Anomalies in Anand's conroe's gaming benchmarks ?

[OcHungry]

Check the conroe benchmarks in the links and compare them to the same gaming benchmarks for E6700 in the conroe motherboards guide- You will see the discrepancies. The FPS are higher in the conroe/AM2 review compared to the FPS in the motherboards guide review. Why?
I am not a gamer and not an expert in graphic card benchmarks, but I am sure some of you can give us your expert opinion on why there are such discrepancies.
I am not accusing Anad, but I am curious if the benches were conducted in favor of conroe or were unintentional errors.
One thing I know is that, AMD's IMC was not utilized to its fullest. I believe Anad should have used DDR2 677 3-3-3-xx 1:1 ratio on AM2's. Something like 330x9=2970mhz.
There is no mention of Temperature during the load, or no stress test conducted for stability.
What if those numbers are only good for benchmarking but when it comes to stability and/or temp, conroe may not be as good as we think. What if conroe is not stable if overclocked?. This factor alone should give AMD the advantage. Because then, you can overclock AM2's and compensate the scores. So, I believe it was imperative that Anad had conducted stress test on both, stock and overclocking
Those who are interested in buying conroe would want to know if conroe is stable during load when overclocked, and want to know what temp. What about raid? Why didn't Anad test conroe in all raids settings?

I hope anadtech conduct more reviews w/ above concerns in mind.
anyway below are the 2 links. Please analyze and let us know.
Thanks.

Motherbords gaming benchmarks
Conroe/AM2 gaming benchmark

 

[BlueWeasel]

Anand
Anandtech
Got it?

 

[AyashiKaibutsu]

They didn't test the games the same way in the two different reviews. Quake 4 and BF2 uses different time demos. Half life 2 was tested with lost coast in one and episode one in another. You can't expect the FPSes to match up.

 

[SlitheryDee]

They were testing two different pieces of hardware. You can't expect comparable number between reviews of different types of hardware. Each one is designed to stress different parts of the computer.

 

[theprodigalrebel]

The Original Conroe Review is testing using X1900 Crossfire @ 1600x1200 WITHOUT AA.
The Motherboard Roundup is using both SLI 7900GTX/X1900 Crossfire but all of them are with 4X AA.

BTW, the graphs in the Motherboard Roundup are seriously difficult to read.

 

[Conky]

Isn't it pitiful how some fanboys must try to run down established sources of good information?

I will take Anand's word over the AMD4evarfanboy.com's opinion. I'm funny like that... only listening to established authorities on stuff like this.

Only idiots are arguing that the Conroe is not all it's advertised to be at this point.

 

[Vegitto]

Well, it depends. If you've sold your soul to a company, then yes, maybe it was cooked. If you're mentally sane, like most of us are, then you'll see that a lot of reputable sites come to the same conclusion and Anand does.

 

[Geomagick]

Also it was very likely that the testing for the different reviews and the machine setup was done by different people. The important thing is that within a single test the number of variables is kept to a minimum. Certainly from previous experience, indeed with 7 years of following Anandtech I have always found it to be an excellent and very much unbiased source of information.

 

[Ackmed]

Originally posted by: Crazyfool
Isn't it pitiful how some fanboys must try to run down established sources of good information?

I will take Anand's word over the AMD4evarfanboy.com's opinion. I'm funny like that... only listening to established authorities on stuff like this.

Only idiots are arguing that the Conroe is not all it's advertised to be at this point.

Well, you certainly lived up to you name, crazyfool.

The OP doesnt link to any sight, he simply thought they were run at the same settings. He doesnt say what CPU he has, or which company he likes better. The only one that looks like a "fanboy", is you.

 

[zsdersw]

On the contrary, I've regularly observed OcHungry's posts.. and always found them to have an AMD fanboy tint to them.

Particularly telling in this case is the title of the thread "Has Anandtech cooked conroe's benchmarks?" (spelling corrected) and this part from his original post: "I am not accusing Anand ..." (spelling corrected). He's doing what he said he's not. By framing the question in the way he did, he's making an accusation. If he were honest about his intentions, he would've titled the thread something like this: "Anomaly in Conroe benchmarks?"

