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

What are benchmarks for then ?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

trickson

Member
Joined
Nov 30, 2004
Benchmarks show how well your computer will perform in a given circumstance right ?
Well then when you ask the question what is better for games a Quad Core or a dual core why is it that some are more than adamant about dual cored CPU's when every benchmark I have seen shows that the Quad Core CPU kills the dual core ..??
No matter what I have the E8200 and I clocked it to 3.7GHz the ONLY benchmark it "won" was Super Pie 1.5 mod . every thing els was a wash out even the benchmarks in my games was averaging lower FPS than with the Quad Core Q6600 ! So when you say the the games do not use all 4 Cores I have to say then if this is really true why is it I get higher FPS with the Q6600 then with the E8200 ? Explain that ...
:confused:

I hope that there can be some good talk and no fanboy bashing . This is just what I have seen with all the test site out there X-bit Labs , Tom's Hardware and other's that test CPU's for us you see this as well , every thing is faster FPS in games , rendering and encoding all of it .
 
Biggest thing is the L2 cache difference...Intels thrive on the l2 cache unlike AMD...the bigger the L2 the better the performance...

The E8---'s have a 6mb cache while the quads have 12mb cache...this is the main reason why the quads can beat the duals in most benchmarks excpet SP...SP uses CPU speed to process the data...the L2 is hardly used...
 
In multithreaded benchmarks, the quad will obviously win (compared to a similar dual). GPU drivers are multithraeded too so a dual will always beat a single-thread cpu for 3d benches and games.

In the benches we do on the benching team,

A quad will beat a dual in: wprime, PCMark05, 3DMark06, vantage
A dual can be better (because of higher clocks) in 3DMark01, 3DMark03, 3DMark05, SuperPI, pifast.


Biggest thing is the L2 cache difference...Intels thrive on the l2 cache unlike AMD...the bigger the L2 the better the performance...

The E8---'s have a 6mb cache while the quads have 12mb cache...this is the main reason why the quads can beat the duals in most benchmarks excpet SP...SP uses CPU speed to process the data...the L2 is hardly used...

Actually, SuperPI is extremely cache-dependant. Cache can mean the difference between more than a few seconds in 1m time.

Wouldn't the 12mb be split between each pair of cores so each pair only has 6mb available to it? ;)
 
Biggest thing is the L2 cache difference...Intels thrive on the l2 cache unlike AMD...the bigger the L2 the better the performance...

The E8---'s have a 6mb cache while the quads have 12mb cache...this is the main reason why the quads can beat the duals in most benchmarks excpet SP...SP uses CPU speed to process the data...the L2 is hardly used...
Ok, you mention cache. That is a point I didn't think of.

The main two reasons people say to get the dual core is because you can overclock them more and most games are not multi-threaded. That will beat cache any day :)
 
In game benchmarks where its FPS not some synthetic gerneration, the dual excel because of their ability to clock higher then the quadcores.

Only games I know of that give you better frames with quad are supcom and FSX (with patch), Crysis is supposed to but I do not recall actually seeing any improvement.
 
In game benchmarks where its FPS not some synthetic gerneration, the dual excel because of their ability to clock higher then the quadcores.

Only games I know of that give you better frames with quad are supcom and FSX (with patch), Crysis is supposed to but I do not recall actually seeing any improvement.

No if this was true than ever single test that X-bit labs and the like do for "US" would show the dual core CPU smacking the Quad core up side the chip . this is not happening and when I run the benchmarks in F.E.A.R. and lost planet and HL2 I would have also see this AND I DO NOT my Quad smacks the E8200 clocked t o3.7GHz right up side the Chip ! and this is what I do not understand you all scream dual core for games when the facts seem to point to quad core even if the game can not supposedly use the quad core . every thing I have seen points to the Quad as the better choice .
 
No if this was true than ever single test that X-bit labs and the like do for "US" would show the dual core CPU smacking the Quad core up side the chip . this is not happening and when I run the benchmarks in F.E.A.R. and lost planet and HL2 I would have also see this AND I DO NOT my Quad smacks the E8200 clocked t o3.7GHz right up side the Chip ! and this is what I do not understand you all scream dual core for games when the facts seem to point to quad core even if the game can not supposedly use the quad core . every thing I have seen points to the Quad as the better choice .

I think the confusion comes into play when thinking benchmarks and games are the same when, if fact, bechmarking programs are far far more intensive than games are currently. Benchmarks are designed to push the system to the limits and back again. Games are not geared for totally stressing out your system so you will see an improvement with a dual over a quad in benchamarks......make sense? That's my understanding anyway, If I'm wrong, somebody please correct me.
 
