+- +-
Seti is down again by Mike
09 Aug 2017, 10:02:44 am

Some considerations regarding OpenCL MultiBeam app tuning from algorithm view by Raistmer
11 Dec 2016, 06:30:56 am

Loading APU to the limit: performance considerations by Mike
05 Nov 2016, 06:49:26 am

Better sleep on Windows - new round by Raistmer
26 Aug 2016, 02:02:31 pm

MB/AP Bench Test Instruction by William
17 Jun 2016, 04:13:45 am

Author Topic: approx. how fast should AP units complete with the optimized app?  (Read 2117 times)

codeman05

  • Guest
Hey guys,

System is a Q6600 @ 3.0GHz running Debian x86.
I just switched one of my headless linux boxes over to AP using boinc 6.6.20 and the linux SSE3 AP application shown on the homepage.

I'm crunching my first AP units as we speak...but I'm impatient ;) and was wondering what you guys would expect the time frame for an average AP unit to complete on a rig similar to mine with this ap.


BOINC is currently estimating 35 hours (I was thinking something closer to 15hrs), but since I'm only a few percent into it I'm sure that is probably ncorrect


Thanks for all the work on this app!

Offline Josef W. Segur

  • Janitor o' the Board
  • Knight who says 'Ni!'
  • *****
  • Posts: 3112
Re: approx. how fast should AP units complete with the optimized app?
« Reply #1 on: 30 Apr 2009, 09:24:29 pm »
Hey guys,

System is a Q6600 @ 3.0GHz running Debian x86.
I just switched one of my headless linux boxes over to AP using boinc 6.6.20 and the linux SSE3 AP application shown on the homepage.

I'm crunching my first AP units as we speak...but I'm impatient ;) and was wondering what you guys would expect the time frame for an average AP unit to complete on a rig similar to mine with this ap.


BOINC is currently estimating 35 hours (I was thinking something closer to 15hrs), but since I'm only a few percent into it I'm sure that is probably ncorrect


Thanks for all the work on this app!

I'd agree with the 15 hour estimate. What you're probably seeing is the Duration Correction Factor causing an over estimate.

The Astropulse progress is quite linear, though there's a small sawtooth effect in 5.03 builds. Divide the time used so far by the fraction done to get a better full runtime estimate than BOINC will be able to provide.

The optimized code has a feature which can make the runtime even shorter for some AP WUs, once the limit of 30 repetitive pulses is reached the search code for those is bypassed. More often than not if that happens at all it is quite early but it can be later, the progress rate inflects upward whenever it happens.
                                                                             Joe

 

Welcome, Guest.
Please login or register.
 
 
 
Forgot your password?
Members
Total Members: 97
Latest: ToeBee
New This Month: 0
New This Week: 0
New Today: 0
Stats
Total Posts: 59551
Total Topics: 1671
Most Online Today: 113
Most Online Ever: 402
(15 Nov 2018, 02:33:34 pm)
Users Online
Members: 0
Guests: 80
Total: 80
Powered by EzPortal