Pentium D 945 Dual core (3.4Ghz) vs Celeron D 3.46Ghz

RGPilling

Distinguished
Oct 24, 2009
6
0
18,510
I just swapped a Pentium D 945 processor (dual core) into my MS7222v3.1 MB with v5.5 Bios update.

I had written an Excel loop to see how much faster the dual core might be, and to my surprise, what the Celeron did in 26.2 minutes, it took the Pentium D 46.3 mnutes to accomplish. Is it just me, or does that seem counter intuititive?

The benchmark for the Celeron was after upgrading the Bios, and the only thing changed was the processor. I also ran some benchmarks on the processor, and it's coming in about equivialent to a Pentium 4 - 1.7Ghz. (Roughly ½ the clock speed of the chip.)

Any ideas what can be done to rectify this apparent problem, or is this chip just a dog?

Any suggestions are appreciated.
 

RGPilling

Distinguished
Oct 24, 2009
6
0
18,510
The actual chips of RAM that I have in the MB are 1x1Gb and 1x2Gb. Unfortunately, the MB will only handle 1Gb per slot, thus the 2nd one is used as a 1Gb for all intents & purposes. The chips are both PC2-5300. I haven't set anything specifically for the RAM speed, that's just what comes up. (It's all the same as when I was running the Celeron processor, except that it all runs slower now.)
 

ainarssems

Distinguished
Mar 4, 2008
723
1
19,010
Two things that I would consider aside from those already mentioned:

1. If You just swapped CPU from single core to dual core, You are probably still using uniprocessor HAL instead of multiprocessor which makes You only use on core of Pentium D while other is doing nothing. This would explain why You do not get any improvement but not why Your performance is cut in half. Check task manager if it shows two graphs for CPU or only just one. Also there is a way to change HAL from uniprocessor to multiprocessor without reinstall I would recommend to reinstall Windows when changing from single core to multiple.

2. Pentim D probably needs more power and either PSU or motherboard cannot provide enough juice for it so it is working with errors which are selfdetectable and not critical so it keeps looping doing the same thing until it does it without error.
 

RGPilling

Distinguished
Oct 24, 2009
6
0
18,510
I checked the task manager, and it's showing 2 processors working. I'm not sure what HAL is.

The PS is a 450 Watt, and it has the supplementary 4pin line attached to the MB, so it's probably getting enough power.

It's just baffling to me as to why it's running so poorly.

If there's anything else that you can think of, I'm all ears.
 
OK look at the comparison below in this link and tell me what you see?

http://ark.intel.com/Compare.aspx?ids=27520,27132,

http://ark.intel.com/Product.aspx?id=27520

http://ark.intel.com/Product.aspx?id=27132

Your FSB ... 800 (200) for the 945 (3.4Ghz) = 17 multiplier

Your FSB ... 533 (133) for the 360 (3.46Ghz) = 26 multiplier

Will your ram run at 800FSB ??

Go into the bios and take a peek?

Right click on My Computer and see what clock speed is showing under WinXP ... or try CPUID (a little utility).

The 945D should smoke the Celery as it has a heap of extra cache, a faster FSB and most importantly two CORES ... not one ... albeit it runs slightly slower.

 
yes thanks ... missed that so did my own calcs.

3:4 divider but running 266FSB ... comes in at 798 (800) so that's fine.

I can only think that you might have poorly installed the cpu and HSF and it is throttling under load (thermal throttling is common on the Prescott line above 3.2Ghz.

Shutdown, remove the HSF, clean the base of the HSF, reapply thermal grease and likewise clean the top of the CPU and reassemble.

Run a decent application to put the cpu under load and monitor the temps with a utility to see if it is throttling back.


http://www.hardwaresecrets.com/article/104

Hope this helps.
 

RGPilling

Distinguished
Oct 24, 2009
6
0
18,510
Thanks for all the advice. You have the 2 chips bang on (from what I remember). Actually, I've had the chip in and out so many times, it's likely spinning from the turnarounds. Each time, I've cleaned the chip & the heatsink with alcohol, and then reapplied the cement. The pattern from the cement comes out even each time before removal, so I think I'm getting it right.

I'll give it one more try, and then I think I'll go back to the Celeron if it doesn't improve. I'm a Chartered Accountant, and with the amount of time I've spent on this "quick fix", I could have earned enough money to buy 3 or 4 new units. Dang I hate my curiosity!!!

Thanks again.