AMD Gets a Beating, Intel Takes Performance Crown

Aggressor Prime · Post by **Aggressor Prime** » Tue Mar 07, 2006 3:39 pm

Ever since the launch of the K8, AMD has made huge progress in the CPU industry. They have had the gaming crown for 3 years now. Well, today, Intel slaps them silly. Intel's Conroe not only beats the FX-60, but the upcoming FX-62 (at least an overclocked FX-60 that resembles the FX-62) by a large margin. More here.

Intel now has the best IPC (instructions per clock) and performance CPU.

fliptw · Post by **fliptw** » Tue Mar 07, 2006 3:43 pm

don't forget to disinfect your hand.

and then wake me when there is a comparison using a AM2 system - if you didn't notice, the intel-provided system is using DDR2, and the AMD system is not. regardless of what the conclusion says, it wasn't a fair test.

MD-2389 · Post by **MD-2389** » Tue Mar 07, 2006 3:55 pm

The very fact that the x2 stayed pretty close to the intel chip running faster RAM says something.

Aggressor Prime · Post by **Aggressor Prime** » Tue Mar 07, 2006 4:42 pm

First, there is a 40% differance in games like FEAR.
Second, you must remember that AM2 adds only DDR2 support which with higher latencies won't give much of a performance boost.

MD-2389 · Post by **MD-2389** » Tue Mar 07, 2006 7:42 pm

Yes, a 40% difference between a DDR1 based system and a DDR2 based system. That is NOT a true comparison. Thats like saying a system with PC133 is faster than a system with PC100. Of COURSE it is! What we're saying that its an unfair comparison between the two because they aren't comperable systems. Lets see what happens when we're comparing apples to apples instead of apples to oranges.

BUBBALOU · Post by **BUBBALOU** » Wed Mar 08, 2006 6:28 am

Zaphod Beeblebrox... figures

Immortal Lobster · Post by **Immortal Lobster** » Wed Mar 08, 2006 7:31 am

your also comparing a CPU that has an iintegrated memory controller which mkaes the RAM speed mostly independent from the CPUs overall performance. Ive heard rumors that intel used only one core on the AMD FX60 though, wheres they used 2 on their conroe. I believe it seeing as they can release that mother of all FUD document.

Besides, Conroe is 6months away, at earliest, and its intel, so its really 8-10months away

MD-2389 · Post by **MD-2389** » Wed Mar 08, 2006 1:15 pm

BUBBALOU wrote:Zaphod Beeblebrox... figures

Thats not him. Trust me, I checked a long time ago. Thats ccb056's brother.

Top Wop · Post by **Top Wop** » Wed Mar 08, 2006 8:10 pm

Doesnt matter who it is, poster is still retarded.

Ferno · Post by **Ferno** » Wed Mar 08, 2006 8:42 pm

this is like comparing a ferrari with a tempo.

Immortal Lobster · Post by **Immortal Lobster** » Wed Mar 08, 2006 9:56 pm

VooDoo PCS Rahul has this to say, and it all puts intels latest PR stunt in perspective. and I was right, it falls right in line with thier mother of all fud document

http://voodoopc.blogspot.com/2006/03/if ... lytech.com

Matrix · Post by **Matrix** » Thu Mar 09, 2006 5:26 pm

You guys all missed the most important thing.... They are comparing an Intel CPU that doesn't come out for another 6 months to an AMD CPU that has been out. You don't think AMD will have anything to match this in 6 months?

You know AMD is back in there labs reading this, laughing like crazy \"Yeah you show off your stuff intel, in 6 months when you release that we'll hit you with our new FX-85! Quad Core 45nm!! what?!!\"

Also the AMD system in that review is using an ATI chipset, they are known to blow ass for AMD systems.

Iceman · Post by **Iceman** » Thu Mar 09, 2006 5:53 pm

Top Wop wrote:Doesnt matter who it is, poster is still retarded.

I must concur

Read his intro post then read Matrix's post ... no further explaination needed.

Duper · Post by **Duper** » Thu Mar 09, 2006 6:37 pm

really thought, on the whole, the market has stayed really competitive which is great for us. I don't think that either company has dominated the other for more than a month or two for quite some time. (if ever) And personally, I hope it stays that way. Means better products for us.

Vindicator · Post by **Vindicator** » Thu Mar 09, 2006 9:46 pm

Indeed. By the time that chip comes out I'll probly be lookin to upgrade and I might just be able to skip dual core altogether and jump to quad-core

Neo · Post by **Neo** » Thu Mar 09, 2006 10:30 pm

Maybe if you AMD fanboys actually read the article, you'd know that it's just a performance preview; Anand already mentioned that they didn't compare Conroe to an AM2 part. 9_9

Besides, that's only a beta version of a high end Conroe, it's not an extreme edition.

