Intel's 45 nm Penryn CPU: 4 GHz Air Cooled

[muk]

Four cores, 45 nm structures and overclocking up to 4 GHz make Intel's Penryn Core 2 CPUs faster and more efficient. How bad does it look for AMD?

http://www.tomshardware.com/2007/10/29/intel_penryn_4ghz_with_air_cooling/index.html

 

[prolfe]

First Reply!

 

[prolfe]

Actually, something constructive is probably in order....AHEM

What a thorough, well-written review. Although I'm sure the author will get flamed, let me say that I thought it was great. Two MINOR points include the fact that although the system setup mentions only one hard drive, the photo shows two. Also, the wet jeans, orgasmic, repeated, bold print declaration that AMD is in trouble were over the top. Still, if the benchmarketing is to be believed, they are accurate statements.

 

[rodney_ws]

Just another paid Intel pumper supporting the home team!!! I get so sick of this. God, I'm gonna get my tech news from a LEGIT source. Bless you Sharikou!

(please note: the above post was 100% sarcasm)

 

[p05esto]

But I hope AMD keeps going. We need them to keep Intel innovating. If AMD goes away and Intel reigns supreme they will rest on their laurels again. It wans't until AMD challenged Intel did they wake the sleeping giant.

Although AMD got what it asked for it still benefited comsumers. Let's hope they can do it again one of these days.

 

[Slobogob]

Page 19 first chart is a piece of crap as while it does state the cache is 2X4 compared to 2X6 it leaves out the fact that Q6800 is also slower in GHz. IE The QX9650 at 3GHz and the QX6800 at 2.93 and may be almost no improvement over the QX6850.

You think they forgot to adjust multiplier and FSB to set the 6800 to exactly 3Ghz?

 

[elbert]

You think they forgot to adjust multiplier and FSB to set the 6800 to exactly 3Ghz?

Crap just noticed that its the QX6850 against the QX6800 on the chart.

I would have liked to seen QX6850 OC'ed in the benchmarks to see what the exact improvements were up to 4GHz.

 

[Reynod]

The Inq also posted 4Ghz results and stable.

Jeez the first few stats

http://www.theinquirer.net/gb/inquirer/news/2007/10/29/first-inqpressions-intel-qx9650

Jeez ... 80K on Sandra CPU

 

[zenmaster]

I think the article was a little over the top in the praise of the Penryn.

The Penryn showed itself to be what everyone said it was.
A modest improvement over the previous C2D.

7% Gain or so from cache and perhaps enhanced SSE4.
Slightly better thermals.
About EQUAL overclockabiltiy. Perhaps SLIGHTLY improved.

To get their particular CPU to 4.0Ghz, they had to fry it in the process.
So, in essence 3.8Ghz was the limit they could take their chip before they had to do a major voltage increase.

3.6Ghz on the G0 Q6600 with air is very doable.
3.8Ghz on the G0 Q6600 with air has been reported.

So, we are not talking a major improvement.
Perhaps 200Mhz and a little cooler.
This is a nice advance but nothing earth shattering.

Now, I would not be surprised if Intel was able to improve the steppings over the next few months and we will see even better OCs into the low 4.0Ghz range and then it will be really exciting.

 

[allenander]

Now you guys should compare it to a AMD 4x4 System so we can see for certain what the difference in performance is between 4 Intel cores and 4 AMD ones. I mean really.. I only read it to find out where AMD stands, then you do a completely and unbalanced test... I mean really guys..c ome on... thats silly. Hey how about you guys compare a 4 core intel system to a 1 core AMd64 and see how low AMD gets then eh?

 

[Tjik]

For credibility reasons:

You do include the following at least twice:

Also, we want you to understand why we are comparing Intel's quad-core CPU to a dual-core processor from AMD. For now, since we don't have any alternatives, it's the best we can do until Phenom arrives.

A lot of folks don't read all the small print, probably more headings and the text next under, so if you claim in a heading with bold text:

Intel 45% Faster Than AMD

... follow it up with this pretty harsh text:

The performance results are nothing short of appalling for AMD. Intel's new Penryn-based QX9650 pulls ahead of AMD's fastest desktop model, the Athlon 64 X2 6400+, by 45%.

