Q6600 isn't real quad?

[TechnologyCoordinator]

Why?

Why what?

 

[Guest]

four cores on one die will always be better than duck tapping two relic Dual cores together

Oh of course it will always be better....except when it isn't. The theory may be there, the marketing blasting away, but the performance is sorely lacking. I suppose if you're not counting performance it might be better, but I'm afraid most businesses and enthusiasts want something that yeilds performance rather than promises it under different scenarios which aren't really applicable as they simply haven't come about. By the time these theoretical problems begin to spin Phenom ahead of the Core 2 Quads they'll both already be outmodded and outdated peices of junk compared with the designs of that future.

Besides, didn't you make a lot of hot air about Vista's NUMA awareness being a major advantage for the Phenom processor over Windows XP? That turned out to make squat all difference. How do we know all these visions you're making won't simply fall flat on their face as well? 4x4 was supposed to be great too, haven't seen anything from that in 18 months, guess that still has growth room for the future? I think AMD just realised it was a scrapheap and simply cancelled all development on the 4x4 platform. True potential indeed.

 

[TechnologyCoordinator]

Oh of course it will always be better....except when it isn't. The theory may be there, the marketing blasting away, but the performance is sorely lacking. I suppose if you're not counting performance it might be better, but I'm afraid most businesses and enthusiasts want something that yeilds performance rather than promises it under different scenarios which aren't really applicable as they simply haven't come about. By the time these theoretical problems begin to spin Phenom ahead of the Core 2 Quads they'll both already be outmodded and outdated peices of junk compared with the designs of that future.

Besides, didn't you make a lot of hot air about Vista's NUMA awareness being a major advantage for the Phenom processor over Windows XP? That turned out to make squat all difference. How do we know all these visions you're making won't simply fall flat on their face as well? 4x4 was supposed to be great too, haven't seen anything from that in 18 months, guess that still has growth room for the future? I think AMD just realised it was a scrapheap and simply cancelled all development on the 4x4 platform. True potential indeed.

Well said.

Marketing VS. Results

 

[keithlm]

Well said.

Marketing VS. Results

Except for when the results are:

#1: Not as dramatically different as people seem to infer on these forums.

AND/OR

#2: Not accurate. If you purposefully handicap one chip and it scores "almost as good" as the other chip: Guess which chip won the benchmark? And it's not always the chip that scored higher.

(I.e., the benchmark shows the "losing" chip trailing by 5%. But the handicap you put into place can affect the results of the benchmark by anywhere from 5%-30%. So again: WHO WON?)

 

[Guest]

Well said.

Marketing VS. Results

Thank you kindly. I'm sure Baron Matrix will be quick to respond, and it is doubtful that he will accept few if any of my points or reminders. I remember one of the selling points of the Athlon 64s being that you could run a 64 bit OS on them, which is a great feature for the minority that did choose to, but it was ultimately we're still here over 90% of us on 32 bit OSs. Sure the Pentium 4s that went head to head with it didn't have that support, making them 100% incompatible with 64 bit OSs, but it didn't make a significant difference and still doesn't even after they're both hitting the dustbin as people charge to replace them with Core 2 Duos/Quads and Phenoms. Most of the features that make Phenom an 'advanced processor' aren't made for this day today, but that's the problem, its trying to be sold today. With whole computers being bought, used, and dumped so quickly with the rapid pace of development; it is silly to make for a distant Tomorrow because by the time the 'tomorrow' for 64 bit computering on the mainstream or such comes around, the processor may have been replaced by a new archetecture or even have gone through several revisions. Throw a few features on to help it last indeed, but don't make that the main meat of the processor, to be bought today it has to be useful today.

 

[uguv]

Do you know what the main differences are between serves and desktops?
If you know that and remember what have written about scaling, could you see what I am talking about?
There was some shortage on 45nm C2Q because Intel is prioritizing the server market, much more money there and they needed to get all advantage they can to compete with AMD.

Xeon and Opteron is almost (if not the same) same processors as C2Q and Phenom.

I don't think there's any argument that the Opterons scale better in multiple processor setups. Anand recently wrote an article comparing 4P 16-core setups where this is even more evident.

