• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

[Discussion]Should I get a Quad- or Dual-Core?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
the new p45 boards finally allow this.

While they do allow a bit more than previous chipsets, it doesn't seem to be ubiquiteous. Case in point: I haven't seen results from your QX9650 with any more than 445FSB, and that was only stable until you tried to reboot.

I'd rather see a few more people with some Quad + P45 results before we flatly say that the P45 chipset has definitely fixed it versus being some combination of things...
 
While they do allow a bit more than previous chipsets, it doesn't seem to be ubiquiteous. Case in point: I haven't seen results from your QX9650 with any more than 445FSB, and that was only stable until you tried to reboot.

I'd rather see a few more people with some Quad + P45 results before we flatly say that the P45 chipset has definitely fixed it versus being some combination of things...

You didn't see my 500 fsb Kentsfield Q6700 did ya?

http://www.ocforums.com/showthread.php?t=570687

This was bootable all day long, I also had it prime stable with 3.5ghz, never primed for 4ghz tho.
 
You didn't see my 500 fsb Kentsfield Q6700 did ya?

http://www.ocforums.com/showthread.php?t=570687

This was bootable all day long, I also had it prime stable with 3.5ghz, never primed for 4ghz tho.

*sigh* And yet again, the point went entirely over your head.

A Q6xxx can do high FSB's on even a P35 chipset. But you already knew that, and you already knew that I've (and several others) have mentioned the FSB wall on Q9xxx processors which is why you made a point to mention the P45 chipset...

How's your QX9650 doing with high FSB's prime? Not so good as your Kentsfield I gather... And maybe that was the point?
 
Actually I just booted 475 fsb about 5 mins ago, still climbing. Its just a matter of getting the settings right, the CPU GTL Ref is the key.

And while you may think 500 fsb on a kentsfield is easy, I can beg to differ.

Thus far my QX has pulled 445, 450, 475, and I'll try for 485 and then 500 later tonight. But my point is, I don't think it's CPU related entirely, I think the motherboard has alot to do with it.

But back on topic, it still remains to whatever the person is trying to accomplish, do they want absolute highest clockspeed? Or do they want absolute highest benching? Picking a quad or dual really depends on it's application.

For me that application is gaming, I want quadcore at 4ghz with safe 24/7 voltages.... I definitely got what I wanted.
 
Last edited:
Again, as I've said before, I'd rather that you did show that you can get it to higher FSB's on the P45 chipset... However, you've yet to show that with any stability at all, in fact you've yet to show much of anything over 400 x 10 with stability yet. It's not that you can't or that you won't, but until you do, using the P45 chipset as some sort of "out" for Quad overclocking doesn't make logical sense.

Until we can see some repetitive successes with P45 chipset and Q9xxx processors, we have no way to say that the issue is resolved.
 
Quads are harder to OC and now the Q6600's are slower than the New line of wolfdale CPU's but they still rock . If I were to answer this question honestly I would say get a NEW wolfdale a E8600 they are smokin fast easy to OC and are sweet so that said have you chosen what you are going to get ?
 
Your right I haven't shown stability with 445 fsb

qx_stable.JPG
 
I moved from a quad to a dual and have been very happy with the switch. I had a pretty golden q6600 but being on air cooling that thing was a little heat factory. Moving to a dual has made it possible to get very high clocks out of my CPU while staying with air cooling. Not bad for a SFF that I can literally take wherever I want easily.

I moved from hardcore benching to more light duty benching/ gaming and the new series of dual cores was a perfect fit.
 
:beer: 445 is a start, but that's really not that far for someone stuck on a 8x multiplier :) I can run 455FSB all day every day with 100% stability on my Q9450, and I can boot and take screencaps at 463 if I don't mind whacking out the GTLREF settings and feeding it a lot more FSB voltage.

For those who don't have the cash for a $750 unlocked-multi processor, 445FSB really isn't a godsend yet :)
 
Point is, just because I have the unlocked multi doesn't mean its not hard to attain. Hell my X3350 hit 450 x 8 stable with an ASUS Commando P965 chipset.

The p45 boards are showing better quad clocks and that's not just me saying that.

Right now my FSB term voltage is manually set to 1.2 which is the lowest the board allows. My NB is stable at 1.35 for 445fsb.
 
Point is, just because I have the unlocked multi doesn't mean its not hard to attain. Hell my X3350 hit 450 x 8 stable with an ASUS Commando P965 chipset.

The p45 boards are showing better quad clocks and that's not just me saying that.

Right now my FSB term voltage is manually set to 1.2 which is the lowest the board allows. My NB is stable at 1.35 for 445fsb.

And again... I'm doing 450FSB x 8 stable on all-stock volts across the board, even on the CPU, on my X38 chipset. So? We're dealing with processors that have a max multiplier of 8 for all currently available quad penryns that are below $750.

