Re: Mersenne: Comparitive bench, please?
Thanks to all who answered! At 10:23 PM 12/7/99 +, you wrote: >Assuming the processor in your 233 system really is a PII, you would >seem to have a slow clock (perhaps it's running in doze mode? This seems to have been it. Check >the BIOS settings & disable power-saving) or another application >running at a normal or high priority chewing up the majority of the >available CPU time (check the NT Task Manager, which can be brought >up using CTRL+ALT+DEL when running a normal Windows session. Really!) I'm nearly MCSE certified (one test to go), so this, I knew I'd already confimed that the CPU was popping at 100%, 96% or so of it by NTPrime. It HAD to be underclocking somehow, and it was. Dunno WHAT it was, but it was, because allowing the BIOS to save settings for "optimum performance" was all that was needed. >Also check the processor type in Prime95/NTSetup Options. If you have >set the CPU type to something else, then the clock timings reported >by Prime95 will be wrong, since they're obtained from the tick count >divided by the processor speed. I wasn't going on that report. (I run the NT Service version, so that info from the screen is not available in any case). I did it by "doing the math" -- stopping the service to put a "progress report" into the results.txt file, then restarting it and letting it run for an hour, then restopping it. Simple division with seconds elapsed versus iterations during that hour yields the time per iteration. You have to do this for longer than a few seconds, though, to get a good report. However, in addition to the overall slow performace of the machine while in use, this is what tipped me off, from my individual accounts report (edited to 80 chars per line): prime fact currentdays datedate exp.bits iterat. run togo exp. updated assigned computer ID --- --- --- - 8297941 63 4390912 91.8 93.7 60.7 07-Dec-99 14:30 07-Sep-99 SKYWALKER 8591381 64 6815743 60.0 14.9 60.9 07-Dec-99 19:43 08-Oct-99 PALPATINE Palpatine is a Pentium/200 (not even a P-II). Skywalker is the P-II in question. All other machines are running at a normal speed, as well, but this one seemed a bit pokey... _ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Comparitive bench, please?
DING! Give that man a ceegar! While I didn't find a setting for anything like that anywhere in the BIOS, I took a risk and let the BIOS reconfigure itself for "optimal settings" in the hopes it wouldn't FUBAR everything and the system is back to its snappy old self again... Thanks, Jud! At 03:01 PM 12/7/99 -0500, you wrote: >At 01:57 PM 12/7/99 -0500, you wrote: > >>I have a P-II/233 running NT 4.0 that I think is severely >>underperforming. It has 256 MB of RAM, and much of it remains >>free. Some applications take literally MINUTES to load, and I suspect a >>CPU problem, > > >I don't know much about NT systems, but my father recently had a machine >that was running extremely slowly. I finally found the problem - in the >setup, it had inadvertently set to run at "compatibility speed", which is >very slow. That could be the problem. > > >+--+ >| Jud McCranie| >| | >| Where a calculator on the ENIAC is equipped with 19,000 | >| vacuum tubes and weighs 30 tons, computers in the future | >| may have only 1,000 vacuum tubes and perhaps only weigh | >| 1.5 tons.-- Popular Mechanics, March 1949. | >+--+ _ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Comparitive bench, please?
No, it shows the NTPrime service taking 98% of the CPU time. I've run benches on it, too, and it is just plain running slow, like a 66 Mhz or slower machine I'll be looking for that compatibility mode thing shortly. At 03:11 PM 12/7/99 -0500, you wrote: >On Tue, 7 Dec 1999, Jeff Woods wrote: > > > I have a P-II/233 running NT 4.0 that I think is severely > > underperforming. It has 256 MB of RAM, and much of it remains > free. Some > > applications take literally MINUTES to load, and I suspect a CPU problem, > > although PrimeNT reports no calculation errors. I suspect that the CPU is > > simply underclocked for some unknown reason -- the fan is working properly. > > > > It is taking 1.788 seconds per iteration for exponents in the 8,290,000 > range. > > > > Could someone else who has a machine similar to this let me know if > that is > > an unreasonably high time per iteration?It doesn't have to be an > > identical machine, but one reasonably close. > >Jeff, > >Sounds to me as if something is wrong with the PII-233. I have a pretty >lightly loaded PII-266 machine with 128MB RAM running WinNT 4 (SP3) that >is getting iteration times of .398 on an exponent around 8.4 million. > >Does task manager show some other process stealing cycles from Prime95? > >Kel _ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
RE: Mersenne: Comparitive bench, please?
My 233 is running much better than your 1.788 as well. A few million things come to mind... First, are you running the newest software? Prime95 is up to, what... 19.0.2? Something like that... I'm sure it's on the page. Also, if you're in a Microsoft environment, Richard's right... check for things in the Task Manager. "FindFast" is a Microsoft program that likes to eat CPU time, and can be removed from the "StartUp" folder if it's on your system. On the more obscure side of things, I just upgraded my motherboard, and had to do a lot of manual settings to get it to work 'efficiently' with my 233. A 233 is running a 66Mhz clock at a 3.5 multiplier rate... the motherboard I got (an Iwill motherboard) defaulted to something else, and it wasn't readily obvious that this was the problem. (Okay, it was readily obvious if you read the manual, but come on, who does that?). I/O isn't an issue with this software, but you may want to check your cache (L1 and L2), perhaps by disabling/enabling it in the bios to see how that affects performance, and you may want to download some benchmark programs (ZDNet has some good ones, I believe) to check your cache and Ram speeds. Generally those are the problems I find causing software to run poorly. ---Chip On Tue, 7 Dec 1999, Pardoe, Richard (PRDR) wrote: > I have a Pentium 232 running NT 4.0 with only 96 MB of RAM that is getting > about 0.6 seconds per iteration for an exponent in the 8,490,000 range. > > It might be worthwhile to look in the Task Manager to see what is using the > CPU time. My CPU is typically mid-90% of the time w/ Prime95. > > Rich > > -Original Message- > From: Jeff Woods > > I have a P-II/233 running NT 4.0 that I think is severely > underperforming. It has 256 MB of RAM, and much of it remains free. > ... > It is taking 1.788 seconds per iteration for exponents in the 8,290,000 > range. > > > _ > Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm > Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers > \\ ^ // (o o) ---oOO--(_)--OOo | Chip Lynch| Computer Guru| | [EMAIL PROTECTED] || | (703) 465-4176 (w) | (202) 362-7978 (h) | _ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
RE: Mersenne: Comparitive bench, please?
I have a Pentium 232 running NT 4.0 with only 96 MB of RAM that is getting about 0.6 seconds per iteration for an exponent in the 8,490,000 range. It might be worthwhile to look in the Task Manager to see what is using the CPU time. My CPU is typically mid-90% of the time w/ Prime95. Rich -Original Message- From: Jeff Woods I have a P-II/233 running NT 4.0 that I think is severely underperforming. It has 256 MB of RAM, and much of it remains free. ... It is taking 1.788 seconds per iteration for exponents in the 8,290,000 range. _ Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