 

[OcHungry]

Don?t worry about spelling or fanboysm, just answer the concerns.
I have looked at both reviews, and see both have same cpu, same motherboard, and same vid card, but different FPS. answer this, instead of evading the concerns.
Are you not worried about stability, raid, temp?
why use 1200x1600 if interested in finding cpu performance?
why not turn on 4xAA?
How come in the motherboard review Anandtech uses 3-3-3xx DDR2 800 but looser timing in Conroe/am2 review?
I have seen reviews show am2 5000 x2 compete w/ E6600, but Anand says E6300 can compete w/ Fx62 or do better.
Why not use DDR2 667 and use lower multiplier, like 325x9=2.925ghz close to E6800.
1:1 ratio will improve AMD quite a bit. If using DDR2 800 3-3-3xx for motherboard review, why didn?t Anand use the same timings for the cpu review? 400x7=2800mhz would make FX62 just perfect and faster. Did Anand use this setting?
Let's be honest, can you overclock conroe to 4ghz w/ heatpipe?. All the reviews I've read had to use phase change. It is interesting that no "one" review, says anything about temp and stress tests, specially when overclocked.
If you are not afraid to tackle these concerns, do everyone favor and express your concern and thoughts. But if you are not willing to be objective, then go ahead evade the concerns and pick on my spelling and the title of this thread.
But if you are buying conroe, I am sure those concerns I have brought to your attention are worrying you.

 

[zsdersw]

Other people have talked about your "concerns". I, on the other hand, chose to talk about something else that's relevant to this topic.

I'm getting a Core 2 Duo, but I'm not the least bit concerned with overclocking. Nothing is worrying me about Core 2 Duo.

 

[Viditor]

Dear OP...

I'm really not that concerned at all (of course I doubt I will be getting a Conroe either...).
I have found Anand to be one of the most objective and unbiased reviewers on the Web!
That's not to say he doesn't make mistakes (as we all do), but he also admits to them and fixes them when he can (though you may have missed this fact as his mistakes happen so rarely!).
I do object to the tonality of your Topic Title though...suggesting that Anand has cooked the benchmarks is both inflamatory and rude, and it certainly reduces the chance that anyone will read what you have to say with an open mind (I know that I find it hard to do so...).
May I suggest you edit the topic title...

 

[OcHungry]

Originally posted by: Viditor
Dear OP...

I'm really not that concerned at all (of course I doubt I will be getting a Conroe either...).
I have found Anand to be one of the most objective and unbiased reviewers on the Web!
That's not to say he doesn't make mistakes (as we all do), but he also admits to them and fixes them when he can (though you may have missed this fact as his mistakes happen so rarely!).
I do object to the tonality of your Topic Title though...suggesting that Anand has cooked the benchmarks is both inflamatory and rude, and it certainly reduces the chance that anyone will read what you have to say with an open mind (I know that I find it hard to do so...).
May I suggest you edit the topic title...

I fixed the title. I didn?t know it can be fixed. I apologies if I offended you or anyone. That wasn?t my intention- although I am a bit upset about the degree of pumping Intel and the rudeness toward members using AMD .
Same treatment did not occur on AMD's K8, and Tom's hardware never created a table suggesting netburst should drop in price by 80 % to compete w/ K8's price/performance. Empire Strikes back, Game is over, New king is born, are among those titles that ignores sensitivity of forum's members.
In this regard TG and Anandtech were careless and rude also (just as the tile of this thread reflected same degree of insensitivity).
Having said that and changing the tone of the OP,
Let's see if anyone cares to respond to the concerns.
At least point me (and others w/ same concerns) to some kind of explanation why there are such differences in FPS, and provide link(s) to some kind of stability test on conroe (specially 4ghz on air).

 

[HoptoIt]

Originally posted by: OcHungry

Originally posted by: Viditor
Dear OP...

I'm really not that concerned at all (of course I doubt I will be getting a Conroe either...).
I have found Anand to be one of the most objective and unbiased reviewers on the Web!
That's not to say he doesn't make mistakes (as we all do), but he also admits to them and fixes them when he can (though you may have missed this fact as his mistakes happen so rarely!).
I do object to the tonality of your Topic Title though...suggesting that Anand has cooked the benchmarks is both inflamatory and rude, and it certainly reduces the chance that anyone will read what you have to say with an open mind (I know that I find it hard to do so...).
May I suggest you edit the topic title...

