Support NeoGAF

[Unknown Soldier]

Other than the fact that I'm very sad that Steve stopped welcoming us back to Hardware Unboxed awhile back, this video shows in rather hilarious fashion how benchmark engineering works at AMD and also why the Ryzen Non-AI 5800XT and 5900XT didn't need fake benchmarks. AMD could have just released the things and people would be like sure fine that's nice get rid of old dies you have no other use for headpats all around but NO. That's not how AMD rolls

Leonidas GET IN HERE ASAP

 

[Dacvak]

I think this thread was already posted and locked earlier.

 

[Leonidas]

AMD benchmarks are completely meaningless at this point.

7950 XTX.
5000 XT CPUs.
and soon 9950X.

All with contrived first party benchmarks in an attempt to mislead consumers.

 

[Leonidas]

Funny how this thread barely has any post, if any other company made such dishonest and contrived claims they'd be raked over the coals on this forum...

 

[Chiggs]

I think this thread was already posted and locked earlier.

Indeed it was.

Funny how this thread barely has any post, if any other company made such dishonest and contrived claims they'd be raked over the coals on this forum...

Not true. Nobody made a thread over this disaster:

Intel accused of inflating CPU benchmark results

SPEC says Intel's Xeon processors were using a compiler that artificially inflated the results of its industrial benchmark by as much as 9%.

www.pcworld.com

 

[Leonidas]

Not true. Nobody made a thread over this disaster:

Intel accused of inflating CPU benchmark results

SPEC says Intel's Xeon processors were using a compiler that artificially inflated the results of its industrial benchmark by as much as 9%.

www.pcworld.com

This is a gaming forum, so of course no one made such a thread of an industrial benchmark.

 

[Chiggs]

This is a gaming forum, so of course no one made such a thread of an industrial benchmark.

Oh, you mean servers, the one area where AMD has TRULY been annihilating Intel, making the cheating that much more egregious?

 

[Skifi28]

Funny how this thread barely has any post

I'm not worried. I'm sure over half will be yours.

 

[Leonidas]

Oh, you mean servers, the one area where AMD has TRULY been annihilating Intel, making the cheating that much more egregious?

Your link says they used compilers that increased perf. by 9%, way less agregious than what AMD is doing (which misleads dumb consumers into thinking the CPUs could be 30% faster in gaming than the reality), and again, this is a gaming forum. AMD is lying about gaming numbers, which is what this thread is about, not some server benchmarks that your link is talking about.

 

[Chiggs]

Your link says they used compilers that increased perf. by 9%, way less agregious than what AMD is doing (which misleads dumb consumers into thinking the CPUs could be 30% faster in gaming than the reality), and again, this is a gaming forum. AMD is lying about gaming numbers, which is what this thread is about, not some server benchmarks that your link is talking about.

So your position is: AMD is lying about benchmarks; Intel would never do a thing like that.

Is that correct?

If so, please type: "That is correct." That way we can move this discussion forward and I can start laughing now, rather than later.

 

[Leonidas]

So your position is: AMD is lying about benchmarks; Intel would never do a thing like that.

Nope. My position is this is a fucking gaming forum. AMD lied about gaming numbers.

Intel increasing their shit by less than 10% is OT, it has nothing to do with gaming.

 

[SoloCamo]

AMD's marketing has always shot themselves in the foot. I just don't understand why they never learn... people that actually watch these events are not idiots.

 

[Chiggs]

Nope. My position is this is a fucking gaming forum. AMD lied about gaming numbers.

Intel increasing their shit by less than 10% is OT, it has nothing to do with gaming.

I've bookmarked your post for future reference and use. I don't think I'll be waiting that long to remind you of it.

 

[Zathalus]

Your link says they used compilers that increased perf. by 9%, way less agregious than what AMD is doing (which misleads dumb consumers into thinking the CPUs could be 30% faster in gaming than the reality), and again, this is a gaming forum. AMD is lying about gaming numbers, which is what this thread is about, not some server benchmarks that your link is talking about.

