Seti@Home optimized science apps and information
Optimized Seti@Home apps => Windows => Topic started by: Crunch3r on 10 Aug 2007, 08:07:27 pm
-
Howdy,
the thread title is kinda ironic.... while having a look at the "Users online" i can see a"dramatic" increase of users online here. .
Seems as if something is free.
Well it is of course, however i'd like to encourage you all to sign up here and post some comments for ALL apps you've downloaded here. No matter what OS.
In the past i've noticed that most of the information we got back here regarding the apps, came only from the pre-release testers ( which i might add, did a great job !!!). However, a bit more response from those that download the apps and run them IS welcome.
Especially, now that we've added plattforms that haven't been (officially) supported before, like Windows on IA64 or Linux IA64/ALPHA or HPUX, SOLARIS etc.
It is in YOUR interest to sign up here and actively participate on further development of the optimized apps.
So take a minute, sign up here and leave us a note of your experience with optimized apps.
Thanks.
-
I did download the optimized apps for C2D/Windows, but I'm now trying to play with the source code myself in WinXP/CUDA.
I have tried to compile the standard nightly source in Visual Studio Express (I'm not entirely sure about the Intel compilers/files for Windows, and I plan to work on this for a lot longer than 30 days so the trial is useless to me), but that fails at setibioncdb.
Despite trying the guide by Urs on how to get the code to compile, I'm using code from a year after the guide was written and it seems to have changed. :)
I have downloaded the new apps, but it's too hot to run seti much - my CPU gets hotter than I'd like, and it makes the computer kick out too much heat. :)
-
...
Despite trying the guide by Urs on how to get the code to compile, I'm using code from a year after the guide was written and it seems to have changed. :)
...
Hi,
like i stated in that file, it was for last years seti@home beta version.
Now the devs at seti@home have made a new version, including many more changes to that old code and adding new codeparts. So, of course that file is now part of history.
Many of these adaptions to the VS8 compiler are now already in the code and therefore are obsolete. Other adaptions will have to be made to make it compile with VS8 again.
Maybe its time for a reworked guide based on the actual source code.
Until, take a look into the optimized sources (http://lunatics.at/windows/optimized-sources.0.html) thread by seti_britta. It includes many, many hints how to work things out.
-
These are the closest AR's I could find for comparison of Stock app and 2.4 on Pent M.
CPU time 15195.25
setiathome_enhanced 5.27 DevC++/MinGW
WU true angle range is : 0.405720
Flopcounter: 16411814961316.320000
Spike count: 0
Pulse count: 0
Triplet count: 0
Gaussian count: 0
Granted credit 54.1526303700828
CPU time 11540.171875
Version: Windows SSE2 32-bit based on seti V5.15 'Ni!'
Rev: (R-2.4|xB|FFT:IPP_SSE2|Ben-Joe)
CPUID: 'Intel PM Pentium M (Dothan)
True angle range: 0.405888
Spikes Pulses Triplets Gaussians Flops
3 0 0 0 16409086762502
Granted credit 54.1436311031597
The decrease in crunch time is 11540.171875 / 15195.25 * 100 = 75.95%
Andy
-
And for my C2D
Not quite same AR
CPU time 8749.125
<core_client_version>5.10.7</core_client_version>
<![CDATA[
<stderr_txt>
setiathome_enhanced 5.27 DevC++/MinGW
Work Unit Info:
WU true angle range is : 0.405730
Flopcounter: 16413617069912.094000
Spike count: 2
Pulse count: 0
Triplet count: 0
Gaussian count: 0
Granted credit 54.1555901686006
* * * * * * * *
CPU time 6129
Version: Windows SSE3 32-bit based on seti V5.15 'Ni!'
Rev: (R-2.4|xT|FFT:IPP_SSE3|Ben-Joe)
CPUID: 'Intel C2D Core 2 Duo (Conroe)'
cpus: 1 cores: 2 threads: 1 cache: L1=32K L2=4096K L3=0K
features: mmx sse sse2 sse3 ssse3
speed: 2742 MHz -- read megs/sec: L1=10340, L2=8871, RAM=5757
Work Unit Info
True angle range: 0.405538
Spikes Pulses Triplets Gaussians Flops
3 0 0 2 16417320628577
Granted credit 54.1707914252839
The decrease in crunch time is 6129 / 8749.125 * 100 = 70.05%
Andy
-
Thanks for the info. I assume your tests were run on Win XP ?
-
Correct assumption.