I fixed the title. I didn?t know it can be fixed. I apologies if I offended you or anyone. That wasn?t my intention- although I am a bit upset about the degree of pumping Intel and the rudeness toward members using AMD .
Same treatment did not occur on AMD's K8, and Tom's hardware never created a table suggesting netburst should drop in price by 80 % to compete w/ K8's price/performance. Empire Strikes back, Game is over, New king is born, are among those titles that ignores sensitivity of forum's members.
In this regard TG and Anandtech were careless and rude also (just as the tile of this thread reflected same degree of insensitivity).
Having said that and changing the tone of the OP,
Let's see if anyone cares to respond to the concerns.
At least point me (and others w/ same concerns) to some kind of explanation why there are such differences in FPS, and provide link(s) to some kind of stability test on conroe (specially 4ghz on air).

Have you written any of the editors yet? Are you doing this just create trouble? Have your visited HardlyOCP lately to see what should be considered rude or careless?

"The whole "real world" "H" benchmarking process is about as accurate as Kyle trying to measure the differences in velocity and volume of expelled gas generated each time his dog farts. "

 

[Viditor]

Originally posted by: OcHungry

Originally posted by: Viditor
Dear OP...

I'm really not that concerned at all (of course I doubt I will be getting a Conroe either...).
I have found Anand to be one of the most objective and unbiased reviewers on the Web!
That's not to say he doesn't make mistakes (as we all do), but he also admits to them and fixes them when he can (though you may have missed this fact as his mistakes happen so rarely!).
I do object to the tonality of your Topic Title though...suggesting that Anand has cooked the benchmarks is both inflamatory and rude, and it certainly reduces the chance that anyone will read what you have to say with an open mind (I know that I find it hard to do so...).
May I suggest you edit the topic title...

I fixed the title. I didn?t know it can be fixed. I apologies if I offended you or anyone. That wasn?t my intention- although I am a bit upset about the degree of pumping Intel and the rudeness toward members using AMD .
Same treatment did not occur on AMD's K8, and Tom's hardware never created a table suggesting netburst should drop in price by 80 % to compete w/ K8's price/performance. Empire Strikes back, Game is over, New king is born, are among those titles that ignores sensitivity of forum's members.
In this regard TG and Anandtech were careless and rude also (just as the tile of this thread reflected same degree of insensitivity).
Having said that and changing the tone of the OP,
Let's see if anyone cares to respond to the concerns.
At least point me (and others w/ same concerns) to some kind of explanation why there are such differences in FPS, and provide link(s) to some kind of stability test on conroe (specially 4ghz on air).

Good move on the title...
What I see between the 2 reviews is that they were conducted by different people using different conditions, so comparing one SET of benchmarks against the other doesn't really give a meaningful result. You should only compare within a benchmark run
I agree that what was printed leaves the possibility that the benchmarks were skewed, but that's where trust of the reviewer comes in...there's really nothing to show that they WERE skewed either. If the review were from THG (shiver), then I would be drowning myself in salt while reading...but both Anand and Wesley (and Gary) have a long history of being trustworthy in their reporting practices, and (at least for me) they have earned the benefit of the doubt.
For another fairly trustworthy review and comparison, have a look at the X-Bit review listed here.

 

[DrMrLordX]

Sensitivity of forum members? Cmon. If we're "sensitive" to people ribbing one company or other when we aren't even necessarily employees or owners/stockholders of said companies, then we're overly sensitive in general. Anandtech's article didn't have any inflammatory language in it at all.

I gotta agree with Viditor's assessment of the situation. You're making something out of nothing. Too many people here are taking the shift of technological superiority personally.

 

[KeypoX]

Originally posted by: Crazyfool
Isn't it pitiful how some fanboys must try to run down established sources of good information?

I will take Anand's word over the AMD4evarfanboy.com's opinion. I'm funny like that... only listening to established authorities on stuff like this.