One corporation was actively cheating at benchmarks for years, while the other had some misleading marketing benchmarks. While both are shit things to do, the actively cheating one is obviously worse. Maybe less relevant to a gaming forum, but still worse in the grand scheme of things.

 

[Chiggs]

One corporation was actively cheating at benchmarks for years, while the other had some misleading marketing benchmarks. While both are shit things to do, the actively cheating one is obviously worse. Maybe less relevant to a gaming forum, but still worse in the grand scheme of things.

All of these companies lie and cheat and it's slimy as fuck, which makes Leonidas' stance here all the more curious and short-sighted.

It's going to be, what, maybe 4-6 months before Intel is caught dicking around....again? When does their new architecture launch? This fall?

 

[Leonidas]

I've bookmarked your post for future reference and use. I don't think I'll be waiting that long to remind you of it.

You'll have to wait longer than AMD being proven to have lied yet again when Zen 5 launches in 31 days.

 

[Chiggs]

You'll have to wait longer than AMD being proven to have lied yet again when Zen 5 launches in 31 days.

I really won't, because I'm not an AMD fanboy. They've just objectively been a better option for most people for the last 4 years. Sure, I made a fuckload of money off their stock, so maybe there's a bit of bias....but it's nothing like the bias I have for Arm right now.

Just admit it: you didn't like the Intel Silicon Degradation thread I made, so you wanted to lash out and saw your chance with this news. Then your thread got locked and you were butthurt.

 

[Leonidas]

I really won't, because I'm not an AMD fanboy.

Why'd you post you bookmarked this for when Intel lied about gaming then?

Just admit it: you didn't like the Intel Silicon Degradation thread I made, so you wanted to lash out and saw your chance with this news. Then your thread got locked and you were butthurt.

Didn't know you made one, I don't care about it since my CPU (13600K) is unaffected.

It is odd my thread was locked though, do you think that's good that my thread, the same as this one, was locked. Seems some people just can't handle when AMD lies to its consumers.

 

[SmokedMeat]

Don’t care enough to watch the video, but I’m wondering who buys a product based on the benchmarks of the manufacturer?

Am I the only one who doesn’t buy into them or feels they’re deliberately cherry picked - regardless of company?

 

[Chiggs]

Why'd you post you bookmarked this for when Intel lied about gaming then?

To show you just how pointless the whole volley is.

• Do you think Intel won't be caught fudging their numbers ever again?
  • Be it consumer or business?
• Do you think I actually think AMD will never bend the truth/outright fabricate or distort the facts to sell product?

So why defend their honor?

 

[winjer]

Nope. My position is this is a fucking gaming forum. AMD lied about gaming numbers.

Intel increasing their shit by less than 10% is OT, it has nothing to do with gaming.

It's true that AMD lied with the results of the 5800XT, by limiting performance with a 6600.
But Intel is doing the same, if not worse.

During the Lunar Lake presentation, they claimed a 14% IPC improvement.
The problem is that figure is based on Spec ICX, a benchmark made by Intel, that has been found to use inflated results.
And the rest are a bunch of synthetic benchmarks, all picked by Intel.
Not one single real world benchmark.

We might also remember when Intel used an overclocked CPU, using a chiller, to pretend their performance was much greater than what it really was in the real world.

Intel: We 'Forgot' to Mention 28-Core, 5-GHz CPU Demo Was Overclocked

We met with Intel and asked a few pointed questions about its now-infamous 28-core CPU demo.

www.tomshardware.com

Or maybe we should remember when Intel hired a third party company to make some "independent" benchmarks. And the result was a gigantic lie.


The reality is that these big companies tend to manipulate the truth, so that's why it's important to have real independent benchmarks.

 

[Leonidas]

Don’t care enough to watch the video, but I’m wondering who buys a product based on the benchmarks of the manufacturer?

Am I the only one who doesn’t buy into them or feels they’re deliberately cherry picked - regardless of company?

You'd be surprised how gullible some people are.

To show you just how pointless the whole volley is.

• Do you think Intel won't be caught fudging their numbers ever again?
  • Be it consumer or business?
• Do you think I actually think AMD will never bend the truth/outright fabricate or distort the facts to sell product?

