Intel Stands Behind Controversial Tests That Favored Its CPU Over AMD's

[TerryLaze]

Err what? Putting 4 DIMMS in had nothing to do with the memory issue, its that they allowed it to use the default configuration which is ultra conservative, compared to going into the BIOS and setting XMP in the Intel system thereby giving it faster speeds and more aggressive timings. The memory issue had nothing to do with having 4 dimms or dual ranked ram.

Sorry the argument doesn't hold water. Ryzen and Intel both require the same exact tweaks to get the best performance out of the installed processor and memory, just like they always have and always will, the only way to avoid such tweaks on BOTH systems is to not care about memory speed and use baseline 2133 mhz ram.

Those two points are what the hardwareunboxed video stated, together with the false settings.
Still, the argument here is that correct ram settings would negate the 50% worse performance which is ridiculous.

 

[Rogue Leader]

Err what? Putting 4 DIMMS in had nothing to do with the memory issue, its that they allowed it to use the default configuration which is ultra conservative, compared to going into the BIOS and setting XMP in the Intel system thereby giving it faster speeds and more aggressive timings. The memory issue had nothing to do with having 4 dimms or dual ranked ram.

Sorry the argument doesn't hold water. Ryzen and Intel both require the same exact tweaks to get the best performance out of the installed processor and memory, just like they always have and always will, the only way to avoid such tweaks on BOTH systems is to not care about memory speed and use baseline 2133 mhz ram.

Those two points are what the hardwareunboxed video stated, together with the false settings.
Still, the argument here is that correct ram settings would negate the 50% worse performance which is ridiculous.

I agree, but it has been proven that Ram speed does make a significant performance difference with Ryzen, definitely not 50%, however the fact they also hacked off 4 cores using Game mode which is not meant for Ryzen, is the major issue.

 

[TerryLaze]

I agree, but it has been proven that Ram speed does make a significant performance difference with Ryzen, definitely not 50%, however the fact they also hacked off 4 cores using Game mode which is not meant for Ryzen, is the major issue.

CS:GO only uses 3 threads and the difference is still 50% so game mode might have affected some results(some of them heavily, like ashes) but the outcome is the same anyway.

 

[Rogue Leader]

I agree, but it has been proven that Ram speed does make a significant performance difference with Ryzen, definitely not 50%, however the fact they also hacked off 4 cores using Game mode which is not meant for Ryzen, is the major issue.

CS:GO only uses 3 threads and the difference is still 50% so game mode might have affected some results(some of them heavily, like ashes) but the outcome is the same anyway.

If your only metric of performance is how well it plays CS:GO sure the outcome is the same. But clearly people play other things and use other software with their systems.

 

[littleleo]

"As usual, we advise readers to wait for reviews instead of plunking down their hard-earned cash based solely on vendor benchmarks."

What no just buy recommendation on the new Intel 9th Gen too? They must not have paid enough to Tom's for that.

It will be interesting to see more balanced benchmarks after they actually start shipping the 9th gen chips.

 

[TerryLaze]

If your only metric of performance is how well it plays CS:GO sure the outcome is the same. But clearly people play other things and use other software with their systems.

That's why they said up to while everybody else lost their minds,that's what up to means,in extreme situations,on occasion,they did not claim a 50% avg, they did not say that 50% would even be common, they said up to.

 

[jpe1701]

Yeah it should be a great chip, and I'm sure it will be faster than my 2700x, just sucks for the consumer that is seriously considering one and now the test is tainted. Puts a bad light on Intel when they have enough on their plate trying to get 10nm working and keeping up with 14nm. I can't believe that with Intel's resources that AMDs come back took them so off gaurd but all these little debacles they keep getting into seem like they were caught with their pants down. Competition is a beautiful thing. Now if they would only lower the price, but I digress.

 

[theoldhenk]

The Game Mode testing aspect of the article is confusing.

If the 2700X doesn't actually include a Game Mode setting, how did PT disable four cores, and more importantly, is there any plausible explanation other than they were purposefully attempting to misrepresent the results.

It just seems a bit farfetched that PT would go to such unnecessary lengths to skew results that should already favor Intel. Especially when its obvious that disabling half the Ryzen cores wouldn't go unnoticed.

I'm also curious whether PT was transparent with respect to the testing procedure. If they weren't trying to hide their methodology, I imagine there's more to the story than Intel/PT risking the inevitable bad press and lost consumer goodwill over CPU benchmarks that would almost certainly favor them anyway.

 

[Brian_R170]

Wait, what? Ryzen's Game Mode makes gaming performance worse? Even it it's only on some Ryzen models, it is a disservice to Ryzen customers.

 

[Rogue Leader]

Wait, what? Ryzen's Game Mode makes gaming performance worse? Even it it's only on some Ryzen models, it is a disservice to Ryzen customers.

Its not made for Ryzen only Threadripper which has more than 8 cores. Disabling half of them helps it in games only allowing the remaining ones to turbo higher.

 

[Brian_R170]

Wait, what? Ryzen's Game Mode makes gaming performance worse? Even it it's only on some Ryzen models, it is a disservice to Ryzen customers.

Its not made for Ryzen only Threadripper which has more than 8 cores. Disabling half of them helps it in games only allowing the remaining ones to turbo higher.

Yeah, I know. Like I said, even if it makes performance worse only on some Ryzen models, it's a disservice.

If I play games and my CPU has a feature called Game Mode, I'm going to turn on Game Mode and I'm going to expect some improvement. If it actually makes all games perform much worse on a specific CPU, which is what I'm gathering from this info presented in this article, then the feature either should not be available or should not be called "Game Mode". Maybe they should call it "Waste your time seeing if this makes performance better or worse" mode.