Only idiots are arguing that the Conroe is not all it's advertised to be at this point.

your wisdom is .... well yeah only idiots question "authorites"

 

[OcHungry]

Here is another review that uses identical hardware as Anand's, but shows practically dead even when using 1200x1600 (same as Anand's resolusion) but in 4xAA, AFx8.
Now I see why?
http://firingsquad.com/hardware/intel_core_2_performance/page8.asp

BTW this is the 4th review site I found showing similar results. The gaming is dead even, and the rest of benchmarks, conroe E6800 is 9-10% better performance.
Nothing like the review here showing E6800 25%-30% faster.
I am still waiting fr your responds, and links.
I also would like to see another website official review showing conroe can be overclocked to 4ghz on air.
A months from now I will be looking for those who bought retail version and see if they were able to OC to 4ghz on air w/ proof of prime95.
Good luck.

 

[Bobthelost]

Look at the Oblivion testing situation. If you were to run the benchmark for your CPU near the oblivion gate then you'll see no performance increase. If you run it in town then conroe would spank the Fx-62 like an italian with a german dominatrix after the semifinals...

It's all about the benchmarks themselves. Even with the exact same runs you get variation in the test results, most/all sites run them several times to average them out.

 

[Keysplayr]

Originally posted by: OcHungry
Here is another review that uses identical hardware as Anand's, but shows practically dead even when using 1200x1600 (same as Anand's resolusion) but in 4xAA, AFx8.
Now I see why?
http://firingsquad.com/hardware/intel_core_2_performance/page8.asp

BTW this is the 4th review site I found showing similar results. The gaming is dead even, and the rest of benchmarks, conroe E6800 is 9-10% better performance.
Nothing like the review here showing E6800 25%-30% faster.
I am still waiting fr your responds, and links.
I also would like to see another website official review showing conroe can be overclocked to 4ghz on air.
A months from now I will be looking for those who bought retail version and see if they were able to OC to 4ghz on air w/ proof of prime95.
Good luck.

OP: Here is a quick chart showing similarities, differences, and unknown's between AT, and FS article you linked to.

FS & AT

You will note that ANY difference, can make a significant impact on results. Especially the unknowns, and there could be a dozen differences not listed in the test platform details.
A single difference is enough to discount a comparison between two review sites. Everybody knows this. Well, almost.

 

[Viditor]

Originally posted by: OcHungry
Here is another review that uses identical hardware as Anand's, but shows practically dead even when using 1200x1600 (same as Anand's resolusion) but in 4xAA, AFx8.
Now I see why?
http://firingsquad.com/hardware/intel_core_2_performance/page8.asp

BTW this is the 4th review site I found showing similar results. The gaming is dead even, and the rest of benchmarks, conroe E6800 is 9-10% better performance.
Nothing like the review here showing E6800 25%-30% faster.
I am still waiting fr your responds, and links.
I also would like to see another website official review showing conroe can be overclocked to 4ghz on air.
A months from now I will be looking for those who bought retail version and see if they were able to OC to 4ghz on air w/ proof of prime95.
Good luck.

Of course the reason is that at the higher resolutions, the game becomes GPU bound, not CPU bound...but your point is a valid one. If you play games at high resolution and gaming benchmarks are your reason for upgrade, then the CPU becomes much less important for your decision.
An intersesting thought though, I wonder if this still holds true in SLI...? If it doesn't and SLI takes the bottleneck up another few notches, then I would think that Intel really needs to get an SLI capable platform out ASAP!

 

[Keysplayr]

Originally posted by: Viditor

Originally posted by: OcHungry
Here is another review that uses identical hardware as Anand's, but shows practically dead even when using 1200x1600 (same as Anand's resolusion) but in 4xAA, AFx8.
Now I see why?
http://firingsquad.com/hardware/intel_core_2_performance/page8.asp

BTW this is the 4th review site I found showing similar results. The gaming is dead even, and the rest of benchmarks, conroe E6800 is 9-10% better performance.
Nothing like the review here showing E6800 25%-30% faster.
I am still waiting fr your responds, and links.
I also would like to see another website official review showing conroe can be overclocked to 4ghz on air.
A months from now I will be looking for those who bought retail version and see if they were able to OC to 4ghz on air w/ proof of prime95.
Good luck.

Of course the reason is that at the higher resolutions, the game becomes GPU bound, not CPU bound...but your point is a valid one. If you play games at high resolution and gaming benchmarks are your reason for upgrade, then the CPU becomes much less important for your decision.
An intersesting thought though, I wonder if this still holds true in SLI...? If it doesn't and SLI takes the bottleneck up another few notches, then I would think that Intel really needs to get an SLI capable platform out ASAP!