... it doesn't make the article look objective. The Intel architecture is in no doubt the most efficient and faster, but how comes a dual-core's results become "appalling" in comparison to a quad-core?

Intel or AMD doesn't matter for me, but these kind of spinning of words in Tom's Hardware have made me a very rare visitor. In case anyone now claims that I'm simply complaining because I'm a hidden AMD fan, let me here present some figures which compares a Intel dual-core (E6850) with a quad-core (QX6850):

Cinema 4D = -100 %
3D Studio Max = -100 %
DivX 6.6.1 = -30 %
Mainconcept H.264 = -70 %

Does that lead to the conclusion that a dual-core Intel E6850 perform appallingly bad?

Come on now Tom's Hardware, you can do better than this!

 

[rodney_ws]

Now you guys should compare it to a AMD 4x4 System so we can see for certain what the difference in performance is between 4 Intel cores and 4 AMD ones. I mean really.. I only read it to find out where AMD stands, then you do a completely and unbalanced test... I mean really guys..c ome on... thats silly. Hey how about you guys compare a 4 core intel system to a 1 core AMd64 and see how low AMD gets then eh?

Who do you know that owns an AMD 4X4 system? The article clearly stated that the test was for single processor systems. Virtually no large OEM sells 4X4 systems so the average home user is going to have a hard time acquiring one... it's just not a viable system in terms of numbers. I felt the article clearly explained itself, its purpose and its limitations.

 

[zenmaster]

Reading the Article, I think the reviewers were in a tough spot in regards to AMD.

1) Do you not compare it against AMD systems at all until the Phenom is available? This would have angered AMD fans that AMD was excluded.

2) Include AMD chips which relatively speaking are dated and have them show poor performance. This is what they did and angered AMD fans.

I think perhaps a little less commentary about the relative poor showing of the AMD chips may have been in order. Save the comments for when you have a phenom to compare in a couple weeks.

 

[Scotteq]

Well... To play Devil's advocate: AMD either couldn't or wouldn't provide a new Quad. No production sample. No engineering sample. Nothing. It's pretty hard to do a comparison to something you can't get your hands on. So Tom's was stuck using the best AMD Proc they could get. Toms made this clear at the beginning of the article.

And it's not an isolated ting: There have been complaints from the beginning about the non-availability of AMD's quad. And the ones you *can* get aren't even the 2Ghz speeds. Newegg, one of the largest online retailers, doesn't have any in stock at any price. The fastest (clock speed) AMD desktop proc you can get there is a stinking Althon 3.0. And there is only ONE Barcy processor in stock: A 1.7 Ghz for $214. And the 2.0Ghz has *never* been in stock there.

http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=2+50001028+40000343&Configurator=&Subcategory=343&description=&Ntk=&srchInDesc=Barcelona


So how about changing:

Come on now Tom's Hardware, you can do better than this!

to:

Come on now AMD, you can do better than this!

Quite frankly, I'm highly disappointed: This is putting AMD two generations behind the power curve, as far as manufacturing process. And I'm not seeing anything on the pipeline to make up the difference. So we have a paper launch NON-delivery of a product which is becoming 2 generations old as it hits the market. AMD is getting their a$$es KICKED in the enthusiast market. Period. End of Discussion.

And it's sad. Just sad.

 

[dragonsprayer]

Thank You for the article -THG is an awesume site! (unless your work for amd, live for amd or just a play old amd fantic)

q6600 4ghz on air: http://youtube.com/watch?v=-JJr2Vnfs1w


come on THG i think 4.5ghz is the number!