So why defend their honor?

I'm not defending any company. I'm calling out those who lie in gaming benchmarks on a gaming forum. Something AMD has done at least 3 times in the past 3 years.

 

[RoboFu]

I did not watch this video or any other video they have made in a year or two but I will say ... unboxed has always been shady but they cater to a crowd who wants all their BS to be true.

 

[Leonidas]

Or maybe we should remember when Intel hired a third party company to make some "independent" benchmarks. And the result was a gigantic lie.


The reality is that these big companies tend to manipulate the truth, so that's why it's important to have real independent benchmarks.

Funny you had to go back 5 years to get something that showed gaming benchmarks, and failed to mention that the numbers were later corrected.

AMD lied about gaming 3x in the past 3 years.

 

[winjer]

Funny you had to go back 5 years to get something that showed gaming benchmarks, and failed to mention that the numbers were later corrected.

AMD lied about gaming 3x in the past 3 years.

One is from this month. And it's a big one from Lunar Lake.
And the lying with their compilers is from 2024.
And then there is the benchmarks you post, with slower memory or much higher latency timings on Ryzen CPUs.

BTW, when did AMD lie 3 times, in the past 3 years. Just curious, because you rarely post your sources.

 

[OmegaSupreme]

I did not watch this video or any other video they have made in a year or two but I will say ... unboxed has always been shady but they cater to a crowd who wants all their BS to be true.

Explain

 

[welshrat]

Oh No Anyway GIF - Oh No Anyway - Discover & Share GIFs

Click to view the GIF

tenor.com

 

[Leonidas]

One is from this month. And it's a big one from Lunar Lake.
And the lying with their compilers is from 2024.
And then there is the benchmarks you post, with slower memory or much higher latency timings on Ryzen CPUs.

I didn't see any Lunar Lake gaming results, its OT. Get back to me when you have gaming benchmarks.

BTW, when did AMD lie 3 times, in the past 3 years. Just curious, because you rarely post your sources.

We've been through this, your memory is great when it comes to some things, but it is abysmal when it comes to things we've already discussed.

'22 AMD stated 7600X was 5% faster than 12900K. A lie.
'24 This thread where AMD swinged results 30+% in their favor. A lie, rightfully called out by AMD Unboxed.
'24 9950X launches in 31 days, I will reveal the info you still haven't figured out, when reviews show ancient 10nm 14900K beating 9950X in one of the 6 games AMD claimed a win in. But of course, I'm just making all this shit up, even though I've seen enough evidence to be 100% sure of this conclusion.

Only someone with a much higher than average IQ, such as myself, can make that final claim. Average folk like you will have to wait for the reviews in 31 days, when what I said is proven to be true

 

[winjer]

I didn't see any Lunar Lake gaming results, its OT. Get back to me when you have gaming benchmarks.

We had no one real world benchmark for Intel's claim of 14%. Not from games, not from apps.
Yet you claimed this was the real number. While AMD's claim was false.

We've been through this, your memory is great when it comes to some things, but it is abysmal when it comes to things we've already discussed.

AMD stated 7600X was 5% faster than 12900K. A lie.
This thread where AMD swinged results 30+% in their favor. A lie, rightfully called out by AMD Unboxed.
9950X launches in 31 days, I will reveal the info you still haven't figured out, when reviews show ancient 10nm 14900K beating 7950X in one of the games AMD claimed a win in. But of course, I'm just making all this shit up, even though I've seen enough evidence to be 100% sure of this conclusion.

Only someone with a much higher than average IQ, such as myself, can make that final claim. Average folk like you will have to wait for the reviews in 31 days, when what I said is proven to be true

Yes we have. But those numbers you state are from a reddit aggregate, which includes several reviews that used slower memory on the Ryzen CPUs.
So of course that when you ad reviews that manipulate the numbers to favor Intel, those averages don't match with the numbers AMD published.
So only when you lie outright, does it seem that AMD had the wrong numbers.

