AMD's smoothness factor put to the test by AMD & HardOCP...

[cold2010]

Marketing bad again

http://www.hardocp.com/article/2011/...rmance_review/

[jimbo75]

You can't argue with the results. Fact is, the intel cpu failed to gain even 30% of the total vote. Now had it been 70% of the vote people would have said "yeah ok that's fair enough", but 30% is way lower than what would be expected.

There is obviously something there making people believe the AMD system was performing better, for them. More than 2/3rds of the players didn't believe the AMD system was any worse - that's worth something regardless of how bad BD benchmarks.

[zanzabar]

there are 2 things that people on XS seam to forget every time, 1) the intel parts generally suck at stock clocks since the ram and IO are overall limited and the intel turbo dose not kick in with 4 threaded games like BF3, and 2) the point is not that amd is better, we can all look at benchmarks the point is that people did not tell the difference.

then for the IGP that was a clean win for amd since intel has the worst IGPs of anything ever made, and since that is were most computers are sold that is what should matter to most buyers, and remember that u have to buy the K edition intel to oc but the amd u dont so even here were people are almost always intel all the way if dont have the $300 to spend amd will be better every time by the time u can oc it anyways.

[highoctane]

I buy my parts for tangible performance/benefit in the real world for the tasks I need rather than trying to relying on my not so sharply tuned esp.

I know car analogies are bad but if you where to consider buying two similar cars with different engines, one rated for 20mpg the other rated for 30mpg and you where told to choose which car got the better mileage by test driving it for ten minutes.

[ice_chill]

You can't argue with the results. Fact is, the intel cpu failed to gain even 30% of the total vote. Now had it been 70% of the vote people would have said "yeah ok that's fair enough", but 30% is way lower than what would be expected.

There is obviously something there making people believe the AMD system was performing better, for them. More than 2/3rds of the players didn't believe the AMD system was any worse - that's worth something regardless of how bad BD benchmarks.

I doubt AMD's marketing team would post results of anything where the AMD FX CPU was clearly worse than Intel. Everything they showed in the last few months was hand picked to work in their favour.

[ice_chill]

there are 2 things that people on XS seam to forget every time, 1) the intel parts generally suck at stock clocks since the ram and IO are overall limited and the intel turbo dose not kick in with 4 threaded games like BF3, and 2)

What has stock clocks got to do with RAM and IO ? turbo mode does not improve RAM and IO speed, it only raises the multiplier of the CPU.

And Intel sucked in memory throughput in up to Core2 CPUs when memory controller was on the motherboard, since moving it on the CPU in socket 1156 and 1155 platforms Intel actually has higher memory throughput.

[zanzabar]

I doubt AMD's marketing team would post results of anything where the AMD FX CPU was clearly worse than Intel. Everything they showed in the last few months was hand picked to work in their favour.

i agree amd has been scummy but the windows patches to make the cpu read as a 4core 4 thread (like it really is) have made the chips play games alot better and while amd is not as quick in benches they are normally cheaper so once amd gets on the ball and puts out a chip only or cheaper heatsink at a $200 price point for the 8core it will be a great platform, and when u cannot oc like in an OEM product i still like amd better than intel.

What has stock clocks got to do with RAM and IO ? turbo mode does not improve RAM and IO speed, it only raises the multiplier of the CPU.

And Intel sucked in memory throughput in up to Core2 CPUs when memory controller was on the motherboard, since moving it on the CPU in socket 1156 and 1155 platforms Intel actually has higher memory throughput.

intels stock qpi speed leads to bad memory and IO performance if u stress both (so use the ram and pci-e) and intel needs turbo but they dont work stock with turbo when all cores are stressed, so by the nature of the chips they are the same in gaming. now throw up some wprime on 8 threads and amd wont fair well but for games they are nice right now, just a little costly.

and u can edit instead of double post.

[jimbo75]

I doubt AMD's marketing team would post results of anything where the AMD FX CPU was clearly worse than Intel. Everything they showed in the last few months was hand picked to work in their favour.

I doubt Intel or Nvidia would post results that weren't in their favour too. Let's think straight, no company is gonna show themselves in a bad light.

It's a blind test of BF3 and AMD is coming out ahead. BF3 is one of the biggest titles around - do you think intel or Nvidia would be wrong to showcase winning results in this had the been winning?

This is what is called *good* marketing. Blind tests like these count in the minds of people because it's as close to reality and impartiality as it gets. What made more than half of the people believe AMD was better?

[AliG]

AMD has been pulling this card since the introduction of Phenom. As a guy who used to formally exclusively use AMD products and now has a 2500k, I can definitely say the 2500k is equally as "smooth" and definitely much faster across the board

[SMTB1963]

It was Anandtech.
The test was not about smoothness at all, it was just
something they noticed.

AMD's test in Dallas didn't seem to be about smoothness either. I can't find the word "smooth" (more properly, "smoothe") in the TPU article, the Legit Reviews article, or the [H]ard|OCP article.

AFAICT, the only thing asked of the testers was "what system gave a better gaming experience":



While the i3 vs A8 was kind of a silly comparison, I think the 2700K vs 8150 might've showed a little bravery on AMD's part. The results of that comparison could have very well gone against them, given the unscientific/uncontrolled nature of the test. Of course, one man's bravery is another man's foolishness.

You can't argue with the results.

Sure you can! "Taste test" experiments in completely uncontrolled environments can be counted on to have a lot of statistical noise in the resulting data. It's one of the reasons that they're used so often in advertising.

[naokaji]

You can't argue with the results.

The main reason to doubt the results can be found in the thread title,

AMD's smoothness factor put to the test by AMD

Thats like the most biased source possible.

