100 Processors (50 AMD & 50 Intel) compared on matbe.com

[Pt1t]

Before Nehalem launch date , our Belgian / French website compared no more than 100 cpu's :eek:



http://www.matbe.com/images/biblio/art_comparatif-de-100-processeurs/000000077963.jpg
100 Processors Compared (http://www.matbe.com/articles/lire/1121/comparatif-de-100-processeurs/)



50 AMD cpu's socket AM2 VS 50 Intel cpu's socket 775.



http://www.matbe.com/images/biblio/art_comparatif-de-100-processeurs/000000077962.jpg


It's french review but you can read it here : 100 Processors Compared (http://www.matbe.com/articles/lire/1121/comparatif-de-100-processeurs/)


If you cant read french, graphics are still easy to understand !

[gOJDO]

Well done guys! :toast:
:bows: for the time wasted testing so much CPUS!
Keep up the good work and update these charts with the upcoming Nehalem and Deneb CPUs.

Although the benchmark tests are not so detailed, the article is very useful for comparing various CPUs released in the past few years.

[savantu]

Total ownage , that's how it can be summed up.What is interesting is that Netburst era P4s perform better than equivalent rated A64s as SW has become multithreaded , both in apps and games.The old dogs learned a few new tricks , or should we say , can finally use some of its tricks.

[Papu]

Really puts it in perspective , a couple of years ago P4 was standard , nowadays its beaten 3x by the qx9770 :clap:
wonder if in another few years well have that kind of increase or more.

[Nedjo]

bah, another roundup in sea of similar test, but with 100 processors!

I want to see something different, from another perspective!

Here's a really good example:


http://en.hardspell.com/doc/showcont.asp?news_id=4042&pageid=3301


those guys have "simulated" average Joe usability scenario that includes IE opened with bunch of tabs, and Norton AV running in the background. Then they've fired up standard battery of benchmarks to see which CPU will provide better user experience!

I don't understand why more revievers don't adopt similar aproach... I mean what's good of multicore CPUs if not providing better multithreaded/multitaking usage experience?!

Imagine if those auto reviewers only would test cars by connecting engine on dinamometer, and measuring highest horsepower by taping the gas pedal?

[R101]

Total ownage , that's how it can be summed up.What is interesting is that Netburst era P4s perform better than equivalent rated A64s as SW has become multithreaded , both in apps and games.The old dogs learned a few new tricks , or should we say , can finally use some of its tricks.

Back then I think you could see this difference only in Doom3.

[Trouffman]

IMPRESSIVE ! Great work !!

[Olivon]

Great Work from M@tBe !

Congrats :clap:

[alfaunits]

Because noone will use their computer in the exact same way, surely not an average Joe the Plumber, so Mbe's test holds more ground on usefulness.


Here's a really good example:
http://en.hardspell.com/doc/showcont.asp?news_id=4042&pageid=3301

I don't understand why more revievers don't adopt similar aproach... I mean what's good of multicore CPUs if not providing better multithreaded/multitaking usage experience?!

[Slovnaft]

wow, that was a LOT of benches on a lot of chips.
and that thing about amd's most expensive quad core being slower than intel's cheapest? very interesting that it's totally true.
but man, after spending that much time benching i would've taken the pleasure of smashing a few of those procs...

[AndrewZorn]

I'm pleasantly surprised to see it isn't a Tom's Hardware chart, but they actually retested them, had standards, etc... actually useful.

[Shintai]

Very nice list. Not the usual "reuse" of old test with changed things.

[gOJDO]

bah, another roundup in sea of similar test, but with 100 processors!Yes! That's why the article is very useful. Imagine how many hours per person are wasted for collecting all the data. At least you can give a credit to those guys.


I want to see something different, from another perspective!If you want something different then do it yourself.

[duploxxx]

wow, that was a LOT of benches on a lot of chips.
and that thing about amd's most expensive quad core being slower than intel's cheapest? very interesting that it's totally true.


amd's most expensive quad has more or less equal price to intel's cheapest....


for the differences between dual and quad core that can only mean that the ratio between single and multithreaded benchmarks was not in line with eachother.

not to mention the fact that you run more then 1 app at the same time but that's a discussion going on for a while already.

[STaRGaZeR]

that can only mean that the ratio between single and multithreaded benchmarks was not in line with eachother.

Single or multithreaded apps affect both quads. And this is not servers. With more multithreaded apps the gap would have been even larger.


not to mention the fact that you run more then 1 app at the same time but that's adiscussion going on for a while already.

Let me guess, slower but smoother?

[duploxxx]

Single or multithreaded apps affect both quads. And this is not servers. With more multithreaded apps the gap would have been even larger.



Let me guess, slower but smoother?

i read his post the wrong way, corrected

i was actually referring to the fact that a e7400/e8200 gets a higher total score then certain quad cores.

