PC Games Intel performance testing fake.

[Drollic]

It's funny how no-one was interested in 1080p frame rates with high end hardware until AMD became competitive. Up to that point it was all about trying to achieve 60fps+ at 4K. Now it's all about getting 135fps rather than lagging behind at only 125fps at 1080p.

The majority are running 1080p because they can't afford 4K hardware. 2080Ti performance with a 9900K at 1080p is about as irrelevant as it gets.

Not true. A lot of people run 1080p and 1440p high hz as they want the smoothest possible gameplay. You don't get that with 4k. It's particularly apparent with fps style games. I personally would find 40-60 fps at 4k unplayable. You've got people currently dropping settings just to hit 60 at 4k. Far better off maxing at 1440p higher frame rate for me.

 

[sandys]

PT seems completely reasonable, just a game mode issue really, that is all. The thing on memory is a bit silly, officially supported memory speeds on each platform is what was tested, yes there are XMP profiles etc, but they are clocked beyond officially supported RAM speed spec. They could have run a better cooler on AMD but it would not make enough difference to really matter.

Regarding 1080p, just LOL, I like big screens not gaming through a postcard , 1080p looks like going back in time to VGA gaming with super big pixels, less than 50dpi is a bit yukky on a big screen give me more res, more screen and many GPUs to push it into high frame rates thank you, there are technologies to make lower fps feel good if you can't make it, it is rarely a problem with the best GPUs.

 

[humbug]

Missing the point so hard your post reads like bait.

Anyone capable of reading a graph knows that Intel has faster cpus for gaming. They put their name on a grossly incompetent test that unfairly exaggerates this and are being mocked for it.

To claim "best gaming cpu" they didn't need to hold up lies but did anyway.

This, of course the 8700K is the best gaming CPU, the 9900K will push that a bit further.

That's not in dispute, the point is Intel commissioned and then slapped their name to a test whose architect was in every way it is possible to be grossly incompetent or deliberately set the test up to make the Intel CPU give the best possible results while the Ryzen CPU put out the worst possible results, including and not limited to actually disabling 4 of the 2700X cores, effectively turning it into a Ryzen 2400G.

 

[DragonQ]

PT seems completely reasonable, just a game mode issue really, that is all. The thing on memory is a bit silly, officially supported memory speeds on each platform is what was tested, yes there are XMP profiles etc, but they are clocked beyond officially supported RAM speed spec. They could have run a better cooler on AMD but it would not make enough difference to really matter.

It really doesn't seem reasonable. The fact that they chose dodgy settings everywhere is only a small part of the issue, the biggest problem is they used different settings and configurations for both machines when there was no need. This is supposed to be an apples-to-apples comparison but is far from it. The Intel machine has an aftermarket cooler and RAM set to an XMP profile. The fact that they were running at "officially supported RAM spec" is irrelevant because setting an XMP profile sorts out timings, not just speed. The Intel machine using 2666 MHz XMP will have a significant advantage over the AMD system using 2933 MHz without XMP. Both machines would definitely be capable of running 2933 MHz RAM with XMP enabled (with 2 DIMMs anyway) and even if they weren't, why not disable XMP on both? This is a huge red flag.

As you said, using "game mode" is a huge issue because it disables half the cores! One could argue that AMD has named it badly but the Ryzen Master app makes it clear it's designed for Threadripper chips only. How you can say "PT seems completely reasonable" and "just a game mode issue really, that is all" in the same sentence baffles me. If AMD commissioned a comparison against an Intel chip with half its cores disabled without making it incredibly clear, would you really have the same opinion?

Either PT are completely incompetent or very disingenuous. It could easily be the first considering they used 4 DIMMs in every test for a total of 64 GB of RAM, a configuration almost no gamer would use (not just because it's unnecessary but because it lowers your maximum achievable speeds/timings). Either way, the fact that Intel endorses and trots out these results whilst preventing competent reviewers from fully debunking it for 10 days because of NDAs is awful behaviour.

 

[jigger]

The PC master race are easy targets. Show them a bar chart and...

 

[NinjaCool]

UFD Tech trying to figure out what had to happen for the 2700X to give those PT results.

