Prime 95?
 

Not sure where this goes but I've been running Prime95 on my Intel Core2 Quad CPU @ 2.4 Ghz pretty much full time for over two weeks and the four cores aren't even to 20% yet.

Seems like it's going pretty slow and I was wondering if anyone had any comments? At this rate it's going to take forever to get any work units back to PrimeNet and I might give up! :-)

Thanks,
-TAC

That's normal. Unfortunately, running one LL test takes forever -- even if it can do one iteration in a 20th of a second (50 ms), there's still millions and millions of iterations to be done.

A few points to clarify:

1. What type of work are you doing (and on what exponents?)
2. How many hours per day is the computer running?
3. Are you running other applications that might use the CPU a lot (e.g. games, video processing?

Depending on the answers, the situation reported may or not be expectable.

And of course, welcome to the search!

[QUOTE=lycorn;331016]1. What type of work are you doing (and on what exponents?)[/QUOTE]
Are you doing 1 test per core?

just started the darn thing.
Worker Windows
number of windows to run 4, priority 1, Worker number: all workers, type of work to get: Whatever make the most sense, CPU affinity: Smart assignement..... blah, blah, blah.

[QUOTE=crites;331009]... At this rate it's going to take forever to get any work units back to PrimeNet and I might give up! :-)
[/QUOTE]
You have to be patient. Core2 Quad is not the fastest runner in the bunch these days. "At this rate it's going to take forever" - no obviously not [I]forever[/I], but approximately 8 more weeks based on your own information. Use the "Status" button (3rd line) under the "Test" menu and you will find out the date when the jobs will finish.

Think about it this way: In two months, you will have finished four work units and will have four non-zero chance to win (Just how do you think C.Cooper found his prime? exactly like everyone else). If you quit now, your chances obviously will be zero.

Just in case it's not already clear from the other excellent answers here, note that there is also plenty of [U]factoring work[/U] required for the project, and a work unit of this type takes [I]much[/I] less time than an LL test, generally just hours or a few days at the most. It won't give you any chance of being the discoverer a new prime, but you will turn in valuable work which, if you discover a factor, eliminates the candidate so that people don't have to waste lots of work on an LL test of that number. [U]P-1 factoring[/U] is especially in demand if you have at least 300 megabytes (preferably 500 megabytes) of memory to spare. [U]Trial factoring[/U], on the other hand, doesn't require much memory, but note that GPUs are much faster at this than CPUs so if you don't have a GPU you might not be using your computing power optimally for the project if you select trial factoring.

So when your LL tests have finished, consider putting some of your cores onto factoring work if you're keen on turning in quick results.

@OP: As said, stay on the LL test. You are doing good. Thanks for joining, and welcome to the project.

One point to note on Core 2 Duos: They are not truly full quad core systems. I only run Core 1 and Core 3 on LL or DC assignments since a serious bottleneck will occur otherwise.

Example: Core 1 alone would get ~60ms per iterations while running Core 1 and Core 2 would yeild ~90ms iterations on both.

[QUOTE=bcp19;331335]One point to note on Core 2 Duos: They are not truly full quad core systems. I only run Core 1 and Core 3 on LL or DC assignments since a serious bottleneck will occur otherwise.

Example: Core 1 alone would get ~60ms per iterations while running Core 1 and Core 2 would yeild ~90ms iterations on both.[/QUOTE]

Would you rather get two units of work done in 90ms, or one unit done in 60ms?

A core2duo is dual-core, while a core2quad is indeed 4 cores. The bottleneck you refer to concerns memory access, and does not exist for smaller LL tests- unfortunately, quite a bit smaller than Prime95.

The original poster isn't with us anymore. save your keystrokes. Mine are ok to waste.