4.11.2007

Kicking the Dead AMD Horse - Intel's QX6800

Watching Intel release one product after another is like watching Tiger Woods play golf, or Michael Schumacher in Formula One. It's fun and awe inspiring in the beginning, but it can get boring very quickly.

Intel released last week its fastest Quad Core Extreme edition yet. The QX6800 is priced above $999 just to keep demand in control.

What about the boring stuff, its performance? Just take a look at the graph on the right. The gap between Intel and AMD is getting too big that we may need to stop benchmarking all together.

"You may have gathered that the Core 2 Extreme QX6800 is the fastest desktop processor that money can buy. No nuance is required to discuss this one. The QX6800 nearly swept our entire benchmark suite, and in many cases, it crushed its main rival, the Athlon 64 FX-74." the TechReport.

30 comments:

Anonymous said...

intel looks like it is milking the public again with the steep price.

amd failed the consumer.

Ho Ho said...

It is interesting to note that Intel basically pulled forward that quadcore for almost two quarters. Originally it was not meant to be released before Q3 when there are 65W 3GHz dualcores availiable.

Does anyone else remember how Scientia said it was impossible for Intel to release higher clocking quads any time soon thanks to some magical cascade overheating problems. Now it seems as they are confident enough to release CPUs with massively increased TDP and do not worry.

Not Penix said...
This comment has been removed by the author.
Not Penix said...

I remember from scientia's blog saying how intel was "unable" to release a higher clocking quadcore a few days ago and look what happened. Apparently he can't take no for an answer or stand to believe he might be wrong.

I wouldn't call milking the public with these cpu's, where have you ever read that the extreme edition cpu's were meant for the masses, they are not, they cater to the high end where people buy them because they know exactly what they are.

AMD hasn't failed the consumer, they have been releasing very decently priced cpu's, but really have been dragging their feet as of late.


BTW hoho, having fun not being able to post on scientia's blog becasue you pissed him off by proving him wrong?

Ho Ho said...

not penix
"BTW hoho, having fun not being able to post on scientia's blog becasue you pissed him off by proving him wrong?"

You can say that :)
I made an apology, if you can call it so, post in the thread a few days ago, he removed it together with a post in the newest thread. I replaced the apology one and so far it seems as it has stayed there for more than a day. I guess he accepted my apologies of not knowing that free speech is not allowed in his blog.

Anonymous said...

You are such an Idiot.

I thought that you were smarter.

You are like the others, A BIG LIAR!!!

Why you didn't put this benchmark

http://techreport.com/reviews/2007q2/core2-qx6800/index.x?pg=9

softwares like maxon, maya, 3d studio max that are used in developing animation movies!!!

You posted the benchmark off the benefits of "multimedia" extensions like MMX, SSE, and SSE2

and You, little piece of shit know that AMD DO NOT SUPPORT

SSE, and SSE2 instructions.

You are just a kid.

People he is telling you lies.

Don't listen do this kid!!!

Anonymous said...

Text from

TECHREPORT

''Hey, look, actual drama: the FX-74 was ahead of the Core 2 Extreme QX6700 here, but the QX6800 recaptures the lead for Intel.''

''Shockingly, the QX6800 is not at the very top for once, although it's close. The FX-74 finishes rendering the scene 10 seconds sooner. ''

These are real softwares that requires CPU and MEMORY power.

This is why Intel needed QX6800!

Figthing against 4 years old AMD architecture.

This is just too funny!

Ho Ho said...

"You, little piece of shit know that AMD DO NOT SUPPORT

SSE, and SSE2 instructions"


Yes, it does. It will even support SSE4 before Intel does. Now who is the one who doesn't know what CPU supports what?

Anonymous said...

What ignorant fools the AMD fanbois are... the author from techreport is actually surprised that the Fx-74 managed to win a single benchmark against the Qx6800 that is why the author chose to use "shockingly" and "some drama" to describe it you moron its not used in a good context for AMD....hmmm i wonder what the performance per watt is on the FX-74 something AMD used to be very proud of back in the days but suddenly it doesnt matter anymore.

This is too funny...AMD getting a whooping by an ancient modified pentium 3 Arch......pathetic.

Anonymous said...

If this is released now, imagine what will happen in a few months. Intel is not standing still.

Scientia from AMDZone said...

ho ho

You now seem to be in the same category as Red. That is a shame. Although you are very young, you did have at times an interesting point of view. Good luck on the other blogs.

Scientia from AMDZone said...

roborat

I see you have me linked. That would explain why I see traffic from your domain. BTW, thanks for not referring me as a delusional AMD fanboy.

Roborat, Ph.D said...

Scientia from AMDZone said...
roborat

I see you have me linked. That would explain why I see traffic from your domain. BTW, thanks for not referring me as a delusional AMD fanboy.


i'm sorry, you make too much sense or doesn't match sharikou's 'creative' mind to be labelled 'delusional'.

Now in the spirit of the debate, i say you're once again wrong to think that your blog and my blog are from different domain when you said you see traffic from my domain. you should have said, from my website or blog since our blogs comes from the same domain.
;)

Anonymous said...

But Scientia, you are a delusional AMD fanboy!

Scientia from AMDZone said...

roborat

Right, in Google/analytics yours is in Referring Source rather than Domain which is where Intel shows up. Yours is between TrackingAMD and AMDZone.

I still say your blog shouldn't be a parody of sharikou's. Personally, I think the quality of the comments pull your blog down but that is just my opinion.

Anonymous said...

Really..... the quality of your posts pulls your blog down scientia.....i suggest a shutdown of your blog.

Have you posted how the huge losses at AMD is a good thing for it yet.... if yes please let me know i am interested in knowing what new excuses you will have now.....let me guess the stock market is not running the rite benchmarks again???? Or the investors are still waiting for NUMA to show the great platform 4x4 is???

What a joke!!

Scientia from AMDZone said...

A joke, perhaps. But, do you know when what you post will have any signficance? First of all, when you are not so ashamed of your own posts that you have to post anonymously. And, secondly, when you can respond with a 1,500 word essay of your own instead of a one line insult. Probably not any sooner than that.

Ho Ho said...

It is kind of sad that you took my words that hard, scientia. Others and you yourself have told much worse things about people, most of them didn't care about it that much. I wonder why did you take this one time from one particular person so personally.

That was the only time when I have ever called anyone biased and even that time it wasn't really direct. Others have done the same and much worse. It almost seems as you are afraid of me proving you and others wrong from time to time so you simply cencor me out to make your points more solid. You also have edited many of my posts and removed important parts, that also is kind of suspicious. I'm sorry if this is wrong but that is the image you are drawing to me.

You talk about balanced view of things but you keep out the people whose oppinion you don't like. Do you call that balance?

I can understand removing posts that are meant to be insulting but this is definitely not what I meant to do with mine and you have to work quite hard to read it out to be otherwise. I just wanted to make you prove your point about why PGC is so much superior than others even when it is documented fact that AMD has helped them optimizing it. Throwing in a little controversy isn't that bad, it should only make you work harder about proving your point. Instead you chose not to prove me wrong. If anything this proves my point you so strongly disagreed with.


I wouldn't like to boast about myself but I think I was one of the few on your blog who knew anything about computer architecture. Only reason why I was mostly defending Intel was that there was so many people defending AMD that one more wouldn't make any difference. Whenever there was something positive about AMD I acnowledged it and you should know this.
On Intel side there were mostly people who don't know that much and only give it bad name by talking rubbish. Sad, but true.


One example would be abinstein. He knows a bit about things, quotes books and various other sources but doesn't understand them fully. E.g he said there is enough ILP in programs to have 30-120 instructions in flight in most programs but he failed to see that this was about ideal CPU with 100% accurate branch prediction, unlimited amount of rename registers, zero memory latency and many other things that aren't there in any CPU. The same goes about dynamically speeding up CPU cores when running single threaded applications. He tries but doesn't try enough and I just fix his mistakes. Unfortunately he doesn't understand when things he sais are wrong.


It is also sad that you removed my post in your post "Intel -- The Monopoly Under Siege". There were quite a few things there that would have greatly benefited from better explanation. It even pointed out a few errors.



"And, secondly, when you can respond with a 1,500 word essay of your own instead of a one line insult. Probably not any sooner than that."

My "essays" were often several times longer than your articles and were often the base of the discussion. Do you think it is purely coincidence that while I was posting there you enjoyed posts with mostly >100 replies but the last three have had only a handful of replies? I guess we will find out if you keep on ignoring me. It almost seems as you have started to boost your postcounts by making a lot of replies, roughly one third of posts is made by you in your later stories.

Anonymous said...

Ho ho, don't bother. His blog will go the same way that AMD will;)

Ho Ho said...

I guess you are right. AMD has some money problems and most likely will slow down a bit. It won't die any time soon, it'll just scale back a bit.

Anonymous said...

My one line insult has more truth to it than that 1500 words of garbage in your blog.

When you start calling AMD for all their success AND FAILURES!!!!instead of making excuses as if you are on AMD's payroll then perhaps i will register under a real name and post something.

You are more credible than sharikou i will give you that but thats not saying much......

Axel said...

Ho Ho personally I thought your comments were some of the best that Scientia's blog ever sees and clearly demonstrated knowledge of programming. Abinstein is a joke.

It's sad to know that a bunch of the comments on that blog were either edited, abridged, or removed entirely. I guess blog authors are free to be control freaks if they want; that doesn't make it proper or meritorious.

I find that Scientia's essays mostly ramble on about insignificant details and miss out on the big picture. DTX? AMD64? AMD/ATI developing low power chipsets?

How about the more significant issues, like AMD's looming financial crisis? Or at IDF, about Penryn's significantly improved IPC over Conroe and open talk about 3.33 GHz quadcore being compared with Kentsfield? Essentially nipping Barcelona in the bud and denying AMD profits for the next several quarters unless they scale CAPEX back by much more than $500 MM? These are the important things. Why not spend 1500 words on those issues instead of wasting time with trivial matters like AMD's DTX initiative.

And yes, Scientia is most definitely an AMD fanboy. Every single essay is couched and founded in a pro-AMD context. Just look at the themes of the last few entries:

The Roots of Monopoly
Why buying ATI wasn't a mistake. Of course, no mention of the poor timing and horrendous loan conditions.

The Monopoly Under Siege
ROFL. Clear signs of Intel's monopoly cracking, with AMD simultaneously losing unit & revenue share this quarter? And more next quarter with the 4/22 price drops? Get your head out of the sand.

THG's Soul
Rant about THG's Intel bias, which is probably wishful delusion. THG is an enthusiast site (despite their intent to cater to the general public) and they have always applauded the CPU manufacturer with the faster chips, whether AMD or Intel.

Where Are the Clock Speeds?
Focus of this entry is why Intel's clocks remain <3 GHz so far (and as Scientia wrongly deduces, through the remainder of 2007). The obvious fact of Barcelona clocks also being low are an afterthought added in to make the entry seem balanced. The focus, make no mistake, is Intel's supposed bulk Si headroom limit. Hogwash. Intelligent rebuttal in the comments makes a good case for Intel purposely holding back clock speeds. And guess what, they were right, Intel only needed 2.93 GHz to bring AMD to a cash crunch.

Chipsets and Chess Pieces
Justification for AMD's purchase of ATI. As in the later entry, no mention of the poor timing and horrendous loan conditions.

Mist in the Morning Sun
Similar to the later rant about Intel's clock speed headroom limits.

