AMD Smoother than Intel?

[piesquared]

n=1 experiences aren't useful. Measurements are needed to confirm this. Those measurements are called benchmarks. Last time I checked, they showed that Intel's CPUs are fastest ("smoothest").

I'm pretty sure we've established that fastest has nothing to do with smoothness. Benchmarks are useless as they can be cheated provided enough money is thrown at them. See Futuremark with Physx. See BAPCO. See Cinebench. See ICC. See Agnor Fog. etc. etc. Personal experiences are all that count, and from first hand accounts of reviewers on Anandtech, for 1, AMD systems are smoother. Blind tests also showed this to be the case. My personal experience and those of people I have asked after using both systems, along with others in just this thread alone, also agree that AMD systems are more responsive. First hand account is the only true accurate test. There's really no way to deny that. The end users experience is the only criteria that matters.

 

[podspi]

Other than the rather specific example I already posted, wouldn't the evidence for this show up in DPC benchmarks, and thus been noticed and quantified a while ago?

I agree there are a few ways a CPU could be smoother than another (See the original Phenom and Cool and Quiet for example), but AFAIK (and have observed) Intel's power management is very, very good.

Almost 100% of the stutter I get nowadays is from when I'm hitting a hard drive that has spun down.

 

[piesquared]

There is stutter when the hard drive has spun down agreed. That's true on all systems though. This seems to be something different, although similar. Scrolling windows is where I and other people i've asked notice it the most. It's why i'm leaning towards it being a problem with intel's garbage graphics and drivers, although an issue with hyperthreading is also possible. Many applications require hyperthreading to be disabled as performance takes a big hit with it enabled.

 

[piesquared]

AMD was part of BAPCO until the Bulldozer launch, and they didn't resign because they were questioning the accuracy of the CPU performance measurement, but because they wanted the GPU to play a more significant role on the benchmarks score because, guess what, Bulldozer sucks big time and that would really kill whatever marketing purpose of Sysmark for AMD.

Have in mind that this happened AFTER the benchmark was specified and built, with AMD agreement until that point, so up until the Bulldozer launch AMD did indeed agree with how the measurements were made and how Sysmark 2012 was built.

Nigel Dessau also used to be John Fruehe's boss, and he was fired from AMD just after the Bulldozer launch.

I think your sharp eyes caught something that not even AMD internal tests did not: That Intel CPU causes microstuters and AMD doesn't. You should email AMD ASAP to make sure they build a benchmark to show this, much like Nvidia did with FRAPS and caught AMD GPU driver team with their pants down.

AMD weren't the only ones that bailed out of BAPCO . Do you really want to have a debate about what BAPCo is and where it came from?

 

[TeknoBug]

When SWTOR first came out a friend of mine had a i3 2130(?) with a gtx480 and I had a X6 1090T with a gtx560Ti and was getting frustrated in the games performance yet my friend experienced no issues. That's when I decided to build an Intel for gaming and haven't looked back so far

 

[videogames101]

DPC latency is measurable. Go test your systems before you make this thread. That way we don't have to sit here watching fanboys throw piles of shit at each other.

edit: LOL OP joined just to make this thread...

 

[shady28]

Other than the rather specific example I already posted, wouldn't the evidence for this show up in DPC benchmarks, and thus been noticed and quantified a while ago?

I agree there are a few ways a CPU could be smoother than another (See the original Phenom and Cool and Quiet for example), but AFAIK (and have observed) Intel's power management is very, very good.

Almost 100% of the stutter I get nowadays is from when I'm hitting a hard drive that has spun down.

No one has really addressed this. Given that PCMark is a "total system" benchmark, it seems more relevant than tests that stress the ability to do *uncommon* tasks like cinebench, handbrake, pov-ray, etc.

Those are the benchmarks we always see on "enthusiast" sites, yet I know maybe one person that uses those kinds of apps with any regularity, and even then it is not the bulk of their usage (a photographer, who does filming / photos for events).

Everyone else does normal things, like email, youtube, web browsing (flash, javascript, HTML5 rendering), opening office docs from email, iTunes, etc.

Those types of common tasks are represented here :

I don't think there is any real argument about an AMD "core", in and of itself, being faster at much of anything vs a full i-series Intel CPU.

But there is definitely a case to be made that the whole benchmark / enthusiast community has some kind of blinders on when it comes to what "normal" people do with their computers, and how to measure that.

 

[cytg111]

So if I say that I've found Intel to be 17 times faster than AMD at the same workloads at the same price point, you'll have to take my word for it and I wouldn't have to provide any sources to back up my claims?

Wow, this is neat! Let me see what I can come up with then.

Common dude, search your feelings, I am sure that you'll find smoothness and whatnot somewhere inthere ....

- Something women do often, but for the love of * Dont mix facts with 'feelings'. - its just bad.