Another thing to note is that the performance gap is only 20% in F.E.A.R. Anand found some mistakes in his tests and published a new article.

Duper · Post by **Duper** » Thu Mar 09, 2006 11:01 pm

it doesn't really matter anyways. In a year or so, it will be obsolete. in 3, it will be laughable.

*shrug*

Jeff250 · Post by **Jeff250** » Fri Mar 10, 2006 2:53 am

AMD Gets a Beating, Intel Takes Performance Crown

I'm actually more offended by the use of a mixed metaphor.

JMEaT · Post by **JMEaT** » Fri Mar 10, 2006 8:42 pm

Any time I see a Pentium vs AMD or an IE vs Mozilla thread I instantly think Mobius. I was wrong this time :O

Aggressor Prime · Post by **Aggressor Prime** » Fri Mar 10, 2006 9:03 pm

Yes, the benchmark was flawed with an old bios that only gave the Athlon 64 one core. Here are the tests with the new bios.

Conroe still shows a 19-31% advantage. Now how will AMD respond? You are all saying that the AM2 will be super fast and own Intel. Have you even been in touch with AM2 details? AM2 is just a DDR2 mod. That is it. DDR2 only shows a performance increase at 800MHz. 667MHz DDR2 = 400MHz DDR. Therefore, AMD will have some improvement with the 800MHz DDR2. But that will be a small improvement. In order of how they affect graphical performance: GPUs, CPUs, RAM. With that in mind, Intel had a huge change in performance. They are now ahead 19-31%. You don't see that much of an advantage with a CPU upgrade. That is what you see when you upgrade your GPU. Therefore, this is something powerful and it will take more than a DDR2 change to help AMD. Then comes the question of scalability. You think AMD can scale GHz to well beyond 3GHz with quad cores and 45nm. Again, you are not with the times. If you would have studied the market, you would know that AMD will only launch one FX this year. That FX is the Athlon 64 FX-62 of whose details are already known. Besides being a Socket 940 product (AM2), it has a clock speed of 2.8Ghz and is dual cores. AMD will not release another FX until next year, that is the FX-64, a dual core 3.0GHz FX. AMD will also not have 65nm CPUs ready until the end of this year. At first, their will be 65nm Opterons with the possiblity of quad cores this year. Therefore, the FX-64 can be moved back to late 2006 and have quad cores to compete with Conroe. But there is still one problem. Games only utilize up to quad cores for 2006. Well, we have quad cores, but they are low quality quad cores. They only have a 3-issue core. Here is a quick calculation of how it would perform: (IPC x clock x cores) 3x3.00x4=36. By this, AMD would be more powerful than Intel. Of course this is an assumption that AMD accelerates their existing plans and modifies them a little. Previously, they said we shouldn't see quad cores for desktops until mid 2007. If they can make that radical change, Intel can too, for they have quad core plans for all markets for 2007 as well. But let us look at this from another aspect, using server CPUs for gaming. You could easily use dual Opterons which have dual cores for gaming. However, Intel can also use 2 of their dual core Conroe Xeons. Again, Intel wins. Let me recap by showing you a differance of calculated performances based on IPC x clocks x cores with a maximum of 4 core solution:
Athlon 64 FX-62 (as planned): 16.8
Athlon 64 FX-64 (as planned): 18
Athlon 64 FX-64 (quad core - made up): 36
Conroe 2.66GHz (as planned): 21.28
Conroe XE 3.33GHz (as planned with a late 2006 launch): 26.64
Conroe XE 3.33GHz (quad core - made up): 53.28
Dual Opteron x90 (2.8GHz clocks, dual core): 33.6
Dual Opteron x95 (3.0GHz clocks, dual core): 36
Dual Xeon 5160 (3.0GHz, dual Conroe core): 48

As you see, Intel has AMD beat in all of these calculated solutions which the present benchmarks prove. Now I am not an Intel fan and love AMD like crazy, so how can I benefit from this? Duh, Athlon XP age anyone? AMD will now be forced to slash prices like crazy. I have calculated, so that AMD can maintain a better price/performance ratio, that the Athlon 64 FX-62 will have to be a maximum of $400, preferably $350. This means good news for people who still have Athlon XPs and want to upgrade and bad news for those who already got into the K8 business by the fact that their CPUs just became worthless. Overall, it is still a very positive thing for all of us. Of course you can expect AMD will want to rise again as soon as possible. Since 2006 is already well planned, the chances of AMD just making radical changes to try to beat Intel will not happen. It will not be financially suitable. AMD, however, can plan for 2007 as their redemption year much like how nVidia will probably have 2007 as their redemption year. In 2007, it would be best to fix CPUs to have a 6-issue core. By this, AMD will have some breathing room and will be able to take the 3.33GHz XE Conroe with a 2.4GHz FX. It will also put AMD in the position that they were before (2-Intel:3-AMD IPC ratio).