 

[Leonidas]

Yes we have. But those numbers you state are from a reddit aggregate, which includes several reviews that used slower memory on the Ryzen CPUs.
So of course that when you ad reviews that manipulate the numbers to favor Intel, those averages don't match with the numbers AMD published.
So only when you lie outright, does it seem that AMD had the wrong numbers.

How will you spin AMD losing in actual reviews in one of the 6 games they contrived themselves into winning?

It won't be an outright lie in 31 days, it will 100% be true.

 

[RoboFu]

Explain

oh don't get me wrong I think all these YouTubers are BS. Clickbait titles .. chasing controversy.

 

[winjer]

How will you spin AMD losing in actual reviews in one of the 6 games they contrived themselves into winning?

It won't be an outright lie in 31 days, it will 100% be true.

Like I have already told you several times, we need independent, trustworthy benchmarks.
Unlike you, we all want to see the reviews. We want to compare Arrow lake vs Zen5 and make our own conclusions.
We all know, except you, that these companies massage the numbers.
And that is why we want those reviews.
So maybe Zen5 will win. Or maybe Arrow Lake will win. The fact today, is that we DO NOT KNOW.
So please, stop trying to pretend that Intel has already won. Because, once again, we do not know.

 

[Leonidas]

Like I have already told you several times, we need independent, trustworthy benchmarks.
Unlike you, we all want to see the reviews. We want to compare Arrow lake vs Zen5 and make our own conclusions.
We all know, except you, that these companies massage the numbers.
And that is why we want those reviews.
So maybe Zen5 will win. Or maybe Arrow Lake will win. The fact today, is that we DO NOT KNOW.
So please, stop trying to pretend that Intel has already won. Because, once again, we do not know.

Swinging results 20%+ in your favor in one of the cherry picked benchmarks is more than massaging the numbers, its a lie.

So maybe Zen5 will win. Or maybe Arrow Lake will win. The fact today, is that we DO NOT KNOW.

If either loses to 7800X3D (which was already barely faster than 13900K/14900K) they are a fucking joke.

So please, stop trying to pretend that Intel has already won. Because, once again, we do not know.

Where have I said Intel won? I said, they probably will win, and ever since AMD said Zen5 lost to Zen4 3D, I think the % chance that Intel wins is increased.

Shouldn't be that complicated to understand this reasoning. Of course we have to wait, but the way things look now, unless Arrow Lake is an abysmal uplift in gaming, it should win vs. Zen5 (given AMD said Zen4 3D is faster). Understand?!?

 

[winjer]

Swinging results 20%+ in your favor in one of the cherry picked benchmarks is more than massaging the numbers, its a lie.

If either loses to 7800X3D (which was already barely faster than 13900K/14900K) they are a fucking joke.

Where have I said Intel won? I said, they probably will win, and ever since AMD said Zen5 lost to Zen4 3D, I think the % chance that Intel wins is increased.

Shouldn't be that complicated to understand this reasoning. Of course we have to wait, but the way things look now, unless Arrow Lake is an abysmal uplift in gaming, it should win vs. Zen5 (given AMD said Zen4 3D is faster). Understand?!?

You have said so many times that Intel has already won. The only reason you come to this forum is to tell us that Intel is the best.
You even make up benchmarks and numbers so you can pretend Intel never lies. And wins all benchmarks.

 

[Leonidas]

You have said so many times that Intel has already won. The only reason you come to this forum is to tell us that Intel is the best.

After AMD themselves said it was slower than Zen4 3D, we can't even be sure Zen5 beats the 14900K.

You even make up benchmarks and numbers so you can pretend Intel never lies. And wins all benchmarks.

Ban bet me then. 14900K beats 9950X in one of AMDs cherry picked benchmarks.

You, and no other AMD fanboy on this forum has the balls to take me up on that.

 

[winjer]

After AMD themselves said it was slower than Zen4 3D, we can't even be sure Zen5 beats the 14900K.

Ban bet me then. 14900K beats 9950X in one of AMDs cherry picked benchmarks.

You, and no other AMD fanboy on this forum has the balls to take me up on that.

Here you go again....

 

[Leonidas]

Here you go again....

no balls, thats you.

He's definatey lying, but I won't take him up on that bet...