 

[dragonsprayer]

I think the article was a little over the top in the praise of the Penryn. (are you kidding mr amd fantic - 4w idle, free 7% increase, 4-5ghz possible with water easy, blows amd out of the water 45-55%)

The Penryn showed itself to be what everyone said it was.
A modest improvement over the previous C2D. (no its much more overclock-able - you must be blind?, note it runs the same temperature at 3.8ghz as the qx6850 runs at 3ghz - the article under plays the benifts) (you need to get smoking the green part on your cpu box! you ever even seen an intel chip)

7% Gain or so from cache and perhaps enhanced SSE4.
Slightly better thermals.
About EQUAL overclockabiltiy. Perhaps SLIGHTLY improved. (at give speed, overclockablity is huge - i don't know what rock you crawled out from but dell, hp, gateway etc now overclock there highend stuff)

To get their particular CPU to 4.0Ghz, they had to fry it in the process. (only thing fried is you - 65c fried?)
So, in essence 3.8Ghz was the limit they could take their chip before they had to do a major voltage increase.
(major voltage change? do even read anything? dell ships there top system at 1.62v)

1.33v is high? what do you run your amd at? i bet its higher!

3.6Ghz on the G0 Q6600 with air is very doable.
3.8Ghz on the G0 Q6600 with air has been reported. (yes i agree not the point of the article - mine is running 3.72-3.85) i ship systems at 3.6ghz

So, we are not talking a major improvement. (yes we are you just did not pay attention)
Perhaps 200Mhz and a little cooler. (perhaps alot cooler and possible faster with water - nooboie)
This is a nice advance but nothing earth shattering. (if you work at amd it is they have hard time at 2ghz since only the 1.7ghz rejects can be found)

Now, I would not be surprised if Intel was able to improve the steppings over the next few months and we will see even better OCs into the low 4.0Ghz range and then it will be really exciting. (you woke up - thg did not push the chip hard they called 1.45v very high but its not, 1.5-1.6 is high - they had to have a part 2 (noobie)

 

[Tjik]

Scotteq, if you read my comment and added comparison between what a difference you see between two per core equal CPU:s, you should have understood that my criticism wasn't directed at how Tom's Hardware once more proved the well known and accepted truth: Intel is at the moment the clear winner. I've got no argument about that.

My argument has more to do with journalism. Tom's Hardware isn't any obscure exclusively known by some few nerds, it has become a larger player on the net influencing consumer decisions. Why then should articles have a touch of yellow journalism? There's no need for half-truths or twisted facts.

That AMD needs to present something more impressive than the old K8 architecture is undoubted. It's also correct that the quad-cores aren't available yet. Nevertheless it's a lot easier, I can't even see that there should be any argument about this, to produce a review, than to produce the object for the review. Who really knows, maybe the next generation AMD won't compete with these Intel ones in pure performance, but should that change the substance of articles?

 

[zenmaster]

Dragon,

Do you understand the slightest technical aspects?
I think you fail to grasp that this chip is built using a smaller process and using new materials.

What is appropriate voltage for larger 90nm or 65nm may not be good for 45nm chips with it's different materials. Since they can operate with less power, supplying old levels may harm the chip.

Also, Water Cooling would have done very little to assist with an OverClock. The chip was hitting a Voltage limit, not a heat limit.
Increasing the voltage further would have simply further endangered the chip.

Your desire to bring AMD into the discussion shows the weakness of your arguments. The point of discussion was comparing the new Intel Chips to the old Intel Chips. To try and prove your point, you compared the new Intel chips to AMDs.

A classical philosophical fallacy employed by those who need to divert readers from the weakness of their arguments on a given topic by discussing a different one.

 

[Scotteq]

Scotteq, if you read my comment and added comparison...

I did read your comment - And Tom's covered their journalistic behind by clearly stating in the beginning of the article that they were unable to obtain a Barcy Quad. Engineering sample or otherwise. Yet the readership/membership *really* want to know how these things stack up against each other. So what do you do? Leave AMD out entirely and get flamed for poor journalistic standards because you didn't include them? Or do you do your best to provide the content your readers demand with the best you can get your hands on, and then get flamed for poor journalistic standards because you were forced to include a poor~er performing product in the comparison? Do you run the comparison and risk AMD's marketing folks getting p*ssed off at you for somehow "making" their stuff look bad? If AMD gets mad, will they not provide product for future testing? Or do you use your position with the readership at large to make the case that AMD *really* need to get performing product into consumer hands?