 

[AnandThenMan]

The only people who say AMD systems feel smoother are those attempting to justify their poor purchase decisions.

I don't need to justify my purchases, I buy what I like. I happen to own an Ivy Bridge and FX system, I like the AMD system better for heavy workloads. It just feels faster and more responsive.

I don't care if others think different or demand benchmarks that prove otherwise. Own both systems before you judge.

 

[witeken]

I'm pretty sure we've established that fastest has nothing to do with smoothness. Benchmarks are useless as they can be cheated provided enough money is thrown at them. See Futuremark with Physx. See BAPCO. See Cinebench. See ICC. See Agnor Fog. etc. etc. Personal experiences are all that count, and from first hand accounts of reviewers on Anandtech, for 1, AMD systems are smoother. Blind tests also showed this to be the case. My personal experience and those of people I have asked after using both systems, along with others in just this thread alone, also agree that AMD systems are more responsive. First hand account is the only true accurate test. There's really no way to deny that. The end users experience is the only criteria that matters.

Personal experiences are all that count? Not really.

Please provide scientific evidence for this while smoothness thing, so we'll have some real facts to talk about.

 

[tential]

Citation needed for what? Our personal experiences? If you want the results of the blind tests, search for them yourself. There's enough people around that remember the results that I don't need to prove anything. As for reviewers opinions, search Anandtech. There's enough people that remember that also.

Intel systems are smoother. It's common knowledge. Reviewers regularly say it. If you want their opinions, search anandtech.

You see how ridiculous that is? You can't just claim that people say something and then not cite ONE SINGLE SOURCE. This is the behavior that makes these types of discussions a joke.

People come on here, make claims as facts and won't cite a single source.
So far, not a single person has cited a source where a reviewer says "AMD is smoother than intel." You need to cite your source if you want to quote what reviewers are saying. Otherwise, it's PERFECTLY fine for you to say that your personal experience is that AMD is smoother. But if you say that Anandtech says that, then provide a source/quote/citation. If you can't back up your claims with a source/quote/citation then don't make them.

 

[piesquared]

Personal experiences are all that count? Not really.

Please provide scientific evidence for this while smoothness thing, so we'll have some real facts to talk about.

Yes really. That's all that matters, and we don't need scientific proof for personal experience. Everything else is just noise. It's been proven time and again that 'scientific proof' can easily be skewed, manipulated and cheated as we've seen from BAPCo, Cinebench, ICC, etc. etc.

Intel systems are smoother. It's common knowledge. Reviewers regularly say it. If you want their opinions, search anandtech.

You see how ridiculous that is? You can't just claim that people say something and then not cite ONE SINGLE SOURCE. This is the behavior that makes these types of discussions a joke.

People come on here, make claims as facts and won't cite a single source.
So far, not a single person has cited a source where a reviewer says "AMD is smoother than intel." You need to cite your source if you want to quote what reviewers are saying. Otherwise, it's PERFECTLY fine for you to say that your personal experience is that AMD is smoother. But if you say that Anandtech says that, then provide a source/quote/citation. If you can't back up your claims with a source/quote/citation then don't make them.

There's more than a single source: Blind tests, Anandtech reviewer, personal experiences. You can search for those yourself, as there are plenty of people that remember those examples. I'd be surprised if you don't remember them as well.

 

[Tom80112]

DPC latency is measurable. Go test your systems before you make this thread. That way we don't have to sit here watching fanboys throw piles of shit at each other.

edit: LOL OP joined just to make this thread...

My Join Date: May 2009.
How can I fake that?
You really can't be this dumb can you?

 

[tential]

I actually don't remember them and if there seems to be a number of forum members who don't. So far in fact, you're the only person who seems to remember them.

There are so many sources yet you won't link one source yourself. I'm not here to validate your claims. If you make a claim you should back it up.

It's COMMON knowledge that AMD is smoother than intel yet we have a 7 page thread debating the fact. With 0 sources posted from every person who has claimed this is true. I didn't know this type of behavior was allowed on this forum but I'm done with this conversation. "Technical" discussions in which 0 people cite sources to back their claims is not a technical discussion at all.

 

[piesquared]

My Join Date: May 2009.
How can I fake that?
You really can't be this dumb can you?

We know he is thread crapping, but that will just get you banned and thread probably locked, regardless of how true or untrue it may be.

 

[piesquared]

I actually don't remember them and if there seems to be a number of forum members who don't. So far in fact, you're the only person who seems to remember them.

There are so many sources yet you won't link one source yourself. I'm not here to validate your claims. If you make a claim you should back it up.

It's COMMON knowledge that AMD is smoother than intel yet we have a 7 page thread debating the fact. With 0 sources posted from every person who has claimed this is true. I didn't know this type of behavior was allowed on this forum but I'm done with this conversation. "Technical" discussions in which 0 people cite sources to back their claims is not a technical discussion at all.