Last edited:
No if this was true than ever single test that X-bit labs and the like do for "US" would show the dual core CPU smacking the Quad core up side the chip . this is not happening and when I run the benchmarks in F.E.A.R. and lost planet and HL2 I would have also see this AND I DO NOT my Quad smacks the E8200 clocked t o3.7GHz right up side the Chip ! and this is what I do not understand you all scream dual core for games when the facts seem to point to quad core even if the game can not supposedly use the quad core . every thing I have seen points to the Quad as the better choice .
Can you link one then? From reading reviews and from personal experience, quad cores do worse when compared to higher clocked duals.
 
Can you link one then? From reading reviews and from personal experience, quad cores do worse when compared to higher clocked duals.

http://www.xbitlabs.com/articles/cpu/display/core2quad-q6600.html

I see what you are saying but even with a Q6600 and higher over clocked Quads you can see that when Over Clocked you can get more performance than with just a dual core CPU even in games (granted this is only when Over Clocked . )
I see that in some apps and games where the dual core is a little bit better and that in it's self speaks volumes as well but even then Quads are clearly the better choice all around and well there is very little that can take down a
http://www.xbitlabs.com/articles/cpu/display/core2extreme-qx9650.html
These are just monster CPU's .
 
That is honestly the first one I've seen where they show that. I'm confused though, the numbers don't seem right. Why would the Q6600 beat the e6850 in applications that support single or dual threaded? They are the exact same processor, the Q6600 just has more.

I don't know, their results just don't seem right. I'm going to drop it at that since we will never be able to discuss it without having a really huge thread. :p
 
Last edited:
OK then what are benchmarks for ? I mean if we can see a Quad core CPU performing better then what is verdict ? you can not just dismiss benchmarks or there would be no frame of reference to guide us as to how well any given computer performs . AM I right ?
 
OK then what are benchmarks for ? I mean if we can see a Quad core CPU performing better then what is verdict ? you can not just dismiss benchmarks or there would be no frame of reference to guide us as to how well any given computer performs . AM I right ?
They give basic information on how good a computer will run under certain conditions. I think more people use it for a hobby (benching members, myself included) or even competition.

It is fun to tweak the system to get more performance out of it. Almost a game within itself.
 
They give basic information on how good a computer will run under certain conditions. I think more people use it for a hobby (benching members, myself included) or even competition.

It is fun to tweak the system to get more performance out of it. Almost a game within itself.

Then it is better to go with a Quad core CPU .
 
Then it is better to go with a Quad core CPU .
Not sure why you keep arguing on this. We have established that you should go with a dual/quad depending on what you will use it for. In some cases, the dual is better (like for WoW since it is heavily CPU dependent). In some cases, the quad is better.

There is no "x is better" comment that is valid regarding dual/quad cores.
 
Not sure why you keep arguing on this. We have established that you should go with a dual/quad depending on what you will use it for. In some cases, the dual is better (like for WoW since it is heavily CPU dependent). In some cases, the quad is better.

There is no "x is better" comment that is valid regarding dual/quad cores.

I am not sure you can say this is arguing at all . I have asked a very poignant question , I feel needs to be answered as there are tons of people that use benchmarks to gauge there computers and validate there power to there user .
And since there is much to do talk about what people should go with and dual cores are often the focal point of what people tell some one to get I mearly wanted to know what the point of benchmarking is then .
See if a Quad Core CPU is better 99%^ of the time then what is the point in telling some one to get a dual core CPU ? seems to me if you tell some one to just look at the benchmarks of any given Quad Core CPU they would see the benefits and draw backs as well right ?

Then for you to say to me I am arguing the point is not true . I am only trying to get and give facts to others .
 
Benches are really just for your own self satisfaction. And to show the world we can spend as much as often as anybody to achieve higher numbers!!!

My 12 year old son runs a P5N-e SLI with a E6600 running at around 2800 mhz he has 2 gigs of G.Skill 6400HZ running around 800mhz and a 3870 vid card.

That is running stock. He goes on call of Duty on line and using just the lowest pistol can headshot everyone in the free for all section and almost always win every damn round and is never out of the top three. He can use only the knife and alway end up in the top 3.

his system is nowhere near the top of the heap but he is on every game. It is maddening!!!!

WeldZilla
 
Last edited:
Benchmarks show how well your computer will perform in a given circumstance right ?

Wrong. Benchmarks show how well your computer will perform under a HYPOTHETICAL circumstance. Even in-game benchmarks like the Crysis benchmark aren't actually going to tell you how well you're going to be able to play Crysis. The Vista benchmark isn't going to actually tell you how well you're going to be able to run Vista. All these benchmarks can tell you is how well your hardware performs with that particular benchmark.

Performance is something you have to gauge for yourself. You can't trust some random number more than you can trust your own two eyes and hands.

The fact is, many games simply aren't optimized to use quadcore yet. Most games have just FINALLY gotten around to using dualcore CPUs. And there aren't even that many of those (comparatively speaking).

So if you can get your E8400 or whatever up to 4.5ghz and you can only get your Q6600 up to 3.5ghz... then the E8400 is going to beat the daylights out of it. And that's because the benchmarks may be optimized to use Quad now... but the games are not.
 
Back