Quite frankly, I am of the opinion that it's far better to publish the results of the best (Product X versus Product Y) that you are able to obtain. Explain clearly what it is, why the comparison was made, and why the results look as they do. Then let the chips fall where they may. {No Pun Intended, I assure you....} This is how people make their buying decisions, after all: We look for the best we can obtain with the money we have to spend.

I understand your point - It may... Hell - almost surely, would have been better from AMD's perspective had Tom's left AMD out of the comparison altogether. But that's not what the member/reader~ship want to see. We want to see Brand A's best going head to head with Brand I's best.

Maybe next time Toms could do a $300 Processor/$150 Mobo/identical memory/identical HDD test, or something. I know it could be made to sound like a backhanded comment, but I would expect to see more comparable results given the way AMD is currently positioning themselves in the market. And that'd go some way towards neutralizing Intel's advantage in outright performance.

 

[bwdsmart]

I personally think the 4x4 system should have been used. I dont consider an extreme edition cpu mass market at 1000$. thats just my opinion. good article though, good to see intel is staying on its toes to stay ahead.

 

[evongugg]

Very few people have that kind of money for a processor.
According to Anandtech.com there isn't that much of improvement over Conroe for the masses (lower cost processors), except in video editing, etc. and overclockability.

 

[kansur0]

I am actually and AMD fan. I am glad that these articles continue to make AMD look foolish and put pressure on them because they really have been behind the curve. No 4x4 system available to test? When did they say that this system was going to be available? Hmmm. I think AMD didn't want to provide a system because it would probably just have been beaten anyways. AMD really likes to blow horns when they have something on paper...but come on AMD...if you are going to tell us about technology make sure you have the hardware available to back up the claims!!!

I WANT PHENOM AND I WANT IT NOW!!!

 

[Guest]

I was here.

 

[Scotteq]

Who was surprised that a 4 core 45nm processor beats a 2 core 65nm processor by 45%?? Considering it has 50% more cores, a die shrink and 6x more cache. lol. Stupidest most pointless comparison test THG has done to date. Of course they are Intel fanboys, so its expected.

Given the entries the result is/was a foregone conclusion, and looking only at the numbers it is disingenuous to even try. But to my mind the real questions are more along the lines of "Why didn't AMD provide a new Barcy Quad when they were clearly asked to do so?" or maybe "Why is Tom's Hardware forced to use old AMD stuff in order to have something to write." and of course "Why, given the very recent years of AMD's performance advantage over Intel, hasn't AMD been able to respond in kind."


As for the Fanboy stuff - If Toms had left AMD out of the comparo altogether, I get the feeling the same people would be using the same phrase. Not like there haven't been recent posts claiming Tom's are Intel Fanboys because of the relative lack of AMD coverage, or anything. Also, given that the commentary a couple years ago from the Intel side was pretty much the same as currently witnessed from the AMD side, I personally have to conclude that Tom's is writing articles on who's stuff beats who's. No More, No Less. That Tom's are rounding out the selection with whatever is newsworthy, from whatever source is providing newsworthy material. No More, No Less. And that the only thing that has changed is which side of the fence happens to be growing the sour grapes. No More, No Less.

And yes, I do notice the irony of "Intel Fanboy" comments from a member who'd rig is Intel based. Quite Amusing, really.

 

[skittle]

Who was surprised that a 4 core 45nm processor beats a 2 core 65nm processor by 45%?? Considering it has 50% more cores, a die shrink and 6x more cache. lol. Stupidest most pointless comparison test THG has done to date. Of course they are Intel fanboys, so its expected.

Most of those applications dont take advantage of multiple cores, also the results are skewed by the %145 percent improvement in 3dsmax and the 115% improvement in cinema 4d %130 improvement in h264.

14% to 30% improvement in GAMES
37% in lame (who uses mp3 these days anyway..)
30.6% in AVG
40.8% in WinRAR
I wouldnt say Im suprised... but the results are not because they used 4 cores (in the non video encoding/rendering tasks atleast). These are single threaded applications, and it they look very good for a chip that isnt on the market yet!