Cya. Plenty of people remember them, i'm not going to troll through search to find them but it was a pretty hot topic.

 

[Tom80112]

I'm glad I started this thread. It confirms what I have suspected all along.

After reading this thread, I also noticed that every poster whom voiced their
experiences of AMD being smoother in day-to-day operation was met with
an onslaught of ridicule by people who went out of their way to reply without
any class.

Hat's off to all those who responded to the onslaught of ridicule with "class".

 

[mrmt]

AMD weren't the only ones that bailed out of BAPCO . Do you really want to have a debate about what BAPCo is and where it came from?

You mean Nvidia or Via? Nvidia has the same interests of AMD in this area: shift much of performance evaluation from the CPU to the GPU, because they are in an even worse position than AMD (AMD has a couple of bad CPU cores, Nvidia doesn't have a CPU core), and VIA, pff...

 

[OCGuy]

Hey guys, it is only silicon. No need to start a war, right?

 

[witeken]

I'm glad I started this thread. It confirms what I have suspected all along.

I'm curious, how can it confirm your suspicion if there isn't 1 professional research about this linked in this whole thread?

Now read this thread again but cherry pick other comments, and you will come to the opposite conclusion.

 

[piesquared]

You mean Nvidia or Via? Nvidia has the same interests of AMD in this area: shift much of performance evaluation from the CPU to the GPU, because they are in an even worse position than AMD (AMD has a couple of bad CPU cores, Nvidia doesn't have a CPU core), and VIA, pff...

21.
intel solved its problem by making it appear as if the Pentium 4 outscored the Pentium III and AMD Athlon processors. It unleashed a multi-tiered strategy that used several different methods to inflate the Pentium 4’s performance scores—methods that Plaintiffs contend constitute unfair business practices that hinder legitimate competition and thus harm consumers.

22. A key element of the strategy was the introduction of two “new" benchmarking tests— called WebMark 2001 and SYSmark 2001—that made the Pentium 4 appear superior to the Pentium III and Athlon. Intel claimed publicly that the new tests were created by a neutral benchmarking company, and thus reflected an objective measure of processor performance; in reality, the tests were written in-house by Intel to favor the Pentium 4 and inflate its scores.

23. Other elements of Intel’s strategy included paying software companies to secretly rewrite software programs used in benchmarking tests so that the Pentium 4 would score higher; using higher-performance memory to inflate the Pentium 4’s performance scores when compared to other processors (which were tested with mainstream memory); and disabling certain features on the Pentium III so that its performance scores would drop and the Pentium 4 would appear better by comparison.


24. The first “new” benchmark program that Intel designed specifically to inflate the Pentium 4’s performance scores was called WebMark 2001.

25. To bypass the skepticism that Intel knew reviewers would harbor against a benchmark Intel had written itself, Intel passed the new benchmark off to the market as one developed by a seemingly independent, objective third-party called Business Applications Performance Corporation, or BAPCo. As Intel’s Performance Marketing Manager told Intel insiders, “It is very important that everybody understand the positioning that this is not an Intel developed benchmark…..rather that it was developed by BAPCo, who is an unbiased industry consortium of leading companies.”

http://www.scribd.com/doc/214203686/Pentium-4-Litigation

Just some of the issues with BAPCo. Now i'm guessing your first response will be something like: 'pff how long ago was that again?', but the important part is that intel still controls BAPCo, and I don't think too many people believe intel's corporate culture has changed much if any. No company is going to support such a biased, deceitful scheme. AMD, NV and Via had no choice but to stick with it because it was the standard, and corporates/businesses used it to evaluate their purchases. When they bailed, they had to bail together.

 

[piesquared]

I'm curious, how can it confirm your suspicion if there isn't 1 professional research about this linked in this whole thread?

Now read this thread again but cherry pick other comments, and you will come to the opposite conclusion.

What kind of professional research? You mean the likes of BAPCo? I'll take personal experience over bought and paid for benchmarks thanks. Reviewers like those on Anandtech and blind tests also confirm the AMD system to be more responsive and smoother.

 

[Sequences]

What kind of professional research? You mean the likes of BAPCo? I'll take personal experience over bought and paid for benchmarks thanks. Reviewers like those on Anandtech and blind tests also confirm the AMD system to be more responsive and smoother.

Do you think personal experience cannot be skewed or biased?

 

[OCGuy]

I just read a report that Intel CPUs are actually 72% more intriguing, 32% more comfortable, and 16% less needy than any other x86 chips on the market. Surely this balances out any smoothness, likeability, or clinginess advantage that AMD may hold.

 

[amenx]

First hand account is the only true accurate test. There's really no way to deny that.

Personally, i've never owned an intel system (other than a Tualatin way back)

:whiste: