mersenneforum.org  

Go Back   mersenneforum.org > Search Forums

Showing results 1 to 25 of 411
Search took 0.05 seconds.
Search: Posts Made By: Karl M Johnson
Forum: GPU Computing 2017-08-11, 08:40
Replies: 3
Views: 1,568
Posted By Karl M Johnson
From what I've pieced together, the...

From what I've pieced together, the aforementioned driver boosts half-precision (FP16) performance for Pascal Titan X, but that's not used much in CUDA programs for GIMPS, right?
Forum: GPU Computing 2017-08-01, 12:40
Replies: 3
Views: 1,568
Posted By Karl M Johnson
Here's the original blog post by Nvidia...

Here's the original blog post by Nvidia (https://blogs.nvidia.com/blog/2017/07/31/titan-xp-drivers-new-levels-of-performance-for-creative-professionals/).
Note them saying "...applications like...
Forum: GPU Computing 2017-08-01, 08:56
Replies: 3
Views: 1,568
Posted By Karl M Johnson
Nvidia unlocking Titan X (P) through drivers

Source (https://www.techpowerup.com/235701/nvidia-unlocks-certain-professional-features-for-titan-xp-through-driver-update)
Can someone test this with CuLu?
Forum: GPU Computing 2017-05-13, 07:58
Replies: 2
Views: 1,215
Posted By Karl M Johnson
It is highly unlikely we will see a Volta Titan...

It is highly unlikely we will see a Volta Titan with full FP64 performance, like the Kepler Titans, so don't expect substantial LL speed gains from non-Tesla GPUs.
Forum: GPU Computing 2016-10-03, 02:44
Replies: 3,434
Views: 310,777
Posted By Karl M Johnson
Nailed it! Turns out, it was not related to...

Nailed it!
Turns out, it was not related to mfaktc, but which EOL worktodo.txt used.
I've realised this only when I've noticed a 'Macintosh' EOL being used n Notepad++ and I'm not sure how this...
Forum: GPU Computing 2016-10-02, 16:35
Replies: 3,434
Views: 310,777
Posted By Karl M Johnson
Found my first bug for mfaktc by getting...

Found my first bug for mfaktc by getting overnight TF work.
It would seem that 128 lines in worktodo file cause some kind of overflow, and mfaktc complains about invalid format in such a case (and...
Forum: GPU Computing 2016-10-02, 14:07
Replies: 3,434
Views: 310,777
Posted By Karl M Johnson
Now I know there's a specialised form...

Now I know there's a specialised form (http://www.mersenne.ca/mfaktc.php#) for this sort of data submission.
I've sent the standardised data, with comment pointing to your request, James.

My...
Forum: GPU Computing 2016-10-02, 12:09
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
Hello, Jerry! Thanks for the new binary, I...

Hello, Jerry!

Thanks for the new binary, I can confirm that on a GTX 1080 it's a bit faster than the old v8.0 RC binary & the exponents match.:smile:
II did a quick test on M1000003, and the...
Forum: GPU Computing 2016-10-02, 07:16
Replies: 3,434
Views: 310,777
Posted By Karl M Johnson
Cheers! Here are some freshly submitted...

Cheers!

Here are some freshly submitted results on a GTX 1080 @ 2ish GHz:

got assignment: exp=137990929 bit_min=70 bit_max=72 (5.20 GHz-days)
Starting trial factoring M137990929 from 2^70 to...
Forum: GPU Computing 2016-10-01, 13:28
Replies: 3,434
Views: 310,777
Posted By Karl M Johnson
Lovely news, Oliver! Now, would someone...

Lovely news, Oliver!

Now, would someone kindly compile a x64 Windows binary? :smile:

Here are the new DLLs (https://www.dropbox.com/s/fy8o5d6jnmsgls6/cudalib.7z?dl=1).
Forum: Hardware 2016-09-17, 14:39
Replies: 313
Views: 33,140
Posted By Karl M Johnson
Intel also wants to put a FPGAs inside future...

Intel also wants to put a FPGAs inside future Xeons, which could result in an extra speed boost if properly utilised.
Forum: GPU Computing 2016-08-29, 11:39
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
I can confirm that the binary Robert_JD...