 

[sandys]

As you said, using "game mode" is a huge issue because it disables half the cores! One could argue that AMD has named it badly but the Ryzen Master app makes it clear it's designed for Threadripper chips only. How you can say "PT seems completely reasonable" and "just a game mode issue really, that is all" in the same sentence baffles me. If AMD commissioned a comparison against an Intel chip with half its cores disabled without making it incredibly clear, would you really have the same opinion?

It is an easy mistake to make, they have said they are doing further testing, which is the reasonable thing to do.

The Intel machine using 2666 MHz XMP will have a significant advantage over the AMD system using 2933 MHz without XMP. Both machines would definitely be capable of running 2933 MHz RAM with XMP enabled (with 2 DIMMs anyway) and even if they weren't, why not disable XMP on both? This is a huge red flag.

XMP profiles don't work on AMD, to this day if I want an unstable mess of a machine that is the option I choose, AMD systems rely on what is in the firmware of CPU and BIOS more so than pre set XMP timings in an DRAM profile setup for Intel.

Either PT are completely incompetent or very disingenuous. It could easily be the first considering they used 4 DIMMs in every test for a total of 64 GB of RAM, a configuration almost no gamer would use (not just because it's unnecessary but because it lowers your maximum achievable speeds/timings). Either way, the fact that Intel endorses and trots out these results whilst preventing competent reviewers from fully debunking it for 10 days because of NDAs is awful behaviour.

4 DIMMs is an issue sure but has impact on Intel dual channel platform as it is on AMD, 4 DIMMs is something people do, whether it is 4x4, 4x8 or 4x16 the load on the memory controller has an impact due to number of sticks as much as quantity.

People getting the knickers in a twist because it is paid for by Intel, but actually PT have detailed processes they have went through and versions software and BIOS etc, it was a fairly well done bit of work, more detail there than you get in many a youtube vid referenced in threads here, how they respond to the backlash over the flaws in testing will be the more important thing and that last GN vid suggests they are doing the work.

 

[jigger]

It is an easy mistake to make

Is it? Is that really an easy mistake to make. Maybe it's just me.

 

[Hotwired]

People getting the knickers in a twist because it is paid for by Intel

Being paid for by Intel requires a grain of salt.

You need more than a grain of salt to swallow the results and the fact that Intel supports them and that they choose to advertise these results weeks before anyone can independently show their testing.

 

[GAC]

got to agree with you DragonQ PT seem either incompetent or disingenuous also or even possibly just too stuck in their ways and attitude of knowing what they are doing.

watching this video, a few times it comes across as they KNOW more than anyone else. the part where he states "iv probably been testing longer than you have been alive" then goes on to say the tests where for the normal user not gamers yet they test like a dozen games and fits a high end cooler to the intel parts. and then the 64 gigs of ram, dont know many normal users blowing £500+ on memory either gotta wonder who came up with the specs and who they aimed them at. then talking about game mode and how various people said it was the wrong thing to do. so again seems more ineptitude than outright shenanigans on that one.

the guy does seem genuinely open but honestly to me it just comes across as in they really didn't know 100% what they where doing when it came to the amd cpu's, the biggest thing that stuck to me was how they talked about normal users yet slap together 64 gigs of ram and then use a high end air cooler on the intel chips.

all in all a rather fluffed set of testing.

 

[sandys]

Is it? Is that really an easy mistake to make. Maybe it's just me.

When TR was launched benchmarks for the 1900x made it look attrocious because similar settings were chose for it as the rest of the TR range cutting cores and of course it is like a 1800/2700 it doesn't have many, a few mainstream sites like Toms etc got it wrong and so would a number of consumers. I can see how it can be genuine.

 

[DragonQ]

It is an easy mistake to make, they have said they are doing further testing, which is the reasonable thing to do.

Are there many "casual users" out there complaining about CPU cores being missing with a Ryzen CPU? I'm not sure it's that easy. I don't even know if it's called "game mode" any more in the settings, I thought it was called "compatibility mode" these days? Either way, Ryzen Master is not even necessary. They say they aren't overclocking so why (a) install it at all, (b) install it but then only change ONE setting? It makes sense for Threadripper but not Ryzen, so clearly they're not that clued up on this.

XMP profiles don't work on AMD, to this day if I want an unstable mess of a machine that is the option I choose, AMD systems rely on what is in the firmware of CPU and BIOS more so than pre set XMP timings in an DRAM profile setup for Intel.