Conclusion: This year Intel will take the performance crown causing AMD prices to drop like crazy. AMD will keep the best price/performance ratio. It will be a short run K7 period. AMD will likely come back with a top performer next year. This is competition in a capitalist society, so love it!

Krom · Post by **Krom** » Fri Mar 10, 2006 9:40 pm

Aggressor, you can and often do see that kind of gaming peformance boost from a CPU upgrade if you are CPU limited. Not all games push only the GPU, if I wasn't clocking my Athlon XP 1700+ at 2.4 GHz this 6800 GT would be no faster then my old 5900 Ultra was unless I cranked them both all the way to my monitors maximum 1920x1440 resolution. In Descent3, this 6800 GT card @ ultra clocks running 1024x768 is no faster then my 5900 Ultra was. I have to run them at 1280x960 with 4x FSAA before the 6800 GT puts any distance between it and the FX, thats a CPU limit, I can't get more then about 480 FPS average timetest at this CPU clock in D3 no matter what video card I have. If I had a CPU with double the power, this 6800 GT card could easly crank out a 900+ FPS average timetest, the FX would probably hit the high 500-600 range. A slow CPU can drag down a game just as much as a slow GPU. Those benchmarks don't even tell the full story either, that 19-31% advantage could still be GPU bound, the conroe chip could have even more performance room then that.

Also, neither AMD or Intel's release schedules are set in stone, Conroe might be early or late, AMD might have something else up ther sleeves, either way as long as they compete equally or near equally it's all good. The big thing I don't really want to see is one side taking a huge lead over the other, I don't feel like putting up with the Intel tax from anyone.

Aggressor Prime · Post by **Aggressor Prime** » Fri Mar 10, 2006 10:27 pm

So you are saying anything could happen? That is the general response from AMD Forums too. We still have 6 months. But its always good to be prepared.

fliptw · Post by **fliptw** » Fri Mar 10, 2006 10:57 pm

I'd like to see a comparison where both systems are using nforce4 chipsets.

Immortal Lobster · Post by **Immortal Lobster** » Sat Mar 11, 2006 2:25 pm

We've got months for conroe to be released, thats a normal 8 months, so it will be released just in time for the xmas season, most likely it will be expensive...and ...well...expensive. AMDs architecture for socket F I though was supposed to utilze RDRAMs X-Ram. which should boost the cache response in the newer opterons, hopefully, and I dont see why it wouldnt, that xram will find its way back to the AM2 Athlon CPUs. In the past 2 years, I cant remember when AMD didnt release something spectacular right after Intel thought they invented the fastest thing in the world, we'll see the outcome of this in October-November.

besides, not like I could afford any of those possible solutions when that arrives, Ill be using my opteron until next christmas...if im lucky.

Aggressor Prime · Post by **Aggressor Prime** » Sat Mar 11, 2006 11:32 pm

Socket F will utilize DDR2. I don't know if they will utilize FB-DIMM right off the back. They probably will though. FB-DIMM allows the connection with any memory type. I think DDR2, DDR3, XDR (FB-XDIMM), and XDR2 (FB-XDIMM) are FB-DIMM ready memories. That is why AMD made that deal with Rambus. Although AMD can use the 8GHz FB-XDIMM, I don't know if they can utilize all that power. And the onboard RAM is called ZRAM. Although these are great latency reducers, there is no latency that I can see that needs reducing except in the 8P+ setups. This might mean that AMD plans to release a higher IPC core, but I don't think it will be socket F summer launch since too much info is already known. They might just move the 2007 products to late 2006 to give us an early 6-issue core.

Duper · Post by **Duper** » Sun Mar 12, 2006 1:16 am

Krom wrote:Also, neither AMD or Intel's release schedules are set in stone, Conroe might be early or late, AMD might have something else up ther sleeves, either way as long as they compete equally or near equally it's all good. The big thing I don't really want to see is one side taking a huge lead over the other, I don't feel like putting up with the Intel tax from anyone.

X2 on both counts.

Personally, I would rather see a product late out on the market than it hit the shelves with some wierd obscure glitch. ... I still haven't forgotten the pentium 90.

