  if(error != cudaSuccess)
  {
    printf("CUDA error: %s\n", cudaGetErrorString(error));
    exit(2);
  }
}

The error is:

CUDA error: too many resources requested for launch

while the environment is:

------- DEVICE 0 -------
name                GeForce GTX 580
Compatibility       2.0
clockRate (MHz)     1594
memClockRate (MHz)  2025
totalGlobalMem      1610285056
totalConstMem       65536
l2CacheSize         786432
sharedMemPerBlock   49152
regsPerBlock        32768
warpSize            32
memPitch            2147483647
maxThreadsPerBlock  1024
maxThreadsPerMP     1536
multiProcessorCount 16
maxThreadsDim[3]    1024,1024,64
maxGridSize[3]      65535,65535,65535
textureAlignment    512
deviceOverlap       1

HTH... thanks.

Luigi

[ET_]

Sorry for the delay... I was dining.

[owftheevil]

Thanks for getting back with that. The only thing I can think of right now is that somehow, either t2 or the threads array have messed up values. I'll look at it over the weekend and get back on Monday.

[ET_]

Originally Posted by owftheevil

Thanks for getting back with that. The only thing I can think of right now is that somehow, either t2 or the threads array have messed up values. I'll look at it over the weekend and get back on Monday.

Thanks

I add that I am using Linux_64, driver 304.88.

CUDA version info:

CUDA version info
  binary compiled for CUDA  4.10
  CUDA runtime version      4.10
  CUDA driver version       5.0

Luigi