mersenneforum.org (https://www.mersenneforum.org/index.php)
-   PrimeNet (https://www.mersenneforum.org/forumdisplay.php?f=11)
-   -   Per iteration time (https://www.mersenneforum.org/showthread.php?t=16037)

 Jwb52z 2011-09-08 19:14

Per iteration time

I hope no one minds my making a new topic. I haven't participated in Prime 95 for nearly a decade back when the V4 was up and running. My current test is M54938133. My computer has an I7 Q720 Quad Core on Windows 7 64 bit. My question is very simple and this thread might be helpful for other users who are completely knew or haven't participated in years as a baseline. This number is currently doing a first time LL set to display every 20 thousand iterations and it reports that as running between 0.07 and 0.08 seconds for those 20 thousand. I'd like to know if that's a good iteration time for that many as I don't remember how to calculate the exact per iteration time for one iteration. Thank you. If Prime95 himself doesn't like this thread or someone else objects, feel free to delete it or move it or whatever you would like to do.

 firejuggler 2011-09-08 19:26

The 0.08 sec is by iteration. that mean you will have to run for about 50 days (0.08 *54938133 =4395050,64 seconds = 1220.84 hours=50.86 days) . I would say that is pretty quick.

 Jwb52z 2011-09-08 19:31

Thank you for the fast response. It's going to take a bit of time for me to get all this back in my brain the way it should be.

 ATH 2011-09-08 19:32

Those are the average iteration times over 20,000 iterations. If you want more accuracy add these 2 lines to prime.txt:

PercentPrecision=6
TimingOutput=4

From undoc.txt:
[QUOTE]You can adjust how many decimal digits of precision are displayed
in lines output to the screen. In prime.txt enter a value between 0 and 6:
PercentPrecision=n

You can alter the way the program outputs timings. In prime.txt set:
TimingOutput=n
Where n=1 for seconds with 3 digits of precision, n=2 for milliseconds
with 1 digit of precision, n=3 for milliseconds with 2 digits of precision,
n=4 for milliseconds with 3 digit of precision.[/QUOTE]

You can check other benchmarks here: [URL="http://www.mersenne.org/report_benchmarks/"]http://www.mersenne.org/report_benchmarks/[/URL]

Your processor at 55M exponent is about 70-75 ms in the chart (0.07-0.075 sec).

 lycorn 2011-09-08 20:42

Welcome back, Jwb52z!
It´s good to see old gimpsters (re)joining the ranks.
Old loves die hard, isn´t it?...

 Christenson 2011-09-09 02:45

Possibly this belongs under information and answers... oh well....

There are some new kids on the prime block you might want to know about in conjunction with openCL and CUDA GPUs: mfaktc, mfakto, and cudaLUCAS. All of them require GPUs. cudaLUCAS is moderately fast; mfaktc and mfakto run circles around CPUs at trial factoring, but still require a significant amount of CPU for part of their operation.

 Jwb52z 2011-09-09 04:06

Thank you Lycorn. As for what the chart says, I think for me that timing would only be if I never used my system because it's now ranging .08 to .09 per iteration.

 All times are UTC. The time now is 07:17.