"C2D Architecture, Cheating Tips From Reviewers"

[9-inch]

Scientia (from AMDzone) posted a great post about some trends we have being seeing from benchmarks on conroe:

...Thus far, I have been completely unable to find any SuperPi scores that show comparable 2MB C2D, 4MB C2D, and FX. Just attempting to judge from some high scores the advantage is at least 15%. One website ran two copies of SuperPi to evenly stress both cores. They used this to show that C2D used less power than K8. Stressing both cores is also a problem since running two copies of the same benchmark would allow C2D to have only a single copy in L2 and maximize its large cache. While the claim is made frequently and loudly that an inexpensive C2D will beat an FX, I have yet to see any review that ran a separate benchmark on each core simultaneously to stress the cores and cache. This would be the best test of real world use but I haven't seen it done. The procedure really isn't that difficult. You run a continuous application on one core and a second timed application on the second core. Then you compare the scores with a single instance and nothing running on the second core. Not difficult. Of course, C2D would take a hit because of cache dividing and another one if the second application accessed memory because of lower memory bandwidth and cache thrashing. The question is how much of a drop. Perhaps a cheat hasn't been found to prevent this yet.

Other tricks are common such as using 4-4-4 timing for AMD because C2D doesn't benefit from the faster timing. Other tricks include praising C2D's higher IPC and then making sure that the C2D chip is clocked higher. These cheats are a continuation from earlier when Yonah chips were overclocked via the FSB which made both memory and processor clock faster and these were compared with FX's which clocked the cpu higher but had lower proportionate memory increases...

Read the whole thread here:
www.amdzone.com

 

[ltcommander_data]

Enough already. If you want multitasking benchmarks, try and spend some time and find them yourself:

http://www.tomshardware.com/2006/07/14/core2_duo_knocks_out_athlon_64/page17.html

You want low memory timings, here:

http://www.anandtech.com/memory/showdoc.aspx?i=2800&p=7

They used the lowest timings possible:

Our Corsair CM2x1024-6400C3 modules were set to the following memory timings on each platform; DDR2-400 - 3-2-2-5, DDR2-533 - 3-2-2-6, DDR2-667 - 3-2-3-7, DDR2-800 - 3-3-3-9, DDR2-1067 - 4-3-4-11, and DDR2-1112 - 5-4-5-14.

I can't believe people are complaining about DDR2 800 with 4-4-4 timings. First, they whined that DDR2 800 with 5-5-5 timings was cheating and told everyone to use 4-4-4 timings. Now that people use 4-4-4 timings, they complain about 3-3-3 timings. Well the results are above and it doesn't help anything. Furthermore, I don't think anyone offers DDR2 800 with 3-3-3 timings anyways. I believe the best is Corsair with 3-4-3.

You say Yonah is cheating with overclocked FSB:

http://www.anandtech.com/memory/showdoc.aspx?i=2800&p=8

Here is a FX62 overclocked to X6800 speeds. This time it's the FX62 that has the overclocked HT advantage. Again, doesn't change much. You really aught to find something better to post instead of spamming the forum with threads of little use.

 

[9-inch]

Enough already. If you want multitasking benchmarks, try and spend some time and find them yourself:

http://www.tomshardware.com/2006/07/14/core2_duo_knocks_out_athlon_64/page17.html

You want low memory timings, here:

http://www.anandtech.com/memory/showdoc.aspx?i=2800&p=7

They used the lowest timings possible:

Our Corsair CM2x1024-6400C3 modules were set to the following memory timings on each platform; DDR2-400 - 3-2-2-5, DDR2-533 - 3-2-2-6, DDR2-667 - 3-2-3-7, DDR2-800 - 3-3-3-9, DDR2-1067 - 4-3-4-11, and DDR2-1112 - 5-4-5-14.

I can't believe people are complaining about DDR2 800 with 4-4-4 timings. First, they whined that DDR2 800 with 5-5-5 timings was cheating and told everyone to use 4-4-4 timings. Now that people use 4-4-4 timings, they complain about 3-3-3 timings. Well the results are above and it doesn't help anything. Furthermore, I don't think anyone offers DDR2 800 with 3-3-3 timings anyways. I believe the best is Corsair with 3-4-3.