Etc. Go back to the 2006 entries, read for yourself, and weep. All articles pro-AMD in theme and conclusions. Many conclusions incorrect, mostly underestimating the magnitude of Intel's resurgence and the impact of C2D. Reminds me of the defunct Van's Hardware.

Anonymous said...

How about the more significant issues, like AMD's looming financial crisis? Or at IDF, about Penryn's significantly improved IPC over Conroe and open talk about 3.33 GHz quadcore being compared with Kentsfield? Essentially nipping Barcelona in the bud and denying AMD profits for the next several quarters unless they scale CAPEX back by much more than $500 MM? These are the important things. Why not spend 1500 words on those issues instead of wasting time with trivial matters like AMD's DTX initiative.

What else do you think AMD fanboy can talk about?

Scientia from AMDZone said...

ho ho

I wasn't referring to you; obviously you don't post anonymously. I disgree with roborat and sharikou180 all the time but they have enough integrity to own what they say.

What I was saying was that any criticism that I get from anonymous posters is of no importance. And, obviously a one line insult from an anonymous poster is trivial beyond words. I suspect that the great majority of anonymous poster who say things like this would not be capable of actually writing a rational rebuttal essay.

I have typically have 30-some Intel employees who read my blog. I would imagine that if it were as big a waste of time as some suggest that they would stop reading.

Anyway, Ho Ho, it wasn't a question of insults. The only point was that you suggested that I wanted the testing to favor AMD and you assumed this based on a three year old press release that you took out of context. In other words, you suggested that I was only pretending to be fair but was actually endorsing PG because I knew that it would favor AMD.

Now, if you don't agree me that is fine. If you think I don't know what I'm talking about that is fine too. However, I am not going to be called a liar on my own blog.

You still haven't dealt with that and that is why I haven't let you post.

Ho Ho said...

scientia
"However, I am not going to be called a liar on my own blog"

I was not calling you a liar and I didn't think your point about PG was only because of that single press release I found. It would be simple to check it but you removed my posts.

My original point was that PG is not the best compiler to use since we don't know how it compiles for different CPUs and it is possible that one or the other CPU manufacturer has given them ways to generate code that runs better on their CPUs*. Later I found out it generates different codepaths for different CPUs making it pretty much useless when comparing different CPUs since they might be running different code.


*) ICC is considerably worse in that. It doesn't check CPU capabilities at all and only makes a decision using vendor information.

One simple example where highly CPU specific code generation doesn't work. Let's say that Intel provided a bunch SIMD optimization tricks for Netbuts for PG. Code is generated with insanely long pipelines in mind. Often extra calculations are made instead of branching. Same code compiled for K8 or Core2 would be quite a bit different.

Ho Ho said...

I just saw the new article on scientias blog about C2D and I must say I haven't laughed so much in ages. There are so many mistakes and made up numbers that it almost looks like some of the older posts of sharikou when he was trying to make up half-real numbers.

Would you like me to generate a few tens of posts in your blog by trying to fix the thing you wrote there? It would be fun tu put your latest article side-by-side with some of your older ones that had similar text but different conclusions.

Unknown said...

OMG that's hilarious!

Benchmark a Celeron to see how C2D performs!

The funnier part is the Celeron will smack around an Athlon 64 :)

Axel said...

Yep, chalk up yet another irrationally pro-AMD blog to Scientia's now impressive catalog. I read that one this morning and was astonished at how deluded he continues to be. I even commented on it, forecasting that yet another of his predictions will turn out dead wrong, similar to his assertion since mid-2006 that despite Core 2 Duo, AMD would magically continue to gain overall volume share through 2007 at the rate of 0.6% per quarter. Where does he pull these numbers from? It's hilarious.

That blog truly is becoming more Sharikou-esque.

Anonymous said...

AMD worshippers. They never learn.

Scientia from AMDZone said...

ho ho & axel

Okay, you get your wish.