backtop

Hardware Conroe Benchmark Criticisms
Brandon Hill (Blog) - March 8, 2006 4:45 PM
Print E-mail del.icio.us 74 comment(s) - last by JumpingJack.. on Mar 11 at 5:13 PM
Intel's new Conroe looks to give AMD a punch to the gut -- or maybe not

Benchmarks posted yesterday of Intel's unreleased Conroe processor have sparked quite a firestorm in forums around the web. Users from both the AMD and Intel camp were quick to examine the numbers provided in the articles, but the end result was still the same -- Intel has a slam dunk with Conroe.

As is the case with benchmarks held outside of a controlled environment and conducted with equipment provided by a party that has something to gain from favorable performance numbers, suspicions have arisen over the benchmark results.

VoodooPC president and occasional DailyTech writer Rahul Sood adds, “The long and the short of it is Intel has crafted an excellent marketing strategy to show off their baby in its first trimester. They are trying to win the hearts and minds of enthusiasts half a year before they have anything to show us. They created these platforms in house, and we can only hope they unknowingly crutched the AMD system by using a chipset and motherboard - with an outdated bios - that no enthusiast supports.”

While the test systems and circumstances surrounding these benchmarks of an unreleased Intel processor and an overclocked AMD processor are far from ideal, this question must be asked -- how often do you get the opportunity to test a fully functional, unreleased product six months in advance?

Comments     Threshold


This article is over a month old, voting and posting comments is disabled
Is this right?
By Questar on 3/8/2006 4:56:44 PM , Rating: 2
quote:
we can only hope they unknowingly crutched the AMD system by using a chipset and motherboard - with an outdated bios - that no enthusiast supports.


Did he say that?

Why would somebody "hope" that Intel goofed the setup of the AMD system?

Frankly he sounds scared that Conroe performs the way it does.




RE: Is this right?
By Bootstrap on 3/8/2006 4:58:17 PM , Rating: 2
I was just going to post this... the quote makes no sense unless he actually wants to see Conroe flop. Typo, maybe?


RE: Is this right?
By Tegeril on 3/8/2006 4:59:41 PM , Rating: 3
I agree, this makes no sense whatsoever. However, Rahul hasn't exactly been unbiased in his commentary on that lame blog in the past.


RE: Is this right?
By alphaaa on 3/8/2006 5:03:08 PM , Rating: 2
he probably does want to see conroe flop. rahul's an amd fanboy, unfortunately.


RE: Is this right?
By Questar on 3/8/2006 5:09:48 PM , Rating: 1
Oh! My GOD! Intel used a bios revision from NOVEMBER!! I mean that's so OLD!!

Heh, wish I could crack open one his machines a to see what bios version he's shipping.


RE: Is this right?
By smitty3268 on 3/8/2006 5:16:30 PM , Rating: 2
We know they crutched the AMD system (although it is debatable how much slower it actually is), but what he is hoping is that Intel did this accidentally rather than on purpose. Or at least that is how I read it. Personally, I don't see why it matters whether it was done accidentally or not, what matters is that the test might not have been completely fair.

The bigger problem, I think, is that the tests were very limited. The only ones done were:

1. Gaming, which is usually VERY dependant on the amount of cache. I think if the X2 had 4MB of L2 cache it would probably be back in the lead.

2. Video/Audio encoding, which was never exactly the strongest point for the X2. Also, 2 of the 3 tests only had Conroe in the lead by about 10%. I figure that is probably due to their better SSE throughput.

I'm not saying that these tests aren't useful, becuase they are. But I'd like to see tests in other apps before concluding that Conroe is the monster most people are right now.


RE: Is this right?
By JackPack on 3/8/2006 5:29:00 PM , Rating: 3
Uh, yeah, Intel "crutched" the AMD system by overclocking it, using 2-2-2 memory, and a DFI motherboard.


RE: Is this right?
By smitty3268 on 3/8/2006 5:41:32 PM , Rating: 3
Try reading the original quote:

quote:
crutched the AMD system by using a chipset and motherboard - with an outdated bios - that no enthusiast supports.


Intel DID do this, and you are using a straw man argument.

