mersenneforum.org  

Go Back   mersenneforum.org > Search Forums

Showing results 1 to 25 of 28
Search took 0.01 seconds.
Search: Posts Made By: jrk
Forum: Operation Kibibit 2011-10-12, 20:59
Replies: 200
Views: 53,150
Posted By jrk
For comparison, here are the two previous...

For comparison, here are the two previous polynomials that were reported for the same stage1 hit:
# norm 6.990167e-17 alpha -10.811225 e 6.655e-17 rroots 4
skew: 23705501.95
c0: ...
Forum: Operation Kibibit 2011-09-04, 22:01
Replies: 200
Views: 53,150
Posted By jrk
Yes. That will work. Thanks for the help.

Yes. That will work.

Thanks for the help.
Forum: Operation Kibibit 2011-09-04, 17:57
Replies: 200
Views: 53,150
Posted By jrk
I am going to continue my run for a while longer,...

I am going to continue my run for a while longer, btw.


That's unexpected. My run had about 0.2% duplication. Are you positive you didn't re-run any of the a6 ranges?


Yes. The hardest part...
Forum: Operation Kibibit 2011-08-30, 16:56
Replies: 200
Views: 53,150
Posted By jrk
Note that there is a per-coefficient time...

Note that there is a per-coefficient time deadline (currently 3200 secs for rsa768). If the search for some coefficient is still going when the time limit hits, it will be stopped and the next...
Forum: Operation Kibibit 2011-08-25, 19:12
Replies: 200
Views: 53,150
Posted By jrk
The randomization should be balanced so that, we...

The randomization should be balanced so that, we could expect to get several million hits total for rsa768 if every 8-digit a6 was tested once. If that was not sufficient to find a good polynomial...
Forum: Operation Kibibit 2011-08-24, 19:22
Replies: 200
Views: 53,150
Posted By jrk
My apologies. And thanks for continuing. The...

My apologies. And thanks for continuing. The updates will start slowing down.

BTW, are you still running on the same set of a6 coefficients each time? The randomization will ensure that a few runs...
Forum: Operation Kibibit 2011-08-24, 17:14
Replies: 200
Views: 53,150
Posted By jrk
And now 630. CPU usage should now hopefully...

And now 630. CPU usage should now hopefully always be negligible.
Forum: Operation Kibibit 2011-08-24, 06:58
Replies: 200
Views: 53,150
Posted By jrk
Newest version of the branch is 626. I made...

Newest version of the branch is 626.

I made an attempt to further reduce CPU overhead, in case it is too high when using the fastest cards.
Forum: Operation Kibibit 2011-08-20, 16:05
Replies: 200
Views: 53,150
Posted By jrk
The most recent revision is 623. When you...

The most recent revision is 623.

When you start running it, I suggest picking a coeff range that starts from at least 7 digits. 8 or 9 digits may be preferable. The internal parameters used for...
Forum: Operation Kibibit 2011-08-19, 18:16
Replies: 200
Views: 53,150
Posted By jrk
Jeff, if you are still running the code, the...

Jeff, if you are still running the code, the latest SVN (623) might be a (tiny) bit better. Thanks for all the help.
Forum: Operation Kibibit 2011-08-19, 07:14
Replies: 200
Views: 53,150
Posted By jrk
But, the discrepancy was apparent only in the...

But, the discrepancy was apparent only in the deg5 root sieve. I thought that deg6 was OK.
Forum: Operation Kibibit 2011-08-19, 06:56
Replies: 200
Views: 53,150
Posted By jrk
It is much the same as the CPU code, except that...

It is much the same as the CPU code, except that the CPU code uses a hashtable to find collisions among the arithmetic progressions, where this GPU code sorts them instead. This really only helps...
Forum: Operation Kibibit 2011-08-18, 11:16
Replies: 200
Views: 53,150
Posted By jrk
Thanks for the testing. If you like, please...

Thanks for the testing.

If you like, please try the newest svn (620). Now, things are run in larger batches, which might boost throughput a little bit. CPU usage should be closer to 0%. Also the...
Forum: Operation Kibibit 2011-08-18, 02:24
Replies: 200
Views: 53,150
Posted By jrk
Ok, thanks. That's not nearly as bad as I was...

Ok, thanks. That's not nearly as bad as I was afraid it would be.
Forum: Operation Kibibit 2011-08-18, 01:00
Replies: 200
Views: 53,150
Posted By jrk
It's good to see that the GPU was faster than CPU...

It's good to see that the GPU was faster than CPU overall but I thought it would be a bit better than that.

The kernels in this new method run so quickly that I'm afraid that the overhead of...
Forum: Operation Kibibit 2011-08-18, 00:22
Replies: 200
Views: 53,150
Posted By jrk
Jeff, can you tell me how much proportion of CPU...

