Sign in with
Sign up | Sign in
Your question

Core 2 duo performance/clock speed rate question.

Last response: in CPUs
Share
September 7, 2006 10:09:27 PM

I was just wondering why the core 2 duo architecture increases clock speed at a higher rate than previous intel stuff and AMD stuff. What i mean is that when AMD has levels of processors they go up by about 200mhz. But Intel's C2D is around 3 going either lower or higher but close to 300mhz. The C2D processors must increase their clock speed higher each time to get the same performance increase rate as A64. Now obviously C2D is better now but i am curious because that architecture seems to have such a spread. Like the e6300 starts at i think about 1.83ghz. Then next is the e6400 at i think 2.1ghz? Then the 6600 at 2.4, 6700 at 2.66 and the 6800 at 2.93. I have seen elmoisevil's sig and he has his 6800 at 4.45ghz and thats very high. I imagin the limit would be around 5ghz and the final stock one will be around 4ghz. But do these huge clock speed differences increase performance in a fairly linear way or do the benefits decrease very rapidly as you get to the extremes. So would elmo's processor be about twice as fast as an e6300? I am curious because of the idea that C2D works harder per clock than A64. Now this is true at lower speeds but I'm wondering how they perform at higher speeds. Does anyone know of any benchmarks for the 3.6ghz FX-62. I'd like to see an overclock of a conroe by a similar percentage to see if the FX closes the gap in the section of speeds on the x6800 or if the gap just widens. I think that if it starts to close the gap then maybe future K8L chips will have higher clock speeds as the main change. It would be interesting to see what clock speed an A64 would need to be to compete with a 5ghz conroe, be it 4.5ghz, or 8ghz. I just want to see the trend. Sorry for my lack of using paragraphs as i did not have any plan as i was writing. 8)
a b à CPUs
September 7, 2006 10:50:07 PM

Quote:
I was just wondering why the core 2 duo architecture increases clock speed at a higher rate than previous intel stuff and AMD stuff. What i mean is that when AMD has levels of processors they go up by about 200mhz. But Intel's C2D is around 3 going either lower or higher but close to 300mhz. The C2D processors must increase their clock speed higher each time to get the same performance increase rate as A64. Now obviously C2D is better now but i am curious because that architecture seems to have such a spread. Like the e6300 starts at i think about 1.83ghz. Then next is the e6400 at i think 2.1ghz? Then the 6600 at 2.4, 6700 at 2.66 and the 6800 at 2.93. I have seen elmoisevil's sig and he has his 6800 at 4.45ghz and thats very high. I imagin the limit would be around 5ghz and the final stock one will be around 4ghz. But do these huge clock speed differences increase performance in a fairly linear way or do the benefits decrease very rapidly as you get to the extremes. So would elmo's processor be about twice as fast as an e6300? I am curious because of the idea that C2D works harder per clock than A64. Now this is true at lower speeds but I'm wondering how they perform at higher speeds. Does anyone know of any benchmarks for the 3.6ghz FX-62. I'd like to see an overclock of a conroe by a similar percentage to see if the FX closes the gap in the section of speeds on the x6800 or if the gap just widens. I think that if it starts to close the gap then maybe future K8L chips will have higher clock speeds as the main change. It would be interesting to see what clock speed an A64 would need to be to compete with a 5ghz conroe, be it 4.5ghz, or 8ghz. I just want to see the trend. Sorry for my lack of using paragraphs as i did not have any plan as i was writing. 8)


Core 2 Duo scales in an almost linear fashion but for games.. you really cannot grasp the true performance of Core 2 Duo as you end up being bottlenecked by the graphics card before you even begin to tap into the extra power.

Cinebench & SuperPi can give you an idea of the performance of the processor.

[/img]


As you can see the SuperPi results scale quite well...
September 7, 2006 11:24:17 PM

I can't say for higher clocks but when I got my C2D I setup a rendering benchamrk to see what improvements over my old system the new chip brought.

C2D E6600 @ 2.4 :6 mins 5 seconds
C2D E6600 @ 3.1 :4 mins 41 seconds

Rendered in Lightwave which has multi-threaded support so it runs both cores at 100% for the duration of the render. Limited data I know but it does show pretty good scaling at least to 3.1ghz.

And just for the hell of it :) 

Athlon XP2400 at stock speed 39 mins 26 seconds.
Related resources
September 8, 2006 12:50:54 AM

I'm not even going to read that until you learn to use paragraphs.
September 8, 2006 1:20:04 AM

Quote:
I'm not even going to read that until you learn to use paragraphs.


Ohhh... come one its not that long for one paragraph. 8) I told you i was just typing, not planning an essay. I don't think it is too hard to understand even so.
September 8, 2006 1:25:07 AM

That one paragraph gives LOTR a run for its money. Come on they teach this in primary school.
September 8, 2006 1:35:54 AM

I reckon Intel has deliberately under-clocked the C2D. May be so that they can easily bring out higher speed CPUs, or may be to allow enthusiasts to OC to their hearts content.

Either way every C2D can be overclocked with stock cooling. And if u have some decent cooling system, u can go bezerk. There is no doubt OCing a C2D to 4ghz will give u a considerable increase in performance. The E6800 already beats the E6300 fairly easily at stock speed. To have E6800 at 4.5ghz will easily give twice as much of E6300.

This article compares OCed C2D with OCed FX-62 and the picture is pretty clear.

THG Tuning Test: Core 2 Extreme vs. Athlon 64 FX-62
September 8, 2006 1:51:30 AM

Quote:
Core 2 Duo scales in an almost linear fashion but for games.. you really cannot grasp the true performance of Core 2 Duo as you end up being bottlenecked by the graphics card before you even begin to tap into the extra power.