You say Yonah is cheating with overclocked FSB:

http://www.anandtech.com/memory/showdoc.aspx?i=2800&p=8

Here is a FX62 overclocked to X6800 speeds. This time it's the FX62 that has the overclocked HT advantage. Again, doesn't change much. You really aught to find something better to post instead of spamming the forum with threads of little use.

You sound like you've being hit hard in the face from someone.

Don't treat your self like that man, just chill.

 

[9-inch]

The problem here is that you are not posting news worthy, or discussion worthy information --- in fact, you are posting material from a previously banned individual, which, if not against the forum rules, should be.

What are you talking about??
That guy never joined this forum (I guess so), but if you're talking about SexBomb, then that's something else. We all know that sexBomb used to "copy-paste" this guy's posts to annoy many of the intel fanboys. (I enjoyed that). Too bad he was savagely banned from this forum without doing anything to anyone (even Action_Man merits being ban for his "spectacular" behaviour), but anyhow.... :roll:

 

[9-inch]

Unless otherwise noted as plagiarism, Scentia is Sexbomb represented by this work.... the collective reasoning is the same.... and posting dribble of this caliber is hardly worthy of discussion.

This thread should be locked.

Why?? because you say so?? :roll:
I didn't know that you turned an admin. Even some members that have more time/posts in this forum than you should merit such a high responsability. Someone open-minded and free of any bias, not an intel fanboy like you.

Anyhow, I'd like to know what the members of this forum think about the above posted (the topic) since we all know that most of these benchmarks are in "some way" kinda rigged (my personal thought).

 

[Crimson_Yachiru]

it should be locked, not becuase what you post isn't discussion worthy, but because you don't defend your positions; you just attack other people without any reason...

a$$hole...

 

[Action_Man]

I love how you STILL pretend that sexbomb was a different person. :lol: 8)

Further proof 9nm is parrot?

 

[RichPLS]

X-Bit Labs wrote: ›The average performance improvement we have seen from Athlon 64 FX-62 equaled 16%, while Core 2 Extreme X6800 demonstrated only 10% average performance boost. This way, there is a certain difference: AMD K8 turns out 6% mode efficient in 64-bit mode than Intel Core.

It is sad when Amdroids are reduced to praising their saviour on the basis that their beloved idol only loses by 10% due to an inefficient CPU design. :lol: :lol:

 

[BaronMatrix]

If you look closely at Anands VERY THOROUGH TESTS where it shows Core2 slightly faster clock for clock with the same memory timings, you will see a more linear scaling with K8 than with Core 2 as it is a lot flatter with clockspeed and load.

Please let next week get here.

 

[jaguarskx]

9-inch, I think you are just posting a SH!T post from some dumbass poster from another site who is desperate to refute anything about Conroe.

But the reviews and articles themselves can be misleading or simply incorrect. For example, Anandtech just posted an OC review. http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2802 The problem is that this review uses stepping 5 and erroneously calls these true production chips.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2802&p=2
We've now seen stepping 6 on XS and they do not OC like stepping 5. The real production chip has only half the OC ability of stepping 5. These lower E6300 and E6400 chips may or may not be stable at 2.88ghz with stepping 6. But, my guess is that Anandtech will never bother to check.

A difference of opinions regarding B1 and B2 stepping. From Xbitlabs:

http://www.xbitlabs.com/articles/cpu/display/core2duo-shootout_10.html

Note that Core 2 Extreme X6800 that we had at our disposal features B1 core stepping. Since the mass production processors acquired B2 core stepping, we would expect the retail CPUs to have even higher overclocking potential. Nevertheless, the results we will obtain today will give us a great starting point for further analysis.

Anadtech posted a beautiful chart earlier where they purport to show that cache does not have much effect on bench speeds. http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2795&p=4
Unfortunately, this chart was created at 1.86Ghz where main memory is more easily able to keep up. Although they now have data at a more realistic 2.88 Ghz showing that their earlier chart is wrong they have not bothered to create a new one.

Here in this http://www.xbitlabs.com/articles/cpu/display/core2duo-e6300_2.html article we again see the claim that cache has no effect. Curiously, they too run the benches at a very low clock speed, 1.86Ghz so as not to put any real demands on memory bandwidth. This of course reduces the importance of the cache.

Those are two really nice charts from Anandtech and Xbitslabs. But there a problem. The only application that is the same on both tests is F.E.A.R. Xbitlabs' result is actually higher than Anandtech's; 5.3% vs. 1.5%. Oops!!!!

Also, while they both have tests for DivX, Anandtech uses DivX 6.1 while Xbitlabs uses DivX 6.2. While DivX 6.2 give better quality than DivX 6.1, it is also slower. The same can be said of DivX 5.1 and DviX 5.2; version 5.1 is faster, but version 5.2 arguably gives better quality.

Other than that all the other benchmark used are different. Thus rending this comparision of the two different result inert and useless.

People can easily find the earlier incorrect one but they will have to dig the the real numbers out of a later article. The top number, SysMark 2004 SE Overall is actually at least 8%, not 3.7%. I have to say, at least 8% because they are overclocking the FSB to reach 2.88Ghz and this contributes some to the speed. Without the overclocked FSB contribution the difference would be greater. But we don't know how much. In one test, a 2.88Ghz E6400 beats an X6800 clocked at 2.93Ghz in DivX encoding. That score shows a 5% contribution which would make the cache size a 13% advantage.

Okay where the hell is the link to Anandtech's updated article? I looked thru the archives, it's not there.

Other tricks are common such as using 4-4-4 timing for AMD because C2D doesn't benefit from the faster timing.

I believe that similar components and setting should be used for testing purposes. While there is the CORSAIR XMS2 CAS3 DDR2 800 RAM with timings of 3-4-3-9; I prefer not to spend $469 on 2MB of RAM.

If you want to read about unfair testing methods then please direct your attention to the following article from

Hardopc.com

The ONLY difference that we experienced is that we did have to lower a couple of settings with the AMD Athlon 64 FX-62 platform compared to the Intel platforms. This was the internal and external shadows. Luckily, the shadow sliders there are “notched” so it is easy to know exactly what position they are in. With the Intel CPUs, we were able to have this 5 notches up, which is in the middle of the slider for those shadow options. When we tried these same settings on the AMD Athlon 64 FX-62 platform, we found performance to be overall lower than the Intel CPUs and not playable. By moving those sliders down a couple of notches to 3 notches on the slider, performance was now playable.

The difference in graphics quality is minor though between notch 3 and 5. We noticed that some objects cast shadows farther in the distance as we approached them with the slider at notch 5. We also noticed that at extreme angles when viewing in third person view zoomed all the way out the character was casting a shadow in more places than with notch 3. That’s all we noticed between the slider positions.

Umm.... all they did was lower the game settings for the FX-62 so that the game will run. But they didn't change the game settings for Conroe.

No big deal right?

Other tricks include praising C2D's higher IPC and then making sure that the C2D chip is clocked higher.

The chart is from:

Intel Core 2 Duo E6300 & E6400: Tremendous Value Through Overclocking

Hmmm....

I can clearly see that the Conroe E6300 is clocked at stock speed of 1.86GHz and beats out the two Athlons 64 clocked at 2.0GHz and 2.2GHz.


- - - - - - --

Who is more stupid?

1. The original poster?

or

2. The person quoting the original poster?

 

[BaronMatrix]

The thing that I find most interesting about that graph is that for a 530MHz/$600 increase you get a lousy 7.25%.

I know I know it's something something limited.

 

[Action_Man]

Do you always have to be stupid?

 

[Action_Man]

Its called spin jack.

 

[JonathanDeane]

Its called spin jack.

Hehehehe I refer to Spud here Moo ? lol

 

[djkrypplephite]

Can't we ban his dumb ass yet? He posts the most inane things, not to mention he himself doesn't ever say anything. CONROE IS BETTER THAN ANYTHING AMD HAS, GET OVER IT. YOU LOSE. I swear to God I'm going to find "Tom" and make him personally IP ban this dumbass.

 

[BaronMatrix]

The thing that I find most interesting about that graph is that for a 530MHz/$600 increase you get a lousy 7.25%.

I know I know it's something something limited.

You get the same roll over curver (price/perf) with AMD:

http://www.xbitlabs.com/articles/cpu/display/core2duo-shootout_13.html

In fact, it is quite abit shallower slope at the high end compared to Intel....

Horrible example you present here.

I wasn't knocking Intel, I already knew that about AMD. I don't buy the highest priced processor and probably wouldn't even with Core 2.

All the time everyone just proves that now there is no REAL difference in processors. Well there won't be tomorrow.