8x 450 is paltry compared to what's available on a dual. So, let's see some 475, 500, and even higher FSB's to drive your point home that a quad can overclock just as well as a dual. You've said it before, let's see it! Unless of course you mean an extreme edition unlocked multiplier quad can overclock as well as a non-extreme, non-unlocked multiplier dual -- in which case, what kind of argument is that really?
 
Well I will say this, I went from a Q6600 to an E8400 this week. I went in knowing I would probably lose points in 3dMark06. What I did not expect was I have lost ground in every benchmark I have tried so far except SuperPi. This is with the 6600 @ 3.6 vs. the 8400 @ 4.0. I am surprised by that immencely and am really baffled by losing so much ground in Wprime 32m and 1024m. Right now I am glad I didn't sell the Quad. The 4GHZ chip is nice and I have not tried any gaming as of yet, but if you like to bench, maybe the Quads are the way to go.
 
Well I will say this, I went from a Q6600 to an E8400 this week. I went in knowing I would probably lose points in 3dMark06. What I did not expect was I have lost ground in every benchmark I have tried so far except SuperPi. This is with the 6600 @ 3.6 vs. the 8400 @ 4.0. I am surprised by that immencely and am really baffled by losing so much ground in Wprime 32m and 1024m. Right now I am glad I didn't sell the Quad. The 4GHZ chip is nice and I have not tried any gaming as of yet, but if you like to bench, maybe the Quads are the way to go.

This is just what I said when I had the E8200 in I was loosing at every benchmark even with the 3.7GHz clock speed , Just one reason why I went back to my Q6600 .

...........i come to visit the forums hoping for a interesting topic to come up...but nooooo...Monday: 'Quad Vs Dual' Tuesday: 'Quad Vs Dual' wednesday: 'Quad Vs Dual' thursday: ahh slightly different..'Dual Vs quad'-lucky me.. friday 'Quad Vs Dual' saturday 'Quad Vs Dual' sunday: day off..monday: 'Quad Vs Dual'....and so on until octo core comes out...(sigh)
:attn::attn::clap::clap:
 
prime81 said:
Thus far my QX has pulled 445, 450, 475, and I'll try for 485 and then 500 later tonight. But my point is, I don't think it's CPU related entirely, I think the motherboard has alot to do with it.

Your right I haven't shown stability with 445 fsb

So what happened to 450, 475, and 485 FSB? 445 is nothing special really, majority of Q9xxx can do it with a good motherboard.
 
Stick with Dual Core, Quads are just un-necessary unless all you do is benchmark or design 3d stuff.
 
what about unreal engine 3? i have a few games that use it, and i heard quad core's are the way to go for such games?

So one or two games in a collection of dozens is enough to make a decision? Seriously, think about what you're asking.

Let's ask it another way just to point out what should be obvious: "should I buy a processor that's slower in almost every other game so that two games can have a 10-15% speedup?"

Does the second question make the point any clearer?
 
So one or two games in a collection of dozens is enough to make a decision? Seriously, think about what you're asking.

Let's ask it another way just to point out what should be obvious: "should I buy a processor that's slower in almost every other game so that two games can have a 10-15% speedup?"

Does the second question make the point any clearer?

I haven't tested more than 1 game, but Crysis saw no difference in performance from 2.7GHz to 4.05GHz.

I just changed the multi so that all other variables remained constant. See my sig for details.

Vista-64 DX10 1680x1050 No-AA All-High settings

Average FPS for the last 2 of 3 runs shown:

6x450=2.7GHz 36.23

7x450=3.15GHz 35.925

8x450=3.6GHz 35.935

9x450=4.05GHz 35.885


The extra CPU speed does nothing in this game.

In this case neither CPU has the gaming advantage. Both cost the same. The quad is hotter, though, and requires better cooling.

*Side note....my current 45nm quad runs cooler and requires less voltage than my old E3110. 3.944GHz (464FSB) at 1.208v 24/7 stable!

I think higher dual-core speeds may come into play w/ x-fire/SLI, but I don't have that kind of setup yet.


I'm not trying to make a blanket statement w/ this post, but I would like others to run some games at different CPU speeds to see if the extra speed of the dual-cores can really make a difference at normal gaming resolutions (1680x1050+).
 
Last edited:
What were the minimum framerates on those benchmark passes Jason?

I don't remember them being any different, but I can't say that for sure. I saw someone else run a similar benchmark, but it was at 1024x768, and the lower CPU speeds definitely affected min fps. I don't think there is any difference at 1680x1050+, though. If the min was lower on the lower speed tests then the avg would have to be affected unless the max was some how higher.

Are you (or anybody else) setup to run this kind of test? My PC is in a state of flux right now, and running this test again will require a few hours of work to get setup.
 
Back