[Revv23]

wow this is quite a feat!

That must have been so boring!

[Sheik]

Wow, thats impressive!

[savantu]

i read his post the wrong way, corrected

i was actually referring to the fact that a e7400/e8200 gets a higher total score then certain quad cores.

I'm pretty sure that in most desktop scenarios an E8200 will perform better than a tricore for example.

[Xope_Poquar]

100 processors in 21 tests. That's 2100 benches. Ugggggh.

Great roundup! The only thing more I would have asked for is a final graph comparing average performance versus current price.

[Kingcarcas]

This is the test i've been waiting for:banana:

[NightCrawler™]

One of the best charts I've seen in a very long time....

/me tips hat...

[azraeel101]

I only cared about gaming benches in those tests, but when I saw 640*480 "low details" I closed the tab and didn't read the rest of the review. For me at least, it's absolutely useless.

[Mad1723]

Really.... amazing.... work! And if you can read french, the article is really interesting too :banana::)

[savantu]

I only cared about gaming benches in those tests, but when I saw 640*480 "low details" I closed the tab and didn't read the rest of the review. For me at least, it's absolutely useless.

Huh ? Both Word in Conflict and FEAR are tested at low/high details.And guess what , it proves that low detail testing is every bit as relevant.It's about CPUs , not GPUs.

[inCore]

Was waiting for somebody to bring that up, the "low detail" settings. Obviously it was done to minimize the effect of the GPU on the benchmarks.

If you're looking out for a new CPU you can find the benchmarks somewhere anyways, doing this testing to suit everybody's desires at once would be like trying to stop a speeding steam train with your arms stretched out in front of you.

[Solus Corvus]

What is interesting is that Netburst era P4s perform better than equivalent rated A64s as SW has become multithreaded , both in apps and games.The old dogs learned a few new tricks , or should we say , can finally use some of its tricks.
I hope you are joking. This review doesn't make netburst look any better in comparison to A64 then it did back in the day. Netburst gets it's ass handed to it in almost every test.

Not that it really matters to me since I won't be using either architecture ever again.

[savantu]

I hope you are joking. This review doesn't make netburst look any better in comparison to A64 then it did back in the day. Netburst gets it's ass handed to it in almost every test.

Not that it really matters to me since I won't be using either architecture ever again.

Huh ?

Back in the days an A64 3000+ constantly beat a P4E 3000MHz in a wide selection of benchmarks.Today it failed to beat a P4E 2800MHz in a single benchmark.

Short memory I guess.

[Solus Corvus]

Never the less, the top netburst still gets raped by a whole raft of A64s in most of the tests. I could care less about the ridiculous ratings.

[xsbb]

Excellent! This info is really useful for a quick overview of old processors!

[azraeel101]

Was waiting for somebody to bring that up, the "low detail" settings. Obviously it was done to minimize the effect of the GPU on the benchmarks.

If you're looking out for a new CPU you can find the benchmarks somewhere anyways, doing this testing to suit everybody's desires at once would be like trying to stop a speeding steam train with your arms stretched out in front of you.

I said it's useless for me, maybe others play at 640*480 and will find this review useful. Unfortunately there are few benchmarks around comparing CPU's at high resolutions/settings, that's why I'm disappointed.

[adamsleath]

e5200 @ 2.5GHz beats AMD 6400+ @ 3.2GHz :rofl: (F.E.A.R. test)

..but on par with phenom x4 @2.6

[HuffPCair]

Lol Intel kicked the whole test. Kind of sad I wish there would of been more comp from AMD in that, but o well. Im just happy my 6600 stayed on top for the most part specially in the gaming area is all I was worried about.

[gallag]

I could care less

So it is important to you, Personally, I couldn't care less:D

[demonkevy666]

they have a spartan as an 8850 BE chip with one core.... :/

[naokaji]

How to condense that test of 100 cpu's:
45nm C2Q = Win.

Positive surprise, the Q8200, didnt expect it to do so well compared to the "old" C2Q's, of course that doesnt change the problem with the far too low multi when it comes to oc.

Negative surprise, phenom, I did expect them to loose, but not that badly.

[munim]

Nice, considering my E6300 at 3.15GHz is comparable to an X6800 I'm pretty happy!

[dogsx2]

Good job, thanks......

Watching the bottom 25% do Cinebench 10 must have been like watching paint dry. :(

[Glow9]

Good stuff

[JumpingJack]

I said it's useless for me, maybe others play at 640*480 and will find this review useful. Unfortunately there are few benchmarks around comparing CPU's at high resolutions/settings, that's why I'm disappointed.

Actually, it should be helpful and you should not be disappointed, it is informative data. Unless you build a completely new rig every new generation, knowing what constraints the CPU will put on a game is important in making decisions.

