Is overclocking really mean equal too?

[TheGoatMan]

From what I've seen people are comparing overclocked processors to higher clocked factory spec processors. Now from what I understand higher clocked newer processors have better technology and are more efficient. So can you really compare a 2500+ overclocked to 3200+ speed? Like for example the 3200+ 400mhz fsb would obviosly be better. So can anybody clear this up? Because I see this like modifying a honda civic to be as fast as a corvette but lacking the refinement.

 

[ceh4702]

On benchmarks pay attention to the Video card used and the memory and chipset and even the BIOS Revision and special settings. Sometimes they turn off features like Sound or specific Video Card options to get the best scores. You would not be wanting to do this to play your average video game. Some people at the ATI and Nvida have also been known to put tweaks in their video cards for specific games to make them faster.

A better judge would be to compare shoppers comments on video cards at sites like

http://www.newegg.com

Even if you have a good brand, there can be a big difference between two different models of video cards.

Overclocking is really a big waste of time as far as I am concerned. You have to significantly increase the clockspeed without causing the Memory timings to degrade to see better performance. Once you invest in additional fans and better CPU coolers, you could just buy a faster processor.

If you do not like my post feel free to point out your opinion or my errors.

 

[cdogg]

GoatMan,

[blue]2500+ was tested on an Abit nf7-s
2600+ was tested on a MSI KT6 delta lsr[/blue]

Your problem is that the tests you've run aren't on equal playing grounds. You have too many variables to account for. In order for you to make a comparison based on benchmarks, you want to isolate the CPU. In other words, keep everything else the same (motherboard, hard drive, operating system, memory, etc). Keep as much as you can the same when testing.

After that, let us know what you find...


~cdogg
[tab]"The secret to creativity is knowing how to hide your sources"
[tab][tab]- A. Einstein

 

[SYAR2003]

That was why i posted the link to
thetechboard.com test.
Because all the cpu's was tested on the same plattform.

 

[cdogg]

GoatMan,

Also, you have to consider AMD's naming scheme. A 2600+ should outperform a 2500+ at stock speeds, regardless if one's a Thorougbred or Barton - assuming they are using the same equipment. Otherwise, if AMD expected the Barton 2500+ to outperform the Thoroughbred 2600+, they would simply choose a higher number rating for the Barton.

As it's been pointed out, the Barton 2500+ manages to outperform the 2600+ in quite a few application benchmarks which favor L2 cache more than CPU speed. Some applications don't. So it depends on the benchmark. Take a look at how the two can reverse roles:

http://www.tomshardware.com/motherboard/20030414/i875p-26.html

The article I just posted a link to also compares Gaming, Synthetic, and Audio/Video benchmarks. You'll find that the Barton is just slightly faster than the Thoroughbred in most of the "non-synthetic" benchmarks. However, synthetic benchmarks that test just the CPU usually favor the Thorougbred. Why? The only thing really being measured is CPU speed (clock cycles), which is hardly fair to the Barton which is running 300MHz slower.

In the end, some benchmarks matter while others just don't tell the whole story...


Overclocking is an entirely different argument and shouldn't be confused with stock speed comparisons. Besides, this forum is mainly to discuss real-world situations in a professional environment. Overclocking doesn't quite fit in that picture...


~cdogg
[tab]"The secret to creativity is knowing how to hide your sources"
[tab][tab]- A. Einstein

 

[TheGoatMan]

Once again you are looking at a 266fsb 2600+. Check the bus speed on that link. But shouldn't the nforce board provide the 2500+ an unfair advantage?

 

[cdogg]

What point are you trying to make? That the Thoroughbred 2600+ 333MHZ FSB is faster than the Barton 2500+ 333MHz FSB?

If so, look at my last post more carefully:
"[blue]A 2600+ should outperform a 2500+ at stock speeds, regardless if one's a Thorougbred or Barton[/blue]"

It makes sense that the 2600+ is faster. That's AMD's MHz naming scheme for ya...


~cdogg
[tab]"The secret to creativity is knowing how to hide your sources"
[tab][tab]- A. Einstein

 

[cdogg]

"[blue]But shouldn't the nforce board provide the 2500+ an unfair advantage?[/blue]"

Why speculate? Why not test the CPU's in the same environment and find out??


~cdogg
[tab]"The secret to creativity is knowing how to hide your sources"
[tab][tab]- A. Einstein

 

[TheGoatMan]

Well there was a significant difference between the 266 and the 333. Also I'm just trying to figure out if I should send my tbred back and pick up the barton thats all. I just figured before I jumped to conclusions because of what i tested I'd just get some opinions first. Thats all. I wasn't implying the 2600+ was faster it just seemed I wasn't getting the results I expected.

 

[cdogg]

I wouldn't. If you have the 333MHz tbred, then getting the Barton 2500+ won't do much if anything for you (with the absence of overclocking).

The 266MHz tbred is just slightly behind the Barton, partly because of the extra L2 cache and also because of the faster FSB. However, the 333MHz tbred makes up for most of it, and is faster in a majority of benchmarks that I've seen (and not just synthetic benchmarks).

 

[TheGoatMan]

I'm not much for overclocking. I only do it on occasion. I have more than enough power to run games and stuff at default speeds. I think I'll just hold out for a few years and go 64 bit.

 

[cdogg]

"[blue]I have more than enough power to run games and stuff at default speeds[/blue]"

I agree.