AMD Throwing the Gauntlet at Intel for releasing biased & unreliable benches.

[Tsukiyomi91]

I too used AMD before but moved to Intel for it's efficiency, despite being pricier than what AMD has to offer. With AMD like this, it's rather painful for me to see AMD stooped this low for a "fight" when they've been lying to their consumers ever since the introduction of the "new" FX processors & R9 graphics series to the masses when they're just 5-6 years old chip with very minimal improvements marketed as brand new. The reason why they lost the interest of consumers & reviewers is not because of their products getting kicked around by other vendors, but it's their own arrogance that kills them for believing that an old dog can still learn new tricks.

 

[cdawall]

Obviously it is since that is exactly what Sysmark measures. It measure raw cpu power, period. PCMark gives a better idea of real-world use, but if you are just comparing rare compute power, the i5 is way ahead.

But Intel shouldn't really be penalized because they created a more powerful processor with a lot more number crunching power just because that extra power isn't likely to show itself in basic computer tasks because those tasks don't use the extra CPU power.

If AMD wants to say Intel and Sysmark is biased then they have to show raw CPU compute based benchmarks that show Sysmark is way off on CPU power. Not use PCMark, which they admit doesn't really use the CPU that much. Do some CPU video encoding benchmarks, some compression benchmarks, some music encoding benchmarks. Those types of benchmarks rely on the CPU power, and will use every bit of CPU power available, and show the real difference in compute power between the two CPUs.

This "look Microsoft Office isn't any faster on Intel, so Sysmark is biased" shit isn't going to fly.

Would you say super pi is biased towards intel? Sysmark tests quite a bit of single threaded performance.

All this being said I looked at an 8800P and promptly purchased a 5500u.

 

[Tsukiyomi91]

@newtekie1 I thought on certain benches, especially on heavy video editing that AMD triumph over Intel & benchers were right but AMD doesn't listen to them? Sure, their chips really do show some processing muscle when it comes to multi-core/threaded tasks where Intel lags behind. But when it comes to computational power, Intel gains back their ground in that league because of one thing: it's efficiency per core it does is exceptional unlike AMD, where they rely heavily on raw speed.

 

[cdawall]

@newtekie1 I thought on certain benches, especially on heavy video editing that AMD triumph over Intel & benchers were right but AMD doesn't listen to them? Sure, their chips really do show some processing muscle when it comes to multi-core/threaded tasks where Intel lags behind. But when it comes to computational power, Intel gains back their ground in that league because of one thing: it's efficiency per core it does is exceptional unlike AMD, where they rely heavily on raw speed.

6th generation intel completely changes that. Multithreading finally caught up to AMD, but for the rest you are more or less correct.

 

[Tsukiyomi91]

If they (AMD) want to save their face, I hope they don't exaggerate on claims that their upcoming Zen & Polaris architecture are performing "worlds apart" while comparing on older chips from Nvidia & Intel while using botched benches to gain popularity & hype in their favor. But, since it's them after all... they'll gonna go on full force with it & shit starts hitting the fan all over again. If that happens, I won't be seeing AMD in any of my new builds.

 

[newtekie1]

Would you say super pi is biased towards intel? Sysmark tests quite a bit of single threaded performance.

No I wouldn't. But I would say SupoerPi hasn't been a relevant benchmark in years.

Sysmark tests things like media creation and number crunching, two very CPU intensive tasks, so it is going to show a bigger advantage to the CPU with more compute power. And, IMO, if you are going to be testing the CPU power, you want to use something that actually uses the CPU as much as possible. PCMark uses relatively low CPU intensive tasks. AMD even neglected to pick the PCMark8 Creative test, and instead went with the Work test, which only tests basic office tasks(word/excel, browsing, and video playback). I mean, shit, a Celeron handles those tasks with ease...

@newtekie1 I thought on certain benches, especially on heavy video editing that AMD triumph over Intel & benchers were right but AMD doesn't listen to them? Sure, their chips really do show some processing muscle when it comes to multi-core/threaded tasks where Intel lags behind. But when it comes to computational power, Intel gains back their ground in that league because of one thing: it's efficiency per core it does is exceptional unlike AMD, where they rely heavily on raw speed.

Yes, that certainly can be true for certain workloads with the desktop processors. But, we are looking at laptop CPUs here. Which for the FX means they had to really dial the clock speeds back, running the 8800p at only 2.1GHz, and it is only a quad-core with no L3 cache. It is a pretty crippled bulldozer based processor.

 

[Bones]

They'll gonna go on full force with it & shit starts hitting the fan all over again.....

I wonder what the budget for each is in buying more fans to either sling it or deflect it somehow?