I doubt a second gen Ryzen system would struggle with 2933 MHz but maybe with the 4 sticks they used. Let's pretend they couldn't get XMP to work. In that case, they have two options: don't use 4 sticks of RAM (sensible), or just don't use XMP on any of the machines, since they claim to want the test to be "fair".

The interview with GN is pretty enlightening because the guy is all over the place with his excuses and reasoning. To be fair, he was put in a difficult situation and I wouldn't expect his answers to be perfect. However, he says within about 30 seconds that they wanted the test to be "fair" by keeping parameters the same but at the same time keep as close as possible to a "normal" or "out of the box" experience. They used an aftermarket cooler on Intel CPUs only ("out of the box" I guess since they don't come with coolers, definitely not fair vs Ryzen stock cooler), and then only used XMP on Intel (definitely not "out of the box" or fair).

People getting the knickers in a twist because it is paid for by Intel, but actually PT have detailed processes they have went through and versions software and BIOS etc, it was a fairly well done bit of work, more detail there than you get in many a youtube vid referenced in threads here, how they respond to the backlash over the flaws in testing will be the more important thing and that last GN vid suggests they are doing the work.

Not purely because it's paid for by Intel, but because of the timing compared to other outlets that have NDAs. I actually think the fact they disclosed so much detail is great, it's a shame they just don't seem to know what they're doing when it comes to gaming testing/benchmarking. The fact that they are trying to respond to the criticism suggests there isn't any malicious intent (compare to CTS Labs who bolted from interviews when the questions got tough). However, the fact that they (so far) completely stand by their testing methodology suggests they are indeed incompetent.

 

[doodah]

Just watching Gamer Nexus' follow up video when he turned up to PT's offices, a tad hostile from the start....definitely some explaining to do.

 

[SiDeards73]

It is an easy mistake to make, they have said they are doing further testing, which is the reasonable thing to do.



XMP profiles don't work on AMD, to this day if I want an unstable mess of a machine that is the option I choose, AMD systems rely on what is in the firmware of CPU and BIOS more so than pre set XMP timings in an DRAM profile setup for Intel.



4 DIMMs is an issue sure but has impact on Intel dual channel platform as it is on AMD, 4 DIMMs is something people do, whether it is 4x4, 4x8 or 4x16 the load on the memory controller has an impact due to number of sticks as much as quantity.

People getting the knickers in a twist because it is paid for by Intel, but actually PT have detailed processes they have went through and versions software and BIOS etc, it was a fairly well done bit of work, more detail there than you get in many a youtube vid referenced in threads here, how they respond to the backlash over the flaws in testing will be the more important thing and that last GN vid suggests they are doing the work.

Sorry but your wrong, just face it, you are completely wrong... The whole debacle was paid by Intel, the memo probably stated "Make our kit look the best possible" you would have to be completely incompetent to use the settings they used. The whole thing was a blatant expertly done hatchet job to make Intel look utterly amazing, why? because Intel know there is a few % difference between their latest and greats with Soldered TIM vs their previous gen, which wasnt massively faster than AMD's best either. So Intel set this up, stacked the deck hugely in their favour to make it look like their chip is some earth shattering, world exploding monster, 50% Faster than Ryzen!!! its pure sensationalism and headling at its finest...

The morons they employed to do this hatchet job for them knew exactly how to gimp AMD to make Intel look great, and they used AMD's own mistakes against them to come off smelling like roses...

Anyone know has not been living under a rock and works in the tech industry knows verbatim 100% that Ryzen likes fast ram, that is your first clue, this is "professional" outfit, whom incidentally Intel has employed previously before for hatchet jobs, so its not like they dont have previous. Then they used the "Game Mode" against AMD, which in all honesty is AMD's own fault as it should never even work on standard Ryzen chips and it definitely could have had a better name. And they used a 1080ti at 1080p to simulate the standard user, well im a standard user, i dont use a 1080ti or game at 1080p, 1070 @ 1440p for me personally, but none the less, the whole charade was done as a professional hatchet job.

