LLRnet/PRPnet rally April 4th-11th

[Rick Reynolds]

Originally Posted by Ralf Recker

The GTX 470 runs at stock clocks: I'm currently testing my tpsieve optimizations on it. I initially developed and tested them on a GTX 460 (GF 104 chip, Compute Capability 2.1) and now trying out a few changes on a GF 100 chip (Compute Capability 2.0) based GTX 470 card. Clocks are the standard: 607 MHz Core, 1215 MHz Shaders (2x core clocks, as usual) and 1674 MHz memory.

So when are you going to release the tpsieve app :) that would be very cool.

[Neo]

Originally Posted by Rick Reynolds

So when are you going to release the tpsieve app :) that would be very cool.

Rick,

It's released. :) It works awesome on Nvidia cards.
See the primesearchteam website; sieving thread.

Neo

[mdettweiler]

Originally Posted by Rick Reynolds

Don't suppose there's any 32 bit cuda enabled llr versions running around for this rally or is that even possible here?

Rick

To further elaborate on what Ralf explained: we do have a CUDA LLR program, though for numbers of this size it's not much use (as it's only about the speed of one CPU core on a modern GPU such as a GTX 460). It is somewhat faster for much larger numbers, i.e. PSP and SoB, but not for what NPLB's testing right now. As a point of reference, the numbers we're doing in this rally are in the vicnity of n=850K; Gary has a GPU that he's currently running on LLRnet port 3500, for another one of our efforts in the vicinity of n=1.3M, and it's about twice the speed of a CPU core.

I do have a binary on hand for llrCUDA, but it's for 64-bit Linux. What operating system are you looking to run this on? There have been some 64-bit Windows binaries posted in the forum (see here for the latest), but I don't believe I've seen any for 32-bit Windows. For 32-bit Linux, however, it should be pretty easy to compile the latest source as long as you have the CUDA toolkit installed.

[Ralf Recker]

Originally Posted by Neo

Rick,

It's released. :) It works awesome on Nvidia cards.
See the primesearchteam website; sieving thread.

Neo

Ken's app has been released but not my modifications .

[Rick Reynolds]

Originally Posted by mdettweiler

To further elaborate on what Ralf explained: we do have a CUDA LLR program, though for numbers of this size it's not much use (as it's only about the speed of one CPU core on a modern GPU such as a GTX 460). It is somewhat faster for much larger numbers, i.e. PSP and SoB, but not for what NPLB's testing right now. As a point of reference, the numbers we're doing in this rally are in the vicnity of n=850K; Gary has a GPU that he's currently running on LLRnet port 3500, for another one of our efforts in the vicinity of n=1.3M, and it's about twice the speed of a CPU core.

I do have a binary on hand for llrCUDA, but it's for 64-bit Linux. What operating system are you looking to run this on? There have been some 64-bit Windows binaries posted in the forum (see here for the latest), but I don't believe I've seen any for 32-bit Windows. For 32-bit Linux, however, it should be pretty easy to compile the latest source as long as you have the CUDA toolkit installed.

Hi Max, I'm running Windows XP 32 bit right now. I've read the posts on the 64 bit testing being done over on prpnet but so far they don't have a 32 bit version planned. I know I should convert to linux but just haven't as of yet. Steve Martin is willing to help me once I start the process. Was just hoping we'd get a 32 bit version I could run and leave my pc os as is.

thanks much, Rick

[Ralf Recker]

I found my first prime at NPLB . I just checked Lennart's last prime to make a correct entry:

"PSieve, Srsieve, NPLB, LLR"

Now I have to wait for the PG admins to sent me my password data (I selected autoreporting over @PG...). I requested a password change manually but I got no e-mail from the Top-5000.

[gd_barnes]

Originally Posted by Ralf Recker

I found my first prime at NPLB . I just checked Lennart's last prime to make a correct entry:

"PSieve, Srsieve, NPLB, LLR"

Now I have to wait for the PG admins to sent me my password data (I selected autoreporting over @PG...). I requested a password change manually but I got no e-mail from the Top-5000.

Congrats Ralf! That is correct on your entry. It may take the top-5000 admin(s) several hours or a day on the password change.

[Rick Reynolds]

Originally Posted by Ralf Recker

I found my first prime at NPLB . I just checked Lennart's last prime to make a correct entry:

"PSieve, Srsieve, NPLB, LLR"

Now I have to wait for the PG admins to sent me my password data (I selected autoreporting over @PG...). I requested a password change manually but I got no e-mail from the Top-5000.

Congrats Ralf!

Rick

[Ralf Recker]

Originally Posted by gd_barnes

Congrats Ralf! That is correct on your entry. It may take the top-5000 admin(s) several hours or a day on the password change.

Thanks for the infos. As I suspected: The e-mail address used was not mine but belongs to one of the PG admins. John forwarded me the necessary infos and I changed the e-mail address and password, created a proof code and submitted the prime number. In addition to that I switched off auto-reporting over at PrimeGrid.

Originally Posted by Rick Reynolds

Congrats Ralf!

Rick

Thank you!

[Neo]

It looks as if the primes finally got going on Port 9000.....

Congrats Lennart, Ralf, and Sashixi!

[Vato]

Originally Posted by Ralf Recker

In addition to that I switched off auto-reporting over at PrimeGrid.

FYI, now that you have the top5000 set up to your email/password, you can turn PG auto-reporting back on. PG can still auto-submit primes that have an existing proof-code, but if you found a prime at PG that needed a new proof-code, you'd have to create that manually.