Sign in with
Sign up | Sign in
Your question

Need good performance testing software - Compare w/ HT

Tags:
  • CPUs
  • Performance
  • Computer
Last response: in CPUs
Share
March 11, 2006 12:16:45 AM

Hey. I've got a friend who just purchased a computer from Dell. Her computer came with a 2.8GHz Prescott with Hyper Threading. It also has 512MB DDR2 Ram. I ran SuperPI on her computer, and it did 1M in 47 seconds.

Compare that with my computer, which is running a 2.6GHz Northwood OC'ed at 3.04GHz. It has 1024MB of P2700 Ram. My computer does 1M in 57 seconds.

I'm curious as to what exactly would cause her computer to run it so much faster. Is it because hers has hyperthreading? It seems to me though that SuperPI wouldn't be able to multi-thread, so hyperthreading wouldn't give her any advantage on that program.

Is there any software that can test JUST the CPU and give it a marking? I would like to be able to see a direct comparison between her 2.8GHz w/HT and my 3.04GHz w/o/HT.

Thanks!

More about : good performance testing software compare

March 11, 2006 1:00:12 AM

There are two reasons I can think of just off the top of my head. The first is HT has improved the performance of executions meaning there is less ‘wasted’ time while processing information. The second is overall performance per clock cycle has been increased slightly, but not consistently. Some software just runs faster while other software runs slower.

I’m no Intel fan nor am I an Intel guru so without going to look up the differences in cores have you looked at the L1 and L2 cache? That may also have an impact.
March 11, 2006 2:43:45 AM

One thing I would add, is the FSB as well. Really hard to say what version your friend has exactly, as well as prescott socket. As far as L2 cache, I don't think your going to see that much of a difference once your have 512KB.

My Northwood 2.66 is 533 FSB, non HT.

Since you don't have HT, I'm guessing your P4 2.6 A which is 533FSB, and your running PC2700 (DDR333).

Now with the Prescott, it may be on a 800FSB, PC3200 (DDR400)
or the 533 FSB.

There are a few of versions of it...

The versions are:

533 FSB
socket 478 -2.8 A

socket 775
511


800FSB

socket 478 -2.8 E

socket 775 - 520 / 521

Not sure if I covered all the prescott versions.. :lol: 

So just saying the 800FSB would increase performance on the 2.8 vs your 3.04 OC.
Related resources
a b à CPUs
March 11, 2006 10:55:16 AM

geez my p4c at 3ghz gets 45sec, that score is poor for both machines!
March 11, 2006 8:25:37 PM

Quote:
geez my p4c at 3ghz gets 45sec, that score is poor for both machines!


:? . o O (The 2.8ghz did it in 47 (2 sec difference) sec, and he considers it poor?)

Comeon man... I'm guessing its a dell hate thing for you.

And the 2.6 OC to 3.04 running at 533FSB gives a clear indication that the 800 FSB does give it a boost for 10 sec. Unless I really underestimated the FSB as a 400. Perhaps I should find this SuperPI program.. heh

Though.. I agree with Jumping Jack, that HT prolly didn't have much to do with it.
March 11, 2006 8:53:56 PM

Thank you all for all your responses. According to CPU-Z however, my FSB on my computer is 100MHz (OC'ed to 117MHz). It then says that the bus speed is 468.1MHz (quad-pumped I understand). Mine isn't 533 or 800 like you guys have been talking about.

Also, so there really is no software to directly compare a single-core no HT with a singlecore w/HT? That kinda stinks. Oh well, I can just tell her that her computer is better period, without any bench testings. It'll make her happy! 8)
March 11, 2006 9:12:31 PM

o.O; Wow, makes me kinda glad for my socket 939 3200+ Winchester at 2.0Ghz. Does 1M in super pi in 46 sec.

Though I'm very interested in Conroe, to see if the performance increase is as much as the Tom's Hardware article on the Pentium M shows. Might have to switch to Intel if it's as good as they say.
March 11, 2006 9:30:38 PM

Quote:
Thank you all for all your responses. According to CPU-Z however, my FSB on my computer is 100MHz (OC'ed to 117MHz). It then says that the bus speed is 468.1MHz (quad-pumped I understand). Mine isn't 533 or 800 like you guys have been talking about.

Also, so there really is no software to directly compare a single-core no HT with a singlecore w/HT? That kinda stinks. Oh well, I can just tell her that her computer is better period, without any bench testings. It'll make her happy! 8)


I just did the test on mine.

I'm on a 533 bus:

2.66 nonHT - stock speed - 19th iteration - 1 min
3.00 nonHT - OC FSB 600 - 19th iteration - 57 sec

So.. either something is up with my system or there no difference on the 400 vs 533 bus.



Though I did forget something - I am using 1 stick of DDR400 rather then 2. I was hoping to get a P4 3c cheap for an upgrade.. hehe
March 11, 2006 9:52:22 PM

3dmark2006 is.. gruesome on my cpu :cry: 

3dmark2006

CPU Score 778 - OC 2.9 - Prime95 Stable
!