On the desktops it's a different story; there generally isn't a FSB bottleneck with desktop applications, and Intel's chips win almost every benchmark.

 

[kassler]

Why what?

You said that AMD scales better on multi-socket motherboards, then I asked you why. Do you know why it scales better on multi-socket motherbards?
If you why it scales better on multi-socket motherboards then you should be able to also see the reason why AMD scales better in other situations. There isn’t any magic with multi-socket motherboards. AMD has something called hypertransport when it talks to other parts on the motherboard, that doesn’t interfere with the memory bandwidth. Intel has only one channel and that is the FSB, both memory and talks to other parts goes through the FSB.
If you know what you are talking about then you should be able to understand this.

 

[kassler]

Besides, didn't you make a lot of hot air about Vista's NUMA awareness being a major advantage for the Phenom processor over Windows XP? That turned out to make squat all difference.

It does! The thing is that the market for selling these applications isn't that big because Intel is so dominating. But if Intel will start selling Nehalem to the desktop market you will see a change in how applications are developed that need performance.
It isn’t possible to create applications scaling to 20 threads on Intel because that’s no meaning because the performance is going down. Creating multithreaded applications on for Intel needs some extra planning.

 

[kassler]

On the desktops it's a different story; there generally isn't a FSB bottleneck with desktop applications, and Intel's chips win almost every benchmark.

If you overclock the FSB most people notice performance gain, even if the processor is working below 50%.

 

[TechnologyCoordinator]

On the desktops it's a different story; there generally isn't a FSB bottleneck with desktop applications, and Intel's chips win almost every benchmark.

DING DING DING.

Very correct.

kassler had been arguing that Intel desktop quad cores don't multi-task well. When asked for proof he provided a link saying that multi-socket opterons scale better than Xeons.

While the information kassler provided is accurate, it doesn't prove his original assertion.

You are very correct. FSB in virtually all cases, is not saturated on an Intel desktop machine.

 

[kassler]

You are very correct. FSB in virtually all cases, is not saturated on an Intel desktop machine.

No it isn't running rendering, complex encoding or single threaded games and those applications is running isolated.

 

[TechnologyCoordinator]

No it isn't running rendering, complex encoding or single threaded games and those applications is running isolated.

Even in most multi-threaded and multi-tasking situations it is not saturated. Regardless of what you try to FUD out there.

 

[keithlm]

and Intel's chips win almost every benchmark.

Unless the benchmarks are actually run on a level playing field. Then the results are not as black and white.

But people blindly accept the results of the benchmarks that were created with one chip handicapped. Even after that handicap is pointed out.

I'm going to laugh my butt off when Intel has to manipulate the masses to convince them that Nehalem is better than Core2.

 

[kassler]

Even in most multi-threaded and multi-tasking situations it is not saturated. Regardless of what you try to FUD out there.

Show one test that uses alot of memory in two or more threads and compare that to AMD where Intel wins.

 

[uguv]

Unless the benchmarks are actually run on a level playing field. Then the results are not as black and white.

But people blindly accept the results of the benchmarks that were created with one chip handicapped. Even after that handicap is pointed out.

I'm going to laugh my butt off when Intel has to manipulate the masses to convince them that Nehalem is better than Core2.

You keep saying this yet you never provide any benchmarks to back up your claims. Provide some data behind your arguments and prove my statements to be incorrect.

 

[keithlm]

You keep saying this yet you never provide any benchmarks to back up your claims. Provide some data behind your arguments and prove my statements to be incorrect.

Provide what?

Just about EVERY available benchmark between the Phenom and Intel uses DDR2-800 memory. This handicaps the Phenom by anywhere from 5%-30%.

Look at any of those benchmarks and add between %5-30% to the results.

(Yes... I know... I'm asking you to actually THINK.)

 

[kassler]

You keep saying this yet you never provide any benchmarks to back up your claims. Provide some data behind your arguments and prove my statements to be incorrect.

here is one review that is a big joke if the writer claims that he is neutral.
http://www.anandtech.com/showdoc.aspx?i=3326&p=1

 

[uguv]

@Kassler - What does the Nehalem preview have anything to do with comparing the core 2 quad to the phenom?

