Anand re-examines Conroe benchmarks

jebo_4jc

[H]ard|DCer of the Month - April 2011
Joined
Apr 8, 2005
Messages
14,566
These numbers make a lot more sense:
http://www.anandtech.com/tradeshows/showdoc.aspx?i=2716

They apparently addressed the old bios issue on the DFI board, and corrected a mistake they made when benchmarking FEAR (they had 12x10 set on the AMD and 10x7 set on Conroe). Now, the numbers basically show Conroe beating AMD by about 20% consistently. This seems realistic to me. Maybe a newer chipset would increase AMD's performance a bit, but I would say 5% max.

Keep in mind that the XE will provide about 12% more clockspeed than the 2.66ghz, so let's say the XE will beat the FX-62 by about 30-ish%.
 
Saw this earlier. This should put an end to the BS from the AMD crowd. Well, most of them. ;)
 
Props to Anandtech for correcting mistakes. A lot of sites would just let it go.

Still very impressive. I can't wait for the ensuing price war.
 
Even with the playing feilds leveled out, Conroe still gives AMD a beating. I know what I'm getting this fall.
 
But noes!!! AMD might release a 3.6 GHz FX-70 the day after Conroe launches!!! Just because they're not bragging about their performance, does that mean it isn't there???? They've probably got an army of 4 GHz quad-core Athlons waiting in the wings!!!!!!!
 
I'll be interested to see how this compairs to the new amd socket cpus comming out soon but those numbers look real impressive. Wonder how it willl overclock?
 
LstOfTheBrunnenG said:
But noes!!! AMD might release a 3.6 GHz FX-70 the day after Conroe launches!!! Just because they're not bragging about their performance, does that mean it isn't there???? They've probably got an army of 4 GHz quad-core Athlons waiting in the wings!!!!!!!

I don't know...I have it on reliable sources that they have used all their profits to keep their plans secret. Apparently, they plan on asking the salvation army for funds to do the development part.
 
swatbat said:
I'll be interested to see how this compairs to the new amd socket cpus comming out soon but those numbers look real impressive. Wonder how it willl overclock?
IMHO I wouldn't expect much for socket AM2. AMD CPUs have never been very memory bandwidth hungry, so DDR2 won't likely introduce much of an improvement.

We will have to see how Conroe will overclock. If I had to guess, I would bank on 10-15%, and if it happens to be an overclocking monster, it will be a pleasant surprise :). The one thing the 2.4/2.6ghz Conroe CPUs have going for them is the XE uses a faster FSB, which means motherboards will already be able to support FSB overclocking for the 1066FSB CPUs, and you can get some fast DDR2, so you will likely only be held back by the absolute maximum of your CPU. That is a nice situation.
 
LstOfTheBrunnenG said:
But noes!!! AMD might release a 3.6 GHz FX-70 the day after Conroe launches!!! Just because they're not bragging about their performance, does that mean it isn't there???? They've probably got an army of 4 GHz quad-core Athlons waiting in the wings!!!!!!!

You do know that you are no better than the amd !!!!!!!s right?
 
Yeah Conroe beats AMD...but...with a 4 core chip. Intel has 2 extra cores and 2mb more of cache to beat a 2 core 2mb cpu by 20%.

I dont' know about you but to me thats not awesome. The conroe gets 10% more performance per extra core. Why is everyone freaking out over that? That is a dismal performance/comparison.

Its 4 cores & 4mb mem Vs. 2 cores & 2mb mem
 
kraggy said:
Yeah Conroe beats AMD...but...with a 4 core chip. Intel has 2 extra cores and 2mb more of cache to beat a 2 core 2mb cpu by 20%.

I dont' know about you but to me thats not awesome. The conroe gets 10% more performance per extra core. Why is everyone freaking out over that? That is a dismal performance/comparison.