As for the argument I've seen several times that Intel has no reason to inflate their numbers because they will be tested again in 6 months... I never understood that. Intel has millions of reasons to hype up their future processors, and nothing really important to lose. I don't think the performance tests were really wrong, but I do think that they carefully selected which apps to let people test and made sure that they were the most flattering.


RE: Is this right?
By JackPack on 3/8/2006 5:46:06 PM , Rating: 2
Outdated BIOS?

Since when has a speed bump required a BIOS update? Uh, yeah, as if performance will be drastically reduced if it does not identify FX-60 during boot. Aside from cosmetic reasons, and perhaps a minor microcode update, an updated BIOS certainly isn't going to affect benchmarking results, given FX-60 is still E6-step.


RE: Is this right?
By smitty3268 on 3/8/2006 6:00:40 PM , Rating: 2
Well, I think you are probably right. Which is why I said I didn't think that was the important issue. Although you sure seem to be focusing on it...

Still, are you sure it doesn't matter? I assume the apps are going to directly query the CPU for its supported instruction sets, but is there any chance this goes through the BIOS? If so, then the AMD benchmarks might have been without SSE2. Or even SSE. Like I said, I don't think this is the case, but it is enough to give me pause.


RE: Is this right?
By tauron on 3/9/2006 3:02:39 AM , Rating: 2
One thing everybody missing is that very soon the sample CPU's will be delivered to the important vendors & partners (also web sites like Anandtech, Hexus, etc.), where Intel will have no chance to lie.

In fact it's at least technically possible to disable SSE support by the help of BIOS but here this is not the case since the performance numbers got by overclocked X2 in those "Intel controlled benchies" are very real life !

Of course people can get better numbers by using better components like DDR500 with low latency, nvidia based mobo and some other optimal settings but still the difference between Conroe scores and that would be dramatic.

Besides that Intel's 975 chipset supports also DDRII 800 like the upcoming AM2 AMD's. And in the tests Conroe used only DDRII-667.


RE: Is this right?
By BigT383 on 3/8/2006 5:38:48 PM , Rating: 2
Personally, I hope that intel was either crippling the AMD system, or AMD comes out with something that can compete by the time we actually get our hands on this thing. Why? Becuase the industry moves the most when the companies are neck and neck (just look at the GHz race). One company keeps trying to outdo the other, and we get speed increases as fast as the companies can put them out. Unlike right now. Today AMD can afford to release updates since they've clearly won this round.

Also- gaming has VERY LITTLE to do with cache at all! In graphics programming, most of the data changes every frame (in an FPS, every time you move the entire scene must be recalculated), thus there is very little that can be cached from one frame to another. Look back at how popular the original celerons were with gamers- they had no L2 cache but still did alright in games (and you could overclock them to be faster than the fastest Pentium 2s). Office apps, which do the same integer ops over and over again and rely heavily on cache, were a different matter. More recently, look at how well Prescotts do vs Northwoods at the same clock speed- doubling the cache doesn't make Prescott blow Northwood out of the water in games. Nowadays, there really is no difference between 512k per core and 1mb per core in games on the X2. Not having any would definately hurt, but the 512k is enough room to cache the small amount of info that stays the same from frame to frame. Plus, with the low latency on-die memory controller, cache misses are less of a problem.

That said, it looks like conroe certainly holds promise. If AMD can't step up, it looks like I may head back to intel for my next upgrade. My last intel processor was a celeron 300a o/c to 450!


RE: Is this right?
By smitty3268 on 3/8/2006 6:05:42 PM , Rating: 2
You're partially right about the cache for gaming - it only makes about a 5% difference on AMD to go from 512K to 1024K. Since the increase in size for the Intel chip is larger and it has no integrated memory controller, I could see the cache being as much as 10% or even 15%, but no more. And maybe only 5% or so, but that is a little more than you seemed to imply.


RE: Is this right?
By ninjit on 3/8/2006 5:13:23 PM , Rating: 2
Morons,

the quote says "we can only hope they UNKNOWINGLY..."

meaning, he hopes intel didn't use a crippled AMD system INTENTIONALLY, because that would make things even worse for them publicity wise.



RE: Is this right?
By TomZ on 3/8/2006 5:26:57 PM , Rating: 2
I disagree - I think he is saying that he hopes the numbers for the AMD system are artificially high so that the difference between Conroe and the AMD are less than was reported. I don't think that he gives a crap about possible Intel publicity problems if they made a mistake.