Unfortunately it has backfired spectacularly because the Internet is a wonderful thing, something which Intel probably forgot existed, doesnt take long after looking at the results Intel gave and a bit of googling to find out somethings not right, dig into the company doing the testing and find out they are paid shills of Intel, ok now things start to add up... then the tech tubers get into it and debunk it for what it is, a sad attempt at trying to gain some sales back from AMD. Something which incidentally seems to have had the opposite effect, but thats mainly down to Intels ridiculous pricing, and their current chip constraint issues.

To stand up and defend Intel on this shows a complete and utter lack of understanding of morals and integrity, and a massive naivety towards the people presenting the info (Intel and their shills) and the current eco-system of hardware in general.

 

[Troezar]

When TR was launched benchmarks for the 1900x made it look attrocious because similar settings were chose for it as the rest of the TR range cutting cores and of course it is like a 1800/2700 it doesn't have many, a few mainstream sites like Toms etc got it wrong and so would a number of consumers. I can see how it can be genuine.

I'll give you the benefit of the doubt and go with naive. Using a third party as a meat shield is as old as the hills. Politicians have been using it for centuries, in fact some would argue politicians are a meat shield themselves protecting the real power that never gets voted out. Intel have a lot of history, it's all out there and this is mild for them.

PT are massively incompetent and about as unscientific as it gets. I audited scientists for years and would have laughed this out of the office if it had been submitted.

 

[SiDeards73]

got to agree with you DragonQ PT seem either incompetent or disingenuous also or even possibly just too stuck in their ways and attitude of knowing what they are doing.

watching this video, a few times it comes across as they KNOW more than anyone else. the part where he states "iv probably been testing longer than you have been alive" then goes on to say the tests where for the normal user not gamers yet they test like a dozen games and fits a high end cooler to the intel parts. and then the 64 gigs of ram, dont know many normal users blowing £500+ on memory either gotta wonder who came up with the specs and who they aimed them at. then talking about game mode and how various people said it was the wrong thing to do. so again seems more ineptitude than outright shenanigans on that one.

the guy does seem genuinely open but honestly to me it just comes across as in they really didn't know 100% what they where doing when it came to the amd cpu's, the biggest thing that stuck to me was how they talked about normal users yet slap together 64 gigs of ram and then use a high end air cooler on the intel chips.

all in all a rather fluffed set of testing.

No thats not incompetence, its clever feigned ignorance, the fact they knew exactly how to gimp AMD to achieve that level of results shows these guys have half a clue how to setup a machine, a lot of what they did was utter nonsense, they knew full well Ryzen (Not threadripper) doesnt like 4 dimm slots populated and will run lower mhz ram to be stable, if they had used 3200mhz c14 ram the results would be a lot better, same with the game mode, they can say "well AMD name it game mode, so its for gaming" and thats a get out of jail for them etc... same with the cooler, well the Ryzen cooler would cost $40 so we can use an aftermarket cooler on the Intel rig as it doesnt come with one etc etc...

That is not incompetence, that is a cleverly thought out and planned hatchet job on your rival.

 

[dragonslayer]

Lmfao if they're truly going out their way to produce misleading results/ fake results then doesn't that just go to show how afraid they truly are of AMD and more importantly how they cannot compete with AMD going forward

 

[sandys]

There are no warnings when you switch Game mode on, it just reboots and disables cores, that is it.

My system after enabling game mode becomes a nice 4c/8t.

 

[DragonQ]

Remember that PT are not Intel. As incompetent as this testing was, if Intel wanted to essentially fund a hatchet job, PT may not have been party to that. Intel may have just scouted out a company that they thought would do the testing in the way they wanted. Hell, they might've even commissioned 20 companies to do it and picked the one with the nicest numbers for them. Intel may also have subtlety or blatantly prodded them in certain directions, e.g. "oh make sure you use the same amount of RAM on each system" or "we want to be fair, so use game mode for AMD's chips since this is gaming test", etc.

 

[jigger]

When TR was launched benchmarks for the 1900x made it look attrocious because similar settings were chose for it as the rest of the TR range cutting cores and of course it is like a 1800/2700 it doesn't have many, a few mainstream sites like Toms etc got it wrong and so would a number of consumers. I can see how it can be genuine.

Well Ryzen looked as good as anything else +/- 10% but the reason for TR performance is obvious from just looking at the configuration of the chip and platform. I'm sure being a new from the ground up platform added a little more complexity when analysing what AMD had to offer initially and what AMD are offering is something Intel are fully aware of.