[CrazyNutz]

This means nothing since the source of the test is AMD, there are way too many things they can do to make the results favor AMD. Using coercive, and/or persuasion techniques on the visitors, and tweaks on the hardware it would not take much to sway the results in one direction.

If this test was done by a completely un-biased source then it would matter.

[gosh]

The CPU that is most balanced between cores should be the smoothest if work depends on the CPU and memory.

[ice_chill]

The CPU that is most balanced between cores should be the smoothest if work depends on the CPU and memory.

What does "most balanced between cores" mean please ? I have never heard that expression.

[thematrixhazune]

We need a reviewer to do this same test. Even though it might bring the fanatics out of the wood work, I myself would be interested in the results...As long as they are done without bias and without the help/aid of the company tested. I have never noticed this "smoothness" from my past intel/amd systems. The only thing I ever noticed was going from HD to SSD or lower to upper end processor.

[dctokyo]

Noob question here, would a AMD video card run better on a AMD platform than Intel? How about a AMD cpu + Nvidia card and a Intel cpu + Nvidia card test to see which one runs smoothest?

[STaRGaZeR]

What does "most balanced between cores" mean please ? I have never heard that expression.

You don't want to ask that to this guy. Trust me

[Final8ty]

http://www.anandtech.com/show/2715

Originally Posted by anand
After playing through the several levels on each platform, we thought the Phenom II 940 offered a better overall gaming experience in this title than the Intel Q9550 based on smoother game play. It is difficult to quantify without a video capture, but player movement and weapon control just seemed to be more precise. Of course, if you have the funds, we would recommend the i7 platform for best possible performance.

[gallag]

I don't trust amd anymore, all to recent they have been caught flat out bsing, if they lie on promotion videos about the specs of the cpus used in comparisons i am sure leaving out a chipset driver or some sort of driver/software manipulation or even just lying about what was in the case ( did they have an open case?). this is just against all reasonable scientific data and the whole smoother argument lost any clout when the advantage of the imc was lost.

[gosh]

What does "most balanced between cores" mean please ? I have never heard that expression.

If you create a multithreaded application, a work that is done in let say four threads. It will not be ready until the last thread is done and all threads are synchronized.
If the cpu has cores that isn't equally fast or can't handle the same amount of work depending on what work it should do. Then there could be some misses and the thread that need to do most work is placed on the slowest core. If this is done one time out of ten there will be some places where the CPU just then need some extra time to finish.
One of the most difficult task doing multithreaded applications is to schedule workloads so that threads in that task finishes at the same time. If you don't then you don't take advantage of the full potential what the CPU has.

[nn_step]

once again proving that engineering goals tend to be meet, while features outside of said goals tend not to be.

aka
AMD put forth effort to improve context switch latency and graphics
Intel put forth effort to improve power consumption and GFLOPs
and they both achieved their respective goals quite well.

The difference is that people don't know enough about what they are doing on a computer to make an informed decision.
[Intel really needs to get their ass in line on the Coreboot issue and AMD needs to tweak their bulldozer cores more.]

[3lfk1ng]

If AMD had posted some legit benchmarks of each rig to show that they are indeed "comparable", the test would be more believable. Also, I would kindly request internal pictures of the computers and a complete spec sheet.

I look at it this way:
If the AMD rig scored ~6k in 3DM11 and the Intel rig scored ~6K in 3DM11 - I would still need more results to make sure the rigs aren't being hand picked for other components.
I would want to see memory benchmarked. Maybe one has better ram or tighter timings.
I would want to see the HDD benchmarked. Maybe one of them is using an SSD or a 5400rpm HDD.
I would want to see if Windows "High Performance" is enabled on both machines.
I would want to see what drivers are being used for both the videocards as well as the settings. Not just in game settings, but video control panel settings where certain attributes can be forced.
I would want to see *sigh*... you get the picture.

There are just far too many variables that need to be listed in order for this "Smoothness factor" to be properly gauged.

[BababooeyHTJ]

http://www.anandtech.com/show/2715

Originally Posted by anand
After playing through the several levels on each platform, we thought the Phenom II 940 offered a better overall gaming experience in this title than the Intel Q9550 based on smoother game play. It is difficult to quantify without a video capture, but player movement and weapon control just seemed to be more precise. Of course, if you have the funds, we would recommend the i7 platform for best possible performance.

You know, I always noticed a few quirks here and there while I was on 775. I wouldn't be surprised if that were true.

I'm not at all buying that bulldozer offers a smoother experience than sandy though. Not at all and I've yet to make one person make that claim. I saw one guy on Rage3d claim that this framerates in Cryostasis doubled when moving from a Phenom 2 to 2600k. We've all seen the Skyrim benchmarks and the Arma2 benchmarks. In a lot of cpu limited games Bulldozer really lags behind. On the other hand Sandybridge really has been a smooth, issue free ride for me. I actually felt the same way with Lynnfield.

[Dainas]

Wouldnt higher FPS "tear" much more due to the 60Hz monitor refresh? I wonder if Intel's is less "smooth" because the FPS are too high...

Marketers are scum and test groups are always comprised of the dumbest trogs they can scrape up. Unless the idiot testers mistake hz tearing for hitching or AMD sabotaged their drivers to run choppier on an Intel system there is no truth to this whole smoothness bull, just like there was with AMD/ATI morons five years ago who pulled the 'smoothness' argument when AMD and ATI were in the darks of the HD2900 and Phenom 1 era.

[paulbagz]

Well Bob Lee Swagger said it himself;

Slow is smooth, smooth is fast.

-PB