You will not find as many high res/settings benchmarks/reviews with the sole focus to compare CPUs since most graphics cards will give you the same result under high res/high quality settings ... all one would see is FPS cap due to the GPU running under your 'realworld conditions' ... no amount of CPU processing horse power will change the performance if the GPU is the limiter. The recent refresh of GPUs this is not always true however, even at high res/high settings, a 4870 X2 is capable of out running the CPU.

Increasing resolution and visual settings actuates the GPU, not the CPU. Running a bench at ubber settings hides any conclusion you could make regarding the CPUs ability to execute game code.

Here is a review more your liking I suspect:
http://legionhardware.com/document.php?id=770

Why, then would one want to see these types of number? Well, if, for example, my build strategy is to put together a system that I can upgrade to the latest GPU in the future ... I would want to know how the CPU may cap a game ... to see that, remove the bottleneck from the GPU and put it on the CPU to see that information.

[dinos22]

so you are saying you tested all these on one testbed
why are the CPUz screens all different versions then........really old ones too
seems as though old info was recycles from previous tests.....?!?!?

it's a nice summary though...lots of work involved

[Pt1t]

so you are saying you tested all these on one testbed
why are the CPUz screens all different versions then........really old ones too
seems as though old info was recycles from previous tests.....?!?!?

it's a nice summary though...lots of work involved

Old CPUZ screenshot are there to illustrate historical part of review. ;)

[Stephane@matbe]

so you are saying you tested all these on one testbed
why are the CPUz screens all different versions then........really old ones too
seems as though old info was recycles from previous tests.....?!?!?

it's a nice summary though...lots of work involved


These CPU-Z screens are there to illustrate. I have taken it from older articles I have written in the past because I did not have time to take (and forgot mostly) cpuz screens while testing for this 100 CPU roundup.
I can assure you that all 100 CPUs have been reviewed on the same motherboards with same configuration and same testing procedure. I have started this back in may 2008 slowly and these last three months I have done intensive CPU testing.

It is always strange to see people that do not believe when there is something unusual done. I do not know you and you do not know me. If you would have known me, you will know that this is not the way I am working. I always want to be the most precise and accurate in my reviews. ;)

So, I repeat, everything has been tested on the same testbenchs (2 configurations, one for AMD and one for Intel of course) from CPU 1 to CPU 100. :)

[T_Flight]

These CPU-Z screens are there to illustrate. I have taken it from older articles I have written in the past because I did not have time to take (and forgot mostly) cpuz screens while testing for this 100 CPU roundup.
I can assure you that all 100 CPUs have been reviewed on the same motherboards with same configuration and same testing procedure. I have started this back in may 2008 slowly and these last three months I have done intensive CPU testing.

It is always strange to see people that do not believe when there is something unusual done. I do not know you and you do not know me. If you would have known me, you will know that this is not the way I am working. I always want to be the most precise and accurate in my reviews. ;)

So, I repeat, everything has been tested on the same testbenchs (2 configurations, one for AMD and one for Intel of course) from CPU 1 to CPU 100. :)

:clap: Great First Post, and :welcome: to XS. :yepp:

[savantu]

These CPU-Z screens are there to illustrate. I have taken it from older articles I have written in the past because I did not have time to take (and forgot mostly) cpuz screens while testing for this 100 CPU roundup.
I can assure you that all 100 CPUs have been reviewed on the same motherboards with same configuration and same testing procedure. I have started this back in may 2008 slowly and these last three months I have done intensive CPU testing.

It is always strange to see people that do not believe when there is something unusual done. I do not know you and you do not know me. If you would have known me, you will know that this is not the way I am working. I always want to be the most precise and accurate in my reviews. ;)

So, I repeat, everything has been tested on the same testbenchs (2 configurations, one for AMD and one for Intel of course) from CPU 1 to CPU 100. :)

Welcome and congratulations for your work.It is outstanding. :welcome:

[dinos22]

These CPU-Z screens are there to illustrate. I have taken it from older articles I have written in the past because I did not have time to take (and forgot mostly) cpuz screens while testing for this 100 CPU roundup.
I can assure you that all 100 CPUs have been reviewed on the same motherboards with same configuration and same testing procedure. I have started this back in may 2008 slowly and these last three months I have done intensive CPU testing.

It is always strange to see people that do not believe when there is something unusual done. I do not know you and you do not know me. If you would have known me, you will know that this is not the way I am working. I always want to be the most precise and accurate in my reviews. ;)

So, I repeat, everything has been tested on the same testbenchs (2 configurations, one for AMD and one for Intel of course) from CPU 1 to CPU 100. :)
fantastic man :up::up::up:

[Stephane@matbe]

Thank you all :)