I have heard that Intel and Nvidia do not have a good relationship. In light of the possible AMD/ATI merger, I hope Intel/NV can work things out. This might be an incentive to do so. I dunno. I wonder if Intel actually has a separate graphics team assembled to get back into the discrete GPU biz. Not that they were ever in it enough to matter, but I wonder.

 

[Bobthelost]

Isn't the X1900XTX crossfire more powerful than 7900GTX SLI though? I got the impression that with the current drivers that the ATI option was starting to use those massive transistor arrays to good effect at last.

In which case Intel and it's Crossfire compatible boards are the way forward

 

[Gary Key]

Originally posted by: OcHungry
Check the conroe benchmarks in the links and compare them to the same gaming benchmarks for E6700 in the conroe motherboards guide- You will see the discrepancies. The FPS are higher in the conroe/AM2 review compared to the FPS in the motherboards guide review. Why?
I am not a gamer and not an expert in graphic card benchmarks, but I am sure some of you can give us your expert opinion on why there are such discrepancies.
I am not accusing Anad, but I am curious if the benches were conducted in favor of conroe or were unintentional errors.
One thing I know is that, AMD's IMC was not utilized to its fullest. I believe Anad should have used DDR2 677 3-3-3-xx 1:1 ratio on AM2's. Something like 330x9=2970mhz.
There is no mention of Temperature during the load, or no stress test conducted for stability.
What if those numbers are only good for benchmarking but when it comes to stability and/or temp, conroe may not be as good as we think. What if conroe is not stable if overclocked?. This factor alone should give AMD the advantage. Because then, you can overclock AM2's and compensate the scores. So, I believe it was imperative that Anad had conducted stress test on both, stock and overclocking
Those who are interested in buying conroe would want to know if conroe is stable during load when overclocked, and want to know what temp. What about raid? Why didn't Anad test conroe in all raids settings?

I hope anadtech conduct more reviews w/ above concerns in mind.
anyway below are the 2 links. Please analyze and let us know.
Thanks.

Motherbords gaming benchmarks
Conroe/AM2 gaming benchmark

1. The benchmark demos and resolution settings used in the CPU article are different. Please read the review setups again and look at the chart titles. The CPU article was at 1600x1200, motherboard article is at 1600x1200 4xAA/8xAF where applicable. The demos used are different also as noted on the chart titles. Until we revise our motherboard demos to the newly developed internal demos this will create a difference in the reported scores. The main video card used in the motherboard testing is a 7900GTX compared to the X1900XT in the CPU article.

2. An upcoming memory article will show what happens when you run the AM2 platform at the same settings as Conroe. We have shown time and time again that the extra bandwidth on AM2 is not fully utilized due to limitations within the CPU at this time.

3. Extensive stress testing was done throughout our article preparation and is continuing on a daily basis. Our upcoming cooling guide will show temps between platforms at various load and overclocking levels as stated in our guide article.

4. It does not matter how much you overclock AM2, it is not going to make up the difference in performance between the two platforms when overclocked and in most cases when Conroe remains at stock levels. Face the facts, Intel has a better overall performing CPU at "this" time. It does not mean Athlon64 is trash, in fact it means we will see terrific prices on what has been and continues to be an excellent CPU family. This is good news for everyone and if you are an AMD fan, then your ability to upgrade just became a whole lot easier. If you are an Intel fan, you no longer have to worry about performance concerns when comparing your system to an Athlon64 unit. If you are perfomance fan, Intel has the CPU for you at this time, just like AMD has for the last three years. This is good news for consumers and who knows, the tables could switch again this time next year.

5. Conroe was extremely stable during the stock voltage overclocks, more so than our current AM2 platform. At the extreme air overclock, the system remained stable enough to complete SuperPI, game benchmarks, and 3DMarks. As I stated in this forum we are finding at the 3.85GHz range our system to be dual prime stable currently. We have an updated Asus P5W-DH designed for extreme overclocking arriving this week for testing and will report our final results in an overclocking article.

6. RAID testing is not looked at in a CPU article. That is for the storage guy (me) and will be upcoming in our second guide as stated in that article.