He also clearly states that AM2 is only 20% slower while 70% faster than PD. 20% clock at 2.2 is 2.6GHz. That's striking distance for 65nm.

 

[BaronMatrix]

Its called spin jack.

Oh by the way with the new prices, AMD shows $700/800MHz offering 18%.

 

[Action_Man]

Look at the xbit page again.

I wasn't knocking Intel, I already knew that about AMD. I don't buy the highest priced processor and probably wouldn't even with Core 2.

:roll: Then why bring it up especially when AMDs is worse. Fanboy.

 

[mjp1618]

Great post.

Who is more stupid?

1. The original poster?

or

2. The person quoting the original poster?

Can we have a poll? :lol:

I choose option 2.

 

[lcandy]

That overclocking article from AnandTech was amazing, I can't believe an overclocked E6300 owns an FX-62, on stock cooling! Just imagine lumping a £30 aftermarket aircooler in there, like a Big Typhoon.

From a neutrals point of view this is a great website, that used to recomend A64 above all Intel had to offer. Now their oppinion couldn't be any different, the benchmarks do NOT lie. Although 9-inch thinks he knows more than the experts? What successful tech-website does he run again?

 

[1Tanker]

That overclocking article from AnandTech was amazing, I can't believe an overclocked E6300 owns an FX-62, on stock cooling! Just imagine lumping a £30 aftermarket aircooler in there, like a Big Typhoon.

From a neutrals point of view this is a great website, that used to recomend A64 above all Intel had to offer. Now their oppinion couldn't be any different, the benchmarks do NOT lie. Although 9-inch thinks he knows more than the experts? What successful tech-website does he run again?

It's dumbassesRus.com.

 

[mad_fitzy]

But Super Pi doesn't matter though. Have you seen all the gaming benchmarks WITHOUT the C2D being clocked higher? It wins. Period

 

[Cabletwitch]

Okaaaay.... 9-Inch, could you be so kind as to get yourself a C2D and an FX-62, and benchmark them yourself? That way, we could see the true results, as you wouldnt EVER hinder either processor to get the results to go your way, would you?

I'm looking forward to the conclusion you arrive at on this one. I guess whoever benchmarks these chips will evidently find that one is going to be better than the other, and in this case, its the Intel one for a change.

So go on, get yourself a pair of rigs, build em yourself, make them as closely setup as you can, then bench them.

Even as the AMD affictionado that I am, its clear to see that this round goes to Chipzilla. More incentive now for AMD to pull out all the stops, and create the next winning design.

 

[Shinobi13]

I have been coming to THG forumz and reading up on the latest and greatest for a while. Had to join and reply to this thread.

Even though I am a AVID AMD fan (never owned an intel product in 12yrs), I have to say that the new C2D has even me wanting to finally retire my trusty 3200+ Athlon XP 6800GS AGP rig (still plays BF2 beautifuly at high settings) and get me an e6700 proc. I still shutter at the thought!

Whats funny is that even though I have preached AMD to all my friends (built countless AMD based rigs for them), I am going to change my tune and recommend C2D. Just when the price comes down of course.

9-inch, face the facts brother....Intel won this round and AMD will have it`s day again.

So F`in what if it is Intel. It will still play your games and encode your downloaded pron. Grow up.

 

[jkflipflop98]

Ever heard of PCmark? It DOES do 2 different tasks at the same time. Stressing both cores, C2D still wins by 30% or so.


How in the bloody hell can you stand looking at a mountain of scientific data from private and professional parties - both have determined that C2D is just plain faster - and still try to dig this crap up?

Anandtech has a new AM2/C2D shootout up. In it, they have both platforms running at the same clocks, same bus, same ram timing, same HD's, same EVERYTHING, and C2D still utterly destroys the best the green team can offer.

We can now say with authority that Core 2 Duo is the faster performer clock-for-clock across the board. At the same 2.93GHz Far Cry is 27.7% faster, Half-Life 2: Lost Coast is 12.4% faster, and Quake 4 is 22.2% faster on Core 2 Duo. Of course AMD does not currently have a 2.93GHz CPU, so we tested by overclocking FX62. This suggests that FX64, or whatever it will be called, will not help much at 3.0GHz with a 200 clock speed.

Linky


The only advantage AM2 has is IMC. And well. . . my friend Penryn says thats nothing special