I can confirm that the binary Robert_JD (http://mersenneforum.org/member.php?u=10788) kindly provided indeed works and gives correct results on stable hardware.
Forum: GPU Computing 2016-08-27, 14:21
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
2.516 ms/iter for M76 makes Titan X faster than...

2.516 ms/iter for M76 makes Titan X faster than Titan Black :smile:
And since the residues match, well, congrats, money well spent!
If only it had 1/2 fp64 performance of fp32, instead of 1/32.
Forum: GPU Computing 2016-08-27, 10:09
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
Hiya Lorenzo! Here's the benchmark data: ...

Hiya Lorenzo!

Here's the benchmark data:

Using threads: square 256, splice 256.
Starting M332220523 fft length = 19208K
| Date Time | Test Num Iter Residue | ...
Forum: GPU Computing 2016-08-27, 05:52
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
It would seem that despite the severely limited...

It would seem that despite the severely limited fp64 performance of high-end Pascal GPUs like 1080 and Titan X, they're a match for the Kepler Titans after all.
I suspect it has something to do with...
Forum: GPU Computing 2016-08-27, 03:52
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
Cheers! Some benchmark data with the new...

Cheers!

Some benchmark data with the new version below.
It looks consistent with airsquirrels' timings (3.65ms/iter on same GPU and Mersenne candidate), but is a bit faster due to factory...
Forum: GPU Computing 2016-08-25, 10:33
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
Thanks, but I'm on Windows. Are you sure that...

Thanks, but I'm on Windows.
Are you sure that the Titan X(P) is producing correct results?
As far as I've read, Nvidia didn't want us to have another non-Tesla GPU with Teslas' fp64 performance, so...
Forum: GPU Computing 2016-08-05, 06:11
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
Nailed it! Everything is prime. 1080 isn't...

Nailed it!
Everything is prime.
1080 isn't meant for fp64 anyways, it should excel at TF when mfaktc/CUDA gets a fix.

As for the benchmarks, I wouldn't even trust the slow benchmarks, not to...
Forum: GPU Computing 2016-08-05, 05:45
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
CuFFT seems broken, though, hence the wrong FFT...

CuFFT seems broken, though, hence the wrong FFT size selected in my first test.
All the timings are 0 :confused2:
Forum: GPU Computing 2016-08-05, 04:37
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
Hmm, despite the residue being wrong, it still...

Hmm, despite the residue being wrong, it still works:
Starting M756839 fft length = 42K
| Date Time | Test Num Iter Residue | FFT Error ms/It Time | ...
Forum: GPU Computing 2016-08-05, 03:06
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
4.2 | Date Time | Test Num Iter ...

4.2
| Date Time | Test Num Iter Residue | FFT Error ms/It Time | ETA Done |
| Aug 05 06:15:11 | M57885161 100000 0x0000000000000000 ...
Forum: GPU Computing 2016-08-04, 13:40
Replies: 2,832
Views: 236,026
Posted By Karl M Johnson
Here's some: | Aug 04 16:19:43 | M57885161 ...

Here's some:
| Aug 04 16:19:43 | M57885161 900000 0x0000000000000000 | 3920K 0.00000 0.0468 4.68s | 43:01 1.55%
| Aug 04 16:19:48 | M57885161 1000000 ...
Forum: Hardware 2015-10-26, 07:52
Replies: 17
Views: 4,424
Posted By Karl M Johnson
Welcome to the club.

Welcome to the club.
Forum: GPU Computing 2015-10-12, 16:37
Replies: 15
Views: 2,070
Posted By Karl M Johnson
Okay, I thought the hardware is unstable, it may...

Okay, I thought the hardware is unstable, it may lose overclocking potential, but as it turns out, it's not entirely related to that.
So far I've gotten no bad residues on C4.2 binaries, but this...
Forum: Software 2015-10-12, 15:56
Replies: 12
Views: 1,635
Posted By Karl M Johnson
Just an overclocked 5960X, nothing special.

Just an overclocked 5960X, nothing special.
Showing results 1 to 25 of 411

 
All times are UTC. The time now is 14:35.

Tue Dec 1 14:35:42 UTC 2020 up 82 days, 11:46, 2 users, load averages: 2.05, 2.28, 2.12

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.