Mr. Perfect · Post by **Mr. Perfect** » Mon Mar 13, 2006 1:24 pm

MD-2389 wrote:The very fact that the x2 stayed pretty close to the intel chip running faster RAM says something.

That's a decent point, but A64s arn't memory bandwith limited. There is a comparison between an 939 A64 and a AM2 A64 engineering sample here. It's in German, so you'll need babelfish or something. The AM2 is only using DDR2 667 while retail ones should have DDR2 800, but considering how it provides no benifit over the DDR400 you get a good idea of how little it cares about bandwith. It actually has taken a small hit from the fucked up latencies of DDR2.

Also consider that the FX was at 2.8GHz, while the intel was at only 2.66GHz. This reminds me somewhat of when we all laughed at Intel when the lower clocked A64s where whooping up the netburst architecture.

If AM2 is really so pedestrian as that early sample, and Conroe is at least as good as it's early sample, I may buy my first Intel somewhere down the line. Brand loyalty can kiss my ass.

Neo · Post by **Neo** » Wed Mar 15, 2006 3:40 pm

Mr. Perfect wrote:Brand loyalty can kiss my ass.

I second that motion. =P

Although, I was never loyal to AMD...A lot of people just act so fanboy-ish. =P

BTW, "Aggressor Prime," you called the server CPU based on the Core micro-architecture a "Conroe Xeon"...heh...the server CPU is code-named Woodcrest. ^_~

Also, Core CPUs use a 4-issue core, so your numbers should be different.

BUBBALOU · Post by **BUBBALOU** » Thu Mar 16, 2006 5:10 am

Neo.............................................

Your twin brother from another mother, Aggressor Prime copies info from websites verbatim.... so if there are typos he copies the typo's. No original content or views ...boring

Aggressor Prime should have his nick changes to \"Aggresive CTRL-C\"

Aggressor Prime · Post by **Aggressor Prime** » Thu Mar 16, 2006 6:14 am

But I copied the name from Descent 4...

MD-2389 · Post by **MD-2389** » Thu Mar 16, 2006 11:05 am

BUBBALOU wrote:Neo.............................................

Your twin brother from another mother, Aggressor Prime copies info from websites verbatim.... so if there are typos he copies the typo's. No original content or views ...boring

Aggressor Prime should have his nick changes to "Aggresive CTRL-C"

What do you expect from someone that sucks on the teat of all the BS that spews from Tom's Hardware?

Neo · Post by **Neo** » Thu Mar 16, 2006 3:34 pm

BUBBALOU wrote:Neo.............................................

Your twin brother from another mother, Aggressor Prime copies info from websites verbatim.... so if there are typos he copies the typo's. No original content or views ...boring

Aggressor Prime should have his nick changes to "Aggresive CTRL-C"

NGA, pleaze. 9_9

BUBBALOU · Post by **BUBBALOU** » Thu Mar 16, 2006 8:39 pm

Neo Prime

Isaac · Post by **Isaac** » Sat Mar 18, 2006 1:33 pm

my amd gives me super powers. Intel can't.

Krom · Post by **Krom** » Sat Mar 18, 2006 1:39 pm

Nah, I'm pretty sure the super powers come from the race of alien space invaders you could hide in that wig.

Isaac · Post by **Isaac** » Sat Mar 18, 2006 5:28 pm

they know...

MD-2389 · Post by **MD-2389** » Sat Mar 18, 2006 8:00 pm

Hey, maybe that AMD chip gave you that bald spot for a superpower chip.

Neo · Post by **Neo** » Mon Mar 27, 2006 12:28 pm

BUBBALOU wrote: Neo Prime

whateva, bubbacookiecutter =P

Isaac wrote:my amd gives me super powers. Intel can't.

heh...Well, apparently, Conroe can sm0ke that dual-core X2 that you have there. 9_9

Ferno · Post by **Ferno** » Mon Mar 27, 2006 1:52 pm

Neo wrote:
BUBBALOU wrote: Neo Prime
whateva, bubbacookiecutter =P

this coming from someone who is not with the matrix, but thinks he is.

Diedel · Post by **Diedel** » Wed Mar 29, 2006 7:12 am

AP,

you trust a speed comparison conducted by one of the competitors? Some ppl seem to need some external device to make up for their lack of self confidence, don't they? Some use cars, you need a CPU.

AMD Gets a Beating, Intel Takes Performance Crown

AMD Gets a Beating, Intel Takes Performance Crown

Re:

Re:

Re:

Re:

Re:

Re:

Re:

Re:

Re: