Message boards :
Number crunching :
new joking WU-size/type ps_new_11 and ps_new_13
Message board moderation
Author | Message |
---|---|
Send message Joined: 8 Jan 09 Posts: 58 Credit: 53,161,741 RAC: 0 |
MW is joking again ! The new WU-types ps_new_11 and ps_new_13 are done is less than 4s with ATI GPUs .... |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
One of the new people working with us (Matt) started his first new searches and they had a bad size. I've stopped them, so hopefully the next batch will be ok :) |
Send message Joined: 12 Dec 08 Posts: 56 Credit: 269,889,439 RAC: 0 |
Thanks...with 4 seconds on a 4850, there is no way the server could keep up... Click to help Seti City. |
Send message Joined: 8 Jan 09 Posts: 58 Credit: 53,161,741 RAC: 0 |
the new ps_new_15 got the same problem - done in >4s on a slow 4830 |
Send message Joined: 8 Jan 09 Posts: 58 Credit: 53,161,741 RAC: 0 |
ps_newtest_2... same problem - done in >5s on a 4830 |
Send message Joined: 24 Dec 07 Posts: 1947 Credit: 240,884,648 RAC: 0 |
But they gave good credit....shame I didn't get more of them! |
Send message Joined: 2 Jan 08 Posts: 122 Credit: 69,480,257 RAC: 1,364 |
But they gave good credit....shame I didn't get more of them! They are still there GG, as I am still getting them. So far I have seen ps_new_9_1 runs for 197 to 224 sec on CPU ps_new_11_1 runs for 185 sec on CPU ps_new_11_2 runs for 187 sec on CPU ps_new_11_2 runs for 3 to 4 sec on GPU ps_new_13_2 runs for 200 sec on CPU ps_new_13_2 runs for 1 sec on GPU ps_new_15_2 runs for 112 sec on CPU ps_new_15_2 runs for 2 sec on GPU ps_newtest_2 runs for 225 sec on CPU ps_newtest_2 runs for 3 sec on GPU ps_new_16_2 runs for 31 to 41 sec on a GPU The CPU is an AMD Opteron 285 @2.6GHz and the GPU is an ATI 4870 at standard clock. |
Send message Joined: 29 Aug 07 Posts: 327 Credit: 116,463,193 RAC: 0 |
I just burned thru 24 of the ps_new_13 WUs in less than a minute. ;) And the ps_newtest_2 WUs don't take any longer than the others. Calm Chaos Forum...Join Calm Chaos Now |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
It looks like these are older 1-stream workunits so they're going to be quicker. Hopefully there aren't too many of these searches to run :D At least it looks like the server is holding up. |
Send message Joined: 26 Jul 08 Posts: 627 Credit: 94,940,203 RAC: 0 |
It looks like these are older 1-stream workunits so they're going to be quicker. Hopefully there aren't too many of these searches to run :D At least it looks like the server is holding up. But barely, we see a lot of time outs lately. By the way, the mentioned searches (and the ps_new_15_2) also have very small integral sizes (only 40 nu_steps!, the other dimensions are smaller too). That is also a reason the server is hammered that much, they take just two seconds or so. |
Send message Joined: 24 Dec 07 Posts: 1947 Credit: 240,884,648 RAC: 0 |
Looks like things are improving with the longer wu's. |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
Just got 5 of these for the first time. Gives a 30 min completion time and finishes in 100 secs on my P4. Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 22 Mar 08 Posts: 65 Credit: 15,715,071 RAC: 0 |
YEAY for desktop supercomputers! lol Hey, anyone remember the first tasks on stock vanilla CPU app? Before the first round of optimizations...My how far we've come.. Can't wait to turn some cuda loose on MW's and see how they do compared to 4850. Will be interesting to run MW on one box with 4850+260 |
Send message Joined: 21 Feb 09 Posts: 180 Credit: 27,806,824 RAC: 0 |
YEAY for desktop supercomputers! lol One of my machines run using a 4850+280 for a while :) One on MW, one on GPUGrid |
Send message Joined: 22 Mar 08 Posts: 65 Credit: 15,715,071 RAC: 0 |
YEAY for desktop supercomputers! lol Yea, running my q6600 like that till it blowd up..lol Thing I am waiting for is being able to crunch one project on both cards for compare. Right now ~5k on seti and 20k here on a 4850 seems bad comparison. :) |
©2024 Astroinformatics Group