Its 4 cores & 4mb mem Vs. 2 cores & 2mb mem
Conroe is actually a dual core processor. The L2 cache is twice as large, but it does not matter; it is one of the features of the Conroe architecture. And chances are the 2 MB one will perform similiar but a bit slower being that when they went from Banias to Dothan, there was not a huge change in performance. AMD's architecture is also a bit more mature. Come the 45nm variant of Conroe, I bet we'll see large performance boosts from the new core...Empoy seems excited about that core but that might be me reading between the lines.
 
kraggy said:
Yeah Conroe beats AMD...but...with a 4 core chip. Intel has 2 extra cores and 2mb more of cache to beat a 2 core 2mb cpu by 20%.

I dont' know about you but to me thats not awesome. The conroe gets 10% more performance per extra core. Why is everyone freaking out over that? That is a dismal performance/comparison.

Its 4 cores & 4mb mem Vs. 2 cores & 2mb mem

Conroe is only 2 cores with 4MB cache
 
^^ the benches were 2 core vs 2 core ? i didnt see any quad core cpu's being compared.......
 
Anyone noticed that the Conroe was running on the Intel 975X "Bad Axe" board


I wonder if this Bad Axe was an engineering sample.. or is it one that is currently in the retail markets?? If thats the case, its an added plus for anyone who has purchased the Bad Axe recently
 
Great follow-up article. This is why I trust Anandtech.com so much because they take their time and they make sure they get everything right. The 2.66GHz Conroe still beats the overclocked FX-60 by at least 20% in gaming benchmarks, and just think what the Conroe XE version will do! :eek:
 
LstOfTheBrunnenG said:
They've probably got an army of 4 GHz quad-core Athlons waiting in the wings!!!!!!!


LOL... doesn't quite work that way chief. If it did, don't you think that Intel, with all their R&D money, would have played the bitch of spades on AMD 2 years ago? Yea... maybe AMD will release a competitive chip next year to combat Conroe maybe they won't. What I do know is that there is NO info on AMD's next gen shit. Maybe they are tight lipped about it.... but my guess is that they just don't have anything that great waiting in the works. Time will tell...
 
jlbenedict said:
Anyone noticed that the Conroe was running on the Intel 975X "Bad Axe" board
I wonder if this Bad Axe was an engineering sample.. or is it one that is currently in the retail markets?? If thats the case, its an added plus for anyone who has purchased the Bad Axe recently


Its here at Newegg if you want it. Id like to give it a try I know that.
 
jlbenedict said:
I wonder if this Bad Axe was an engineering sample.. or is it one that is currently in the retail markets?? If thats the case, its an added plus for anyone who has purchased the Bad Axe recently

From what I understand there is a later SKU of silicon and later board rev that will allow Conroe, even the XE @1333fsb, to run. Not sure if that is currently released, but I do know that the changes to get the regular version of conroe to run is minimal and could probably be done by the consumer if they have the ability to solder well provided they are willing to void the warrenty on their 300 dollar board. :D I think that it's the 1333fsb support that requires new silicon.
 
PaHick said:
Its here at Newegg if you want it. Id like to give it a try I know that.

The Broad Axe Intel motherboard used during testing was a modified 975X motherboard. The new revision 975X motherboards will support Conroe as will "Broad Water" AKA the 965 chipset. Abit was the only motherboard vender showing off a 965 chipset motherboard.
 
The Doc said:
Abit was the only motherboard vender showing off a 965 chipset motherboard.

Bet they didn't have it running did they? :p Do you know what ICH they are using, ICH7 or 8?
 
The Doc said:
Great follow-up article. This is why I trust Anandtech.com so much because they take their time and they make sure they get everything right. The 2.66GHz Conroe still beats the overclocked FX-60 by at least 20% in gaming benchmarks, and just think what the Conroe XE version will do! :eek:

I have a little more respect for him now. Please note though, he or they did what their most AMD Fanbase wanted done. NOT what was fair. Look at their forums? It is just a bunch of AMD Biased folks jumpping every setting they'd thought put AMD at a disadvantage. These same folks routinely condone handicapped Intel systems. Example Wes Finks overclocked Hammer vs 2.4C as he complained about how unstable a 2.4C is when overclocked :rolleyes:

http://www.dailytech.com/article.aspx?newsid=1170

Rahul said:
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.

The AMD system had CAS2 2-2-5 1T RAM and Anand updated the BIOS. It still got spanked, sometimes even worse LOL! This was one of the reasons for the update and why I said he was merely digging a deeper hole.
 
The Bad Axe Intel 975 motherboard revision 303 or higher will support Conroe. The intial revision of the board needed a few MINOR tweaks, caps, etc. to make it conroe compatible. If you buy that motherboard in a few weeks it will probably be the right revision, though I'd want to look at the screenprint to be sure.
 
Still doesn't change the fact that Conroe totally 0wns AMD's best. :D Very interesting times ahead... :)
 
Reality is reality. A few trivial settings, BIOS details, and other errata that Rahul and the rest of AMD-nation chose to cling to as a last hope for some sort of parity with Intel's new speed demon have been invalidated thanks to Anand pursuing the issue appropriately.

Conroe is the real deal, as real as it appeared two days ago. Possibly even moreso.

Going to be a great second half of the year in terms of new hardware to lust over that's for sure. :)
 
Alright I cry foul Why didn't Intel use DDR2 667 @ 3-2-2-8 timings I just thinks its immoral how they are skewing the results of their own conroe .

Also Ananda statement that that DDR800 would have no impact on conroe. Unless he tested it @ DDR800 how can he make such a statement. It impacts the old P4 and thats a fact. So anada is saying that AM2 will have zero affects on the AMD cpu @ DDR2 800 . Since it has zero effect on conroe it surely can not benefit AMD a much inferior cpu.

I got to go back over to AT forum its bring tears of joy to my eye's . To see all the cring going on over there over a cpu that won't be out as they say for a 1 -1 1/2 years . The damage controll is unbelievable . These are the same assholes the cruicified rollo.

If I can work up the guts I might even post but it could give me away.
 
The Doc said:
Great follow-up article. This is why I trust Anandtech.com so much because they take their time and they make sure they get everything right. The 2.66GHz Conroe still beats the overclocked FX-60 by at least 20% in gaming benchmarks, and just think what the Conroe XE version will do! :eek:

wow i would hardly trust them now. what other tests have these brainiacs screwed up? it is certainly not rocket science to test computer hardware.
 
Memory bandwidth affects the Pentium 4 architecture because it had 20 or 31 stage pipelines, which need to be constantly saturated with data.

The Athlon architectures have always had relatively short pipelines (10 stage, I think the Opty/Ath64 pipelines have 11?) which aren't memory bandwidth hungry because a pipeline flush does not affect it as heavily as it would the long pipelines of the Pentium 4 processors.

The Conroe architecture also features a shorter pipeline a la the Athlon architecture, so Anand is just inferring that memory speed won't affect it and the AM2 Athlons much because of historical evidence.
 
BillLeeLee said:
Memory bandwidth affects the Pentium 4 architecture because it had 20 or 31 stage pipelines, which need to be constantly saturated with data.

The Athlon architectures have always had relatively short pipelines (10 stage, I think the Opty/Ath64 pipelines have 11?) which aren't memory bandwidth hungry because a pipeline flush does not affect it as heavily as it would the long pipelines of the Pentium 4 processors.

The Conroe architecture also features a shorter pipeline a la the Athlon architecture, so Anand is just inferring that memory speed won't affect it and the AM2 Athlons much because of historical evidence.
That was the reason I posted as I did . But for Anand to say that is wrong . He knows and I know AM2 will offer no benefits to AMD cpu's . AS you yourself said historical evidence proves this out . because of the added latency of DDR2 it may even hurt.

Were Conroe is concerned there is no historical evidence to back up his claim unless there is a 4 issue cpu tech I am not aware off. and some other improvements made to this cpu never before seen.