 

[cdawall]

No I wouldn't. But I would say SupoerPi hasn't been a relevant benchmark in years.

Sysmark tests things like media creation and number crunching, two very CPU intensive tasks, so it is going to show a bigger advantage to the CPU with more compute power. And, IMO, if you are going to be testing the CPU power, you want to use something that actually uses the CPU as much as possible. PCMark uses relatively low CPU intensive tasks. AMD even neglected to pick the PCMark8 Creative test, and instead went with the Work test, which only tests basic office tasks(word/excel, browsing, and video playback). I mean, shit, a Celeron handles those tasks with ease...

Media creation is something the AMD chips excel at...That should say something about the testing.

 

[qubit]

Oh this should be good! I'll read it when I get home and break out the popcorn.

 

[cdawall]

Oh this should be good! I'll read it when I get home and break out the popcorn.

It isn't even worth the popcorn, just click bait and another shitty marketing ploy by AMD. Why haven't they fired the entire marketing department again?

 

[xfia]

It isn't even worth the popcorn, just click bait and another shitty marketing ploy by AMD. Why haven't they fired the entire marketing department again?

I just asked Galaxy but she didnt know If only I had Watson or Jarvis for the hard ones.

 

[Bill_Bright]

If they (AMD) want to save their face

They don't have anything to save face over! I say again, they make great processors that can be expected to last just as long as other brands. And when used with other properly chosen components, can form the foundation for a great computer - whether a game machine, CAD/CAE, server, POS, office, or plain old home computer.

This is just a marketing/PR assault from Intel, and sensationalize reports of unrealistic (not real-world) benchmarking tests.

To just make an absolute statement you won't be using any AMD products in any of your builds is simply falling for the marketing hype! But not only that, limits your own choices.

When you fall for the marketing hype and restrict yourself to a "sole-source", you feed the monopoly. And monopolies are NEVER good for consumers.

IF Intel, AMD, NVIDA only made one CPU and one GPU each, then and only then would it make sense to boycott one or the other. But because each produce dozens of processors across a wide range of performance capabilities - with all being very reliable, quality products - it makes no sense to not even consider all brands when building or buying a new computer. Why limit your options?

We need AMD! We don't need them to be in head-to-competition with Intel. But we do need AMD to keep nipping at the heels of Intel forcing Intel to keep looking over their shoulder to ensure AMD does not leap-frog over them again! AMD did that once before, totally spanking and embarrassing Intel. It took Intel nearly 8 years to gain the lead again with the Core 2 Duos and Intel cannot afford to let that happen again.

We need Intel to keep advancing the cutting edge but they will lose the incentive if no one is sneaking up behind them. If AMD dies because of bad press and biased comments from IT enthusiasts like us, that will be very bad for all of us.

 

[newtekie1]

Media creation is something the AMD chips excel at...That should say something about the testing.

Not in this case. They've crippled the FX-8800P so much. The 8-Core desktop FX chips are good at media creation, but even still have lagged behind the 4c4t dekstop i7s starting with the Haswell chips.

Now we are comparing a 4-Core 2.1GHz AMD, with no L3, to a 2c4t Intel Broadwell. The AMD chip is too crippled, they've killed the media creation ability.

Again, if they wanted to show a Bias they would have picked appropriate benchmarks. Not picked a benchmark that uses very little CPU power. There are plenty of benchmarks out there that give a good idea of actual CPU power they could have picked, PCMark8 Work is not one of them.

 

[cdawall]

Not in this case. They've crippled the FX-8800P so much. The 8-Core desktop FX chips are good at media creation, but even still have lagged behind the 4c4t dekstop i7s starting with the Haswell chips.

Now we are comparing a 4-Core 2.1GHz AMD, with no L3, to a 2c4t Intel Broadwell. The AMD chip is too crippled, they've killed the media creation ability.

Again, if they wanted to show a Bias they would have picked appropriate benchmarks. Not picked a benchmark that uses very little CPU power. There are plenty of benchmarks out there that give a good idea of actual CPU power they could have picked, PCMark8 Work is not one of them.

Fair enough I guess I didn't realize just how crippled the 8800P was.

 

[Bill_Bright]

There are plenty of benchmarks out there that give a good idea of actual CPU power they could have picked, PCMark8 Work is not one of them.

I think the point they were making is that IF Intel used PCMark8, it would have shown that Intel was not the run-away processor Intel was making it out to be.

 

[newtekie1]

I think the point they were making is that IF Intel used PCMark8, it would have shown that Intel was not the run-away processor Intel was making it out to be.

Sure, but then the argument could be made that PCMark8 Work is just as biased towards AMD since it doesn't actually test the CPU power. It uses low CPU power tasks.

Basically the argument AMD is making is that Intel shouldn't be saying their processor are more powerful because AMD processor are just as good at Word, Excel, and Internet Browsing. Yeah, sure they are, but any modern processor handles those tasks with ease, you aren't going to see really any difference between processors in those work loads.

Fair enough I guess I didn't realize just how crippled the 8800P was.

Just to give an idea of how crippled it is. I use multiple machines for encoding video when I have an whole TV series to encode. Delegating a season to each computer. My A10-6800K is similar to the 8800P in configuration, but still actually more powerful and clocked way higher. My A10-6800K, even at 4.6GHz, is just barely faster at encoding to H264 than my laptop i3-3217u which is an IvyBridge at only 1.8GHz...

When you take the huge L3 away from bulldozer it's performance starts to tank.

 

[cdawall]

If AMD wanted to bitch they would cry about 3dmark being biased. That is a test the 8800P should win, but doesn't.

 

[Bill_Bright]

Basically the argument AMD is making is that Intel shouldn't be saying their processor are more powerful because AMD processor are just as good at Word, Excel, and Internet Browsing. Yeah, sure they are, but any modern processor handles those tasks with ease, you aren't going to see really any difference between processors in those work loads.

Exactly my point as well and way we as advisors, should not disparage AMD just because we prefer Intel. And I note many AMDs work great in gaming rigs too. You don't have to have the most powerful rig to have the same amount of fun.

 

[BiggieShady]

What I hate more is how Intel compiler treats AMD CPU's even if they support all the required extensions. If AMD CPU runs Intel compiled binaries, it'll automatically be gimped.

That whole fiasco made way for third party math library that is optimized for all architectures optimally and based on supported instruction set rather than looking at cpuid ... The library is called Yeppp ... http://www.yeppp.info/benchmarks.html ... performance is just as impressive as library is underused.
Here are some comparative AMD vs INTEL numbers: http://www.yeppp.info/home/yeppp-performance-numbers/

 

[newtekie1]

Exactly my point as well and way we as advisors, should not disparage AMD just because we prefer Intel. And I note many AMDs work great in gaming rigs too. You don't have to have the most powerful rig to have the same amount of fun.

I completely agree. I can't count the number of Athlon 5350 machines I've built for clients just because all they do is Word, Excel, and Internet. They don't need anything more, and a 5350 machine is dirt cheap.

 

[Super XP]

This is old news, of course. Intel always played its tricks on Benchmarks. Especially Synthetics. If Intel needs to actually cheat on its Benchmarks, that's telling a completely different story. Quite pathetic actually, on there part.

 

[terroralpha]

talk about the pot calling the kettle black... how about those bulldozer and fury x benches??? I held on buying a 980 Ti for like a month and a half for no reason! i even ended up getting one but returned it after a week after realizing how worthless it was.

the AMD cheerleaders blamed immature drivers. here we are like 6 months later and the newest drivers only added a few FPS boost on average.

 

[Bill_Bright]

This is old news, of course. Intel always played its tricks on Benchmarks. Especially Synthetics. If Intel needs to actually cheat on its Benchmarks, that's telling a completely different story. Quite pathetic actually, on there part.

Now this is BS! You think Intel has the industry cornered on hype and marketing fluff? This plays exactly into what I was saying above where we as enthusiasts need to keep our biases in check to avoid perpetrating falsehoods as you just did.

No where is there any evidence of "cheating". They never "falsified" any test results - nor did they go "Volkswagen" with their CPUs sent to testing labs.

talk about the pot calling the kettle black... how about those bulldozer and fury x benches???

Exactly! Pick a product, any product and their marketing departments are going to "spin" the data to make their product look best. That's their job! And how well they do it makes the difference between consumers falling for the hype hook line and sinker, or take it with a grain of salt as they should.

 

[newtekie1]

Now this is BS! You think Intel has the industry cornered on hype and marketing fluff? This plays exactly into what I was saying above where we as enthusiasts need to keep our biases in check to avoid perpetrating falsehoods as you just did.

Yep, AMD is pretty much the king right now in marketing BS.

Heck, the video they made and this entire campaign is a perfect example. They're trying to say their processors are just as good as Intel's by running Excel and Word benchmarks... Talk about cherry picking benchmarks.

 

[Bill_Bright]

Yep, AMD is pretty much the king right now in marketing BS.

Heck, the video they made and this entire campaign is a perfect example. They're trying to say their processors are just as good as Intel's by running Excel and Word benchmarks... Talk about cherry picking benchmarks.

And I really don't have a problem with this. Why not tout your strengths? If you provide a product or service, you can't grow your business, let alone stay in business if you don't advertise.