Hi,
CUDALucas.cu:
[code]
cudaSetDevice (device_number);
cutilSafeCall(cudaSetDeviceFlags (cudaDeviceBlockingSync));
cudaGetDeviceProperties (&g_dev, device_number);
// From Iain
if (g_dev.major == 1 && g_dev.minor < 3)
{
printf("A GPU with compute capability >= 1.3 is required for double precision arithmetic\n\n");
printf("See http://www.mersenne.ca/cudalucas.php for a list of cards\n\n");
exit (2);
}
[/code]
Compare to 1.3 is not enough.
We need compare to compile option.(--generate-code arch=compute_XY,code=sm_XY)

[QUOTE=msft;426628]Hi,
CUDALucas.cu:
[code]
cudaSetDevice (device_number);
cutilSafeCall(cudaSetDeviceFlags (cudaDeviceBlockingSync));
cudaGetDeviceProperties (&g_dev, device_number);
// From Iain
if (g_dev.major == 1 && g_dev.minor < 3)
{
printf("A GPU with compute capability >= 1.3 is required for double precision arithmetic\n\n");
printf("See http://www.mersenne.ca/cudalucas.php for a list of cards\n\n");
exit (2);
}
[/code]
Compare to 1.3 is not enough.
We need compare to compile option.(--generate-code arch=compute_XY,code=sm_XY)[/QUOTE]

Do you still support/qualify the code for CC 1.3?

Will CUDALucas warn me if I do not have enough free memory or picked too few threads for a given exponent size?

Thanks!

[QUOTE=TheJudger;426646]Do you still support/qualify the code for CC 1.3?[/QUOTE]
Could you explain that a little more in detail?

Figured it out . Thanks

[QUOTE=msft;426658]Could you explain that a little more in detail?[/QUOTE]

The question is if you still spend time developping/testing CUDALucas on compute capability 1.3. My(!) opinion is that they were good cards in the past but today the just use too much energy for their performance. It's up to the individual user of course.

Oliver

[QUOTE=TheJudger;426732]The question is if you still spend time developping/testing CUDALucas on compute capability 1.3. My(!) opinion is that they were good cards in the past but today the just use too much energy for their performance. It's up to the individual user of course.[/QUOTE]
Although I still run CUDALucas on a very reliable Tesla S1070, I would not object to dropping support for cc 1.3.

[QUOTE=TheJudger;426732]The question is if you still spend time developping/testing CUDALucas on compute capability 1.3. My(!) opinion is that they were good cards in the past but today the just use too much energy for their performance. It's up to the individual user of course.

Oliver[/QUOTE]
Understand.

Does the "BigCarry" option work on both 32-bit and 64-bit executable?

Has the source of the issue that necessitated the option in the first place been identified?

Thank you.

I keep getting the round off >0.35 errors once every 100K-200k iterations on a hundred million digit test I just started (a 332M exponent).

Before starting the test I run both FFT and threads benchmarks from 1024K to 32768K FFT lengths.

Device: GeForce GTX TITAN
CUDALucas2.05.1-CUDA6.5-Windows-WIN32

The parameters CUDALucas picked for this test are:
18432K FFT
512 Threads
32 Splices

The BigCarry parameter is 0.

What would you recommend I do for this test? Restart it with a larger FFT size?

The Titan passed tests and did a few matching double-checks beforehand.

Thanks.

While running CUDALucas on a Windows 10 64-bit machine I keep getting the following warnings in the event log:

[b]Display driver nvlddmkm stopped responding and has successfully recovered.[/b]

The warnings do not seem to coincide with the round-off errors.

Is this anything to worry about?

Thanks

Edit: never mind this post, looks like I asked about this question two years ago (though for a different GPU and different version of Windows). LOL. Apparently this is nothing to worry about.