Cinebench & SuperPi can give you an idea of the performance of the processor.




As you can see the SuperPi results scale quite well...
[/img]When I first saw the SuperPI of 23 seconds I thought it was 1M and I was thinking "what the hell? that's a lot slower than I remembered" :lol: 
September 8, 2006 7:40:34 AM

Quote:
I was just wondering why the core 2 duo architecture increases clock speed at a higher rate than previous intel stuff and AMD stuff. What i mean is that when AMD has levels of processors they go up by about 200mhz. But Intel's C2D is around 3 going either lower or higher but close to 300mhz. The C2D processors must increase their clock speed higher each time to get the same performance increase rate as A64. Now obviously C2D is better now but i am curious because that architecture seems to have such a spread. Like the e6300 starts at i think about 1.83ghz. Then next is the e6400 at i think 2.1ghz? Then the 6600 at 2.4, 6700 at 2.66 and the 6800 at 2.93. I have seen elmoisevil's sig and he has his 6800 at 4.45ghz and thats very high. I imagin the limit would be around 5ghz and the final stock one will be around 4ghz. But do these huge clock speed differences increase performance in a fairly linear way or do the benefits decrease very rapidly as you get to the extremes. So would elmo's processor be about twice as fast as an e6300? I am curious because of the idea that C2D works harder per clock than A64. Now this is true at lower speeds but I'm wondering how they perform at higher speeds. Does anyone know of any benchmarks for the 3.6ghz FX-62. I'd like to see an overclock of a conroe by a similar percentage to see if the FX closes the gap in the section of speeds on the x6800 or if the gap just widens. I think that if it starts to close the gap then maybe future K8L chips will have higher clock speeds as the main change. It would be interesting to see what clock speed an A64 would need to be to compete with a 5ghz conroe, be it 4.5ghz, or 8ghz. I just want to see the trend. Sorry for my lack of using paragraphs as i did not have any plan as i was writing. 8)
I think what you're referring to is that each CPU step is 266MHz, vs K8/PD's 200MHz. That's because of the K8/PD's 200MHz "FSB"...so a 10x multiplier gives 2000MHz, and an 11x multi gives 2200(thus 200MHz increase).C2D uses a 266MHz FSB, so every multiplier increase(of 1x) gives an external speed increase of 266MHz.(266x10=2660,266x11=2926).
September 8, 2006 12:20:02 PM

Congratulations! How much did all the setup cost if it cam be public info? (However I know I'd afford it only if I broke an ATM or something like)
a b à CPUs
September 8, 2006 1:15:45 PM

Quote:
Congratulations! How much did all the setup cost if it cam be public info? (However I know I'd afford it only if I broke an ATM or something like)


Those are from my friends over at Xtremesystems. The 5.4GHz was done using LN2.. so it's not VERY practical. I preffer chillers, Thermo-electric modules or vapochilling instead. More practical and doesn't cost that much if you DIY.
September 8, 2006 1:16:06 PM

Dang, those E6600s and E6700s hold their own against an OC'd FX-62!
a b à CPUs
September 8, 2006 1:28:45 PM

Quote:
Dang, those E6600s and E6700s hold their own against an OC'd FX-62!


Yes... that's why many, such as myself, claim them to be such superior processors (when compared to today's competition). Of course this doesn't mean an Athlon64 X2 is a bad processor... it just means that given the way we use our processors and the fact that we use them in such a manner that exposes all of Core 2 Duo's strengths and none of it's weaknesses is what makes it the best processor.

It's quite clear that the Opteron in 4 to 8 way configs have the potential to be superior.. but sadly we don't have software that can take advantage of such configurations. Therefore we're utilizing all of the K8's weaknesses and none of it's strengths (relative to Core 2 of course) in our current systems.

I like my Athlon64's. I believe they're REALLY good processors.. I also thoroughly enjoy my Sempron64's (I have 3 of them two of them 2600+ and one of them a 3000+). The only processor I've purchased which left a sour taste in my mouth was my AMD Athlon64 3500+. This because I was coming from a Pentium4 3.2Ghz EE and had gotten accustomed to the multitasking abilities afforded to me by Hyper-Threading.

Athlon64 X2 changed all that for me. It gave me the performance of AMD's K8 with the Multitasking performance I had grown accustomed too with HT (and of course more since I now had two true logical cores in a single physical package).

So when people claim I love Intel and dislike AMD, it's false. My logo underneath my user name changes varying on what my MAIN rig's specs are. Few months ago it was an Athlon64 X2 logo... now a Core 2 Extreme logo. I will never use a companies business practises (unless it hurts the consumer directly such as driver cheating by either nVIDIA or ATi as well as hiring shills to enter forums and fill would be consumers with confusion and mis-information) as a means to win a processor argument. In my view.. performance is what matters and the Core 2 delivers this in spades... per Clock, Per Watt and Overall. It's one heck of a nice processor.

I don't doubt AMD will strike back, and I have a feeling K8L will be more powerful then Core 2 Duo (as it should be coming out a year after). If K8L proves to be superior.. I can almost guarantee the logo underneath my user name will be that of a Quad Core K8L.
September 8, 2006 2:38:37 PM

I totally agree with you. I have ran Intel for many years now and there was a time when I had an AMD K62. Back then it was far superior for gaming, so that's what I went with. For per Core 2 that was still the case, but I became a multitasker and not so much a gamer, so back when the northwoods were the newest that's what I went with, because of the HT. I am not a fanboy of either. I buy what gets me the best VALUE.
!