@Keith - I agree that DDR2-1066 will help the Phenom more than the core2 but 30% seems like quite a stretch. Are there any reviews that at least show comparisons between Phenoms using DDR2-800 and DDR2-1066?

 

[thunderman]

Reviewers like to handicap AMD processors…… or use Biased benchmarks. I’ve read various personal opinions of persons who have upgrade from Intel Core2 to Phenom and have been pleased with the results.

Nvidia are Evil Too! Nvidia overcharge! ATI equals performance for Buck! AMD for technology and innovation! AMD technology used for popular Animated films and star wars episode 3!!

AMD4Life!

 

[kassler]

@Kassler - What does the Nehalem preview have anything to do with comparing the core 2 quad to the phenom?

You will be surprised how much information that is manipulated. Sites making reviews need companies that advertise. I don’t think they do this work for free. Who has most money? If you know the technology behind and you read reviews then it is often that one don’t really understand how/why the reviewer can’t see reasons. Sometimes there are good reviews. I think they need to do these to have some sort of reliability.

 

[uguv]

You will be surprised how much information that is manipulated. Sites making reviews need companies that advertise. I don’t think they do this work for free. Who has most money? If you know the technology behind and you read reviews then it is often that one don’t really understand how/why the reviewer can’t see reasons. Sometimes there are good reviews. I think they need to do these to have some sort of reliability.

Ah, so it has nothing to do with what we're talking about. It seems that you're implying that the phenom generally loses to the core 2 in benchmarks not because it's slower but because Intel bought off all the reviewers.

 

[kassler]

Ah, so it has nothing to do with what we're talking about. It seems that you're implying that the phenom generally loses to the core 2 in benchmarks not because it's slower but because Intel bought off all the reviewers.

I don't imply anything. But check this:
http://www.nzone.com/object/nzone_twimtbp_gameslist.html
http://techreport.com/discussions.x/14707

What I am saying is that there is important not to swallow everything. For me it seems very strange if somebody say a lot of positive things about C2Q and Nehalem but not Phenom.
Best is that they explain what each processor is good/bad at but that is rarely done.

 

[keithlm]

@Keith - I agree that DDR2-1066 will help the Phenom more than the core2 but 30% seems like quite a stretch. Are there any reviews that at least show comparisons between Phenoms using DDR2-800 and DDR2-1066?

Yes. But of course I can't find it right now... I don't bookmark them. (I'll go look for it.) You CAN see the difference by just looking at results from a site that uses DDR2-1066 and compare the results with a site that uses DDR2-800... but that is not very scientific.

I'm having Gigabyte BIOS issues or I would just go run some of my own. Perhaps after I get my sb750 board in the next week or two. (I might buy 2 ATI 4850's also.)

I do know from personal experience that going from DDR2-800 to DDR2-1066 will gain about 200 points in 3DMark06.

EDIT: Apparently you accidently read 30% and did not see that I typed a range of 5%-30%. I even typed it twice... so you missed it twice.

 

[spongebob]

Provide what?

Just about EVERY available benchmark between the Phenom and Intel uses DDR2-800 memory. This handicaps the Phenom by anywhere from 5%-30%.

Look at any of those benchmarks and add between %5-30% to the results.

(Yes... I know... I'm asking you to actually THINK.)

After digging around some major eTailers, I noticed that every Phenom system I saw ships with DDR2-800. Universal dealer ignorance, or perhaps (as uguv suggests) the bang isn't worth the buck?

EDIT: Apparently you accidently read 30% and did not see that I typed a range of 5%-30%. I even typed it twice... so you missed it twice.

Don't think he missed anything - I think he was just questioning the 30% high end. IMO, one, maybe two, benchmarks showing a 30% increase is statistical noise.

 

[kassler]

After digging around some major eTailers, I noticed that every Phenom system I saw ships with DDR2-800. Universal dealer ignorance, or perhaps (as uguv suggests) the bang isn't worth the buck?

Or maybe customers doesn't understand that fast memory will add performance for applications that are using a lot of memory on phenom and because of that doesn’t request it. On the Intel side DDR2-1066 is useless.
The need for an IMC and why the FSB is dead