A lot of people are crying foul rig test. I do agree the test is rigged all the evidence points to the fact that Intel did everthing they could to show AMD in its best light. It looks to me like intel pulled the ruins in on the Conroe stallion and didn't let it run at full speed .

That evidence is very clear.
 
Ah yes, now I see your point. Yes, I do agree this test must be taken with giant grains of salt.

I am excited but reserved about these, because they are boxes provided by Intel and we pretty much won't see anything more credible, as you will, until around the launch time. We'll just have to hold on until then, I guess.
 
jebo_4jc said:
I'm pretty sure his post was a joke.

Wow, only one of them got it. I guess you have to be careful what you say during wartime around here.
 
LstOfTheBrunnenG said:
Wow, only one of them got it. I guess you have to be careful what you say during wartime around here.

I wouldn't worry about it. Sometimes you just have to accept, these are tough crowds.
 
BillLeeLee said:
Ah yes, now I see your point. Yes, I do agree this test must be taken with giant grains of salt.

I am excited but reserved about these, because they are boxes provided by Intel and we pretty much won't see anything more credible, as you will, until around the launch time. We'll just have to hold on until then, I guess.

I agree with the other guy. IMHO, Intel bent over backwards to put AMD in the best light. Maybe they remember the AMD white papers verified by the same folks who cooked Enron's books. Then folks, guys they didn't have deals with, talked about how poorly the Intel rigs were setup by AMD and made AMD look even worse. Maybe Intel wanted NO parts of that, hell, AMD doesn't even do that (poor set up) anymore. With Athlon64, they didn't have to.

To shut up Rahul and the loudest of the AMD followers mouths, maybe as many sites as possible should set up a very similar system and cross check numbers? Hell, how long would that take? I think Anand didn't think Intel was cheating and if he did, we'd all know LOL! Of course it's going to be hard to convince folks things have changed. Many didn't believe how much ass Dothan kicked.

I got into a bit a of trouble with Chris and others when I said 75% of the folks here would still be AMD biased even if Conroe kicked the crap out of anything AMD had at that time. I hope I'm proven wrong.

Why do I like [H]?

Front Page said:
While there were some concerns raised yesterday over the wrong BIOS being used on the AMD machines used by Intel, the initial benchmarks run by Anandtech were reran with an updated BIOS and the new results came out basically the same as the initial findings.

Some sites are skipping this news. I'm glad [H] didn't.
 
LstOfTheBrunnenG said:
Wow, only one of them got it. I guess you have to be careful what you say during wartime around here.

Naw, I think a lot of folks saw it as a Joke. Didn't need 10 folks telling this person it was a joke.
 
$BangforThe$ said:
Alright I cry foul Why didn't Intel use DDR2 667 @ 3-2-2-8 timings I just thinks its immoral how they are skewing the results of their own conroe .

Also Ananda statement that that DDR800 would have no impact on conroe. Unless he tested it @ DDR800 how can he make such a statement. It impacts the old P4 and thats a fact. So anada is saying that AM2 will have zero affects on the AMD cpu @ DDR2 800 . Since it has zero effect on conroe it surely can not benefit AMD a much inferior cpu.
...

Intel doesn't do this sort of things.

They always test their system with RAM set at SPD value..

For example , if you look SPEC INT and FP Intel system latencies are 4-4-4-12@533/400 ( for Xeons ) while AMD sets 2-2-2-5-1T

Even when Intel used DDR , they tested at 3-3-3-8 DDR400.
 
jlbenedict said:
Anyone noticed that the Conroe was running on the Intel 975X "Bad Axe" board


I wonder if this Bad Axe was an engineering sample.. or is it one that is currently in the retail markets?? If thats the case, its an added plus for anyone who has purchased the Bad Axe recently

I am wondering the same too. Will the current retail "BadAxes" support Conroe with a BIOS update.

I will probably just wait til the Conroe is out well, no reason to gamble.
 
Back
Top