As has been pointed out by others on the 'net, the AMD numbers are pretty consistent with other benchmarks performed by objective third parties. That only leaves open the question about whether the Conroe numbers are inflated. But again, as has been pointed out by others, it is unlikely the case, since in a few months others will be able to run the same tests and prove/disprove Intel's numbers. Intel would look foolish if there were large differences.


RE: Is this right?
By rahulsood on 3/8/2006 6:03:52 PM , Rating: 4
Yes, sorry guys, I edited it for clarity - I can see where the misunderstanding is.

EDITED FOR CLARITY - If you read the article you'll see that the AMD system is clearly crutched - I am hoping that Intel had no clue that this was the case. I suspect they did this unknowingly.

By the way, I have openly suggested that Intel is going to be back on the rails this year on more than one occasion. We are still a customer of Intels - mostly on notebooks, and I don't have my blinders on.

I am suggesting that the benchmarking is not an accurate representation of what AMD can offer today. If an AMD expert built the AMD machine, just as an Intel expert built the Intel machine we might see closer results. In no way am I discounting the power of Conroe, but I am certainly discounting the quality of the build :)

Just wait until we get some real product in our labs, then you can make an informed decision. Until then we're still talking about a product that is months out. They can't possibly pull it in sooner.

http://voodoopc.blogspot.com/2006/03/if-only-they-...

Rahul.



RE: Is this right?
By TomZ on 3/8/2006 9:40:05 PM , Rating: 5
Rahul, your bias towards AMD is clear. I don't mind bias, but at the same time you deny you have any bias. You're either trying to fool us, or you're fooling yourself.

At the end of the day, your opinion is meaningless anyway. Consumers buy PCs and CPUs, and we will decide the market share between AMD and Intel. It is not for you to decide. Your job is to produce PCs that we want to buy, that include the processors we prefer.

So get to work - you aren't being paid to believe in the power of your dreams.


RE: Is this right?
By rahulsood on 3/8/2006 9:56:53 PM , Rating: 3
Rahul, your bias towards AMD is clear. I don't mind bias, but at the same time you deny you have any bias. You're either trying to fool us, or you're fooling yourself.

Well, it's true my bias is towards better technology. Around 80% of our notebooks go out with Intel processors, however. As far as workstations and desktops go, we'd be doing our customers a dis-service by making poor recommendations. We prefer the Pentium M for desktops over the Pentium 4.

At the end of the day, your opinion is meaningless anyway. Consumers buy PCs and CPUs, and we will decide the market share between AMD and Intel. It is not for you to decide. Your job is to produce PCs that we want to buy, that include the processors we prefer.

I totally hear what you're saying here - but technically the gaming market tends to influence the mainstream. This means that many people visit our site and see our recommendations and base their buying decisions on ours. This includes other people in the channel, people in I.T. departments, and gamers who build their own PCs. Our opinion does matter in that sense. ..but you are correct, our job is to produce PCs that people want to buy. It is also our job to ensure we are giving informed buying advice as I'm sure you can appreciate.

So get to work - you aren't being paid to believe in the power of your dreams.

Thanks :) How did I know that was going to bite me in the ass ;)



RE: Is this right?
By rahulsood on 3/8/2006 10:01:28 PM , Rating: 2
..also on that note I should mention that I've also been accused of being biased towards Nvidia (many many times). If you've read my battles with ATi over the years you'd know that I am only telling it from the inside. When you are dealing with these companies it's completely different from what you all see on the outside.

I am biased towards great technology no doubt, that's why we were first in line to launch Crossfire on RD580 - I love it! I am also looking forward to Nvidia's upcoming launch, but that doesn't take away any positive feelings towards ATi at the moment.

In any case, I do see your point and understand where you're coming from. Thanks for the advice ;)




Yea but...
By Marlin1975 on 3/8/2006 4:55:05 PM , Rating: 2
Was it not Intel that also showed off a p4 ruinning at 4Ghz YEARS ago and here we are with no 4ghz P4's.

So it looks good, but the REAL question is will it come to be as they, intel, want us to think.


Sell the sizzle, not the steak.