(To Mersenne mailing list and known MacGIMPS users)

I could use some benchmark information on Mac computers running MacGIMPS to
help calibrate the kind of work to allocate.  Anyone who is running
MacGIMPS on a PPC Mac who would be willing to run a timing test for several
hours, please read on.  Anyone else, thanks for listening.

To run a timing test:

Please dedicate several hours to this test when your Mac is not doing
anything else -- say overnight.

Restart with no extensions running, and start up no other software.

Set your work-in-progress files aside, and start a MacGIMPS test on the
exponent 7,654,321  (that's not a magic number, I just want all timings to
be comparable and they depend on the exponent. )

Configure the following MacGIMPS options:
-  No checkpoints (set checkpoints to a large number like 1,000,000)
-  Create a new log file and request logging every 200 iterations
-  Responsiveness 0  (fastest compute)
-  No scheduled responsiveness changes

Let the test run for several hours -- overnight would be great.

Then, mail the resulting log file to mailto:[EMAIL PROTECTED] --
attach it or cut&paste into the email.

IMPORTANT:  Include your CPU type (604, G3), clock rate, and the version
number of your copy of MacGIMPS in the email.  Just the model number of
your computer will do if you don't know the CPU and clock rate.  The test
is not useful without this information and if you forget to provide it I'll
ask you.

It's important that this test run undisturbed (no extensions and no other
software) and at priority 0 for consistency.  If you can't do that because
of other demands on your computer time, thanks for your interest but please
don't send in tests.

You can throw the checkpoint file and document file away -- only the timing
information in the log file is needed.

Thanks & regards,
   Richard
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm

Reply via email to