Jeff, can you tell me how much proportion of CPU time is used when you run the GPU program?
Forum: Operation Kibibit 2011-08-16, 20:34
Replies: 200
Views: 53,150
Posted By jrk
Anything newer than my 5-year-old 8800 GTS : ) ...

Anything newer than my 5-year-old 8800 GTS : )

Note that, if you do decide to try it, make sure you have the latest version of the branch, because some bugs have been fixed since the earlier...
Forum: Operation Kibibit 2011-08-11, 18:54
Replies: 200
Views: 53,150
Posted By jrk
A quick benchmark with RSA768 and A6 ~1e7...

A quick benchmark with RSA768 and A6 ~1e7 suggests that the new code is about 100 times as fast as the original CUDA version on my system. I suspect that the CUDA version might be a bit faster than...
Forum: Operation Kibibit 2011-08-10, 08:10
Replies: 200
Views: 53,150
Posted By jrk
yes... I have GPU code that, when running on...

yes...

I have GPU code that, when running on an (ancient) 8800 GTS, can find RSA768 polys at about the same speed as the CPU code does on one core of a (similarly ancient) C2D. No idea yet how it...
Forum: Operation Kibibit 2011-08-07, 07:23
Replies: 200
Views: 53,150
Posted By jrk
I've noticed that polys with similar murphy score...

I've noticed that polys with similar murphy score but greatly different skew might have differences in yield, with the one with greater skew being the better one. But I don't have a whole lot of...
Forum: Operation Kibibit 2011-04-28, 06:17
Replies: 200
Views: 53,150
Posted By jrk
Please try svn 564 now and see if you continue...

Please try svn 564 now and see if you continue getting the errors. Thanks.
Forum: Operation Kibibit 2011-04-28, 00:45
Replies: 200
Views: 53,150
Posted By jrk
Thanks. I'll look at it.

Thanks. I'll look at it.
Forum: Operation Kibibit 2011-04-25, 02:47
Replies: 200
Views: 53,150
Posted By jrk
Post SVN revisions between 1.47 and 1.48 would have...

SVN revisions between 1.47 and 1.48 would have all said v1.48 in the logs, even though they weren't final. The final 1.48 was released on Jan 8th. Does that make sense? Sorry for the confusion.

(I...
Forum: Operation Kibibit 2011-04-25, 01:14
Replies: 200
Views: 53,150
Posted By jrk
From that output it looks like you're using an...

From that output it looks like you're using an old SVN. Please use 1.48 or the latest SVN revision and see if you still get "crap" and post here if you do. Thanks.
Forum: Operation Kibibit 2011-04-24, 20:18
Replies: 200
Views: 53,150
Posted By jrk
There should have been a "poly..." line printed...

There should have been a "poly..." line printed right before that "crap". Can you find it and print it here? Thanks.
Showing results 1 to 25 of 28

 
All times are UTC. The time now is 01:53.


Sun Feb 5 01:53:07 UTC 2023 up 170 days, 23:21, 1 user, load averages: 0.54, 0.70, 0.77

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2023, 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.

≠ ± ∓ ÷ × · − √ ‰ ⊗ ⊕ ⊖ ⊘ ⊙ ≤ ≥ ≦ ≧ ≨ ≩ ≺ ≻ ≼ ≽ ⊏ ⊐ ⊑ ⊒ ² ³ °
∠ ∟ ° ≅ ~ ‖ ⟂ ⫛
≡ ≜ ≈ ∝ ∞ ≪ ≫ ⌊⌋ ⌈⌉ ∘ ∏ ∐ ∑ ∧ ∨ ∩ ∪ ⨀ ⊕ ⊗ 𝖕 𝖖 𝖗 ⊲ ⊳
∅ ∖ ∁ ↦ ↣ ∩ ∪ ⊆ ⊂ ⊄ ⊊ ⊇ ⊃ ⊅ ⊋ ⊖ ∈ ∉ ∋ ∌ ℕ ℤ ℚ ℝ ℂ ℵ ℶ ℷ ℸ 𝓟
¬ ∨ ∧ ⊕ → ← ⇒ ⇐ ⇔ ∀ ∃ ∄ ∴ ∵ ⊤ ⊥ ⊢ ⊨ ⫤ ⊣ … ⋯ ⋮ ⋰ ⋱
∫ ∬ ∭ ∮ ∯ ∰ ∇ ∆ δ ∂ ℱ ℒ ℓ
𝛢𝛼 𝛣𝛽 𝛤𝛾 𝛥𝛿 𝛦𝜀𝜖 𝛧𝜁 𝛨𝜂 𝛩𝜃𝜗 𝛪𝜄 𝛫𝜅 𝛬𝜆 𝛭𝜇 𝛮𝜈 𝛯𝜉 𝛰𝜊 𝛱𝜋 𝛲𝜌 𝛴𝜎𝜍 𝛵𝜏 𝛶𝜐 𝛷𝜙𝜑 𝛸𝜒 𝛹𝜓 𝛺𝜔