 

[Rogue Leader]

Wait, what? Ryzen's Game Mode makes gaming performance worse? Even it it's only on some Ryzen models, it is a disservice to Ryzen customers.

Its not made for Ryzen only Threadripper which has more than 8 cores. Disabling half of them helps it in games only allowing the remaining ones to turbo higher.

Yeah, I know. Like I said, even if it makes performance worse only on some Ryzen models, it's a disservice.

If I play games and my CPU has a feature called Game Mode, I'm going to turn on Game Mode and I'm going to expect some improvement. If it actually makes all games perform much worse on a specific CPU, which is what I'm gathering from this info presented in this article, then the feature either should not be available or should not be called "Game Mode". Maybe they should call it "Waste your time seeing if this makes performance better or worse" mode.

Its a switch in Ryzen Master which actually says its for Threadripper on it. So if you go through the effort to download Ryzen master, look for the switch, ignore that it says its for Threadripper and then turn it on, you're just kind of stubborn ignorant. On the specific CPU it was designed for it makes games work better. Maybe there should be a lock on it to not work on a regular Ryzen CPU, but at some point people need to take responsibility for their actions.

 

[Brian_R170]

Its a switch in Ryzen Master which actually says its for Threadripper on it. So if you go through the effort to download Ryzen master, look for the switch, ignore that it says its for Threadripper and then turn it on, you're just kind of stubborn ignorant. On the specific CPU it was designed for it makes games work better. Maybe there should be a lock on it to not work on a regular Ryzen CPU, but at some point people need to take responsibility for their actions.

Wait, you mean AMD's own software allows it? Well, then you are right. There are people that need to take responsibility for their actions. They all work for AMD.

Surely AMD knows what the feature does to specific CPUs and owning the software means they could have easily disabled it for those CPUs, or if there is still some valid reason for having the feature on a non-Threadripper, they could rename it for it's intended purpose on those CPUs or at least warn the user of the expected performance decrease.

 

[Rogue Leader]

Its a switch in Ryzen Master which actually says its for Threadripper on it. So if you go through the effort to download Ryzen master, look for the switch, ignore that it says its for Threadripper and then turn it on, you're just kind of stubborn ignorant. On the specific CPU it was designed for it makes games work better. Maybe there should be a lock on it to not work on a regular Ryzen CPU, but at some point people need to take responsibility for their actions.

Wait, you mean AMD's own software allows it? Well, then you are right. There are people that need to take responsibility for their actions. They all work for AMD.

Surely AMD knows what the feature does to specific CPUs and owning the software means they could have easily disabled it for those CPUs, or if there is still some valid reason for having the feature on a non-Threadripper, they could rename it for it's intended purpose on those CPUs or at least warn the user of the expected performance decrease.

I mean, its a known feature as to what it does, it says its for threadripper. If you're so inclined to tinker enough with your system to download Ryzen master and go in and start clicking stuff, you (the user) should take some responsibility for the mistake.

Sorry I don't agree that AMD needs to handhold everyone through overclocking or configuring their system. Configuring a PC requires a modicum of research and common sense.

 

[wownwow]

"Intel Stands Behind ..."

Intel stands behind their designs with spec violations that caused the security problems of "Meltdown" and "Foreshadow"; the processor chips function as INTENDED, according to the ex-CEO!!!

 

[cryoburner]

Well tom's already confirmed 33% difference in both games and apps between the stock i7-8700K and the stock 7 2700X,
(same performance with 33% less cores for the intel)
the 9900k has yet another 33% more cores than the 8800k.
https://www.tomshardware.com/reviews/cpu-hierarchy,4312.html

What are you even talking about? You seem to have a misunderstanding about how CPUs work. Adding more cores doesn't automatically equate to adding more performance, since few applications will even make use of those extra cores. You linked to the CPU hierarchy chart, where Tom's results show the stock i7-8700K performance being quite similar to that of the 2700X overall. But nearly all applications don't even use all of the cores and threads that these processors have as it is, so the 2700X's extra cores are just sitting around not doing anything in most of the tests. Likewise, adding two more cores to an 8700K will do very little for performance, aside from in a few heavily multithreaded applications like CPU-based video encoders and file extraction utilities. And while games are slowly becoming more multithreaded, I wouldn't expect 8 cores + SMT to provide any significant benefit to gaming compared to 6 cores + SMT anytime soon. The slightly higher clock rates might help a little in high-refresh rate scenarios, but for the most part performance should be more or less indistinguishable between these CPUs to the end user.

To further this point, just look at the Ryzen 2600X, which has the same number of cores and threads as the i7-8700K. In that comparison you linked to, it performed just 2% behind the 8700K at gaming, and just 18% behind it in their application benchmarks. Overall performance tends to be quite similar between a 2600X and a 2700X in the vast majority of applications, and similarly, real-world performance in most applications and practically all current games should be quite similar between an 8700K and a 9900K.

 

[jpe1701]

I have to say I am with Rogue Leader, it says Threadripper right on it. They consolidated the software this time around so that it's one download for both Threadripper and Ryzen. It tells you what CPU you have right on the same screen so anyone who can read should be able to figure it out. It's pretty straight forward software. I don't necessarily disagree though that maybe they should disable it on Ryzen but still if you're just hitting switches without knowing what it does and disregarding the label it's on you.

 

[alextheblue]

"As usual, we advise readers to wait for reviews instead of plunking down their hard-earned cash based solely on vendor benchmarks."

What no just buy recommendation on the new Intel 9th Gen too? They must not have paid enough to Tom's for that.

It will be interesting to see more balanced benchmarks after they actually start shipping the 9th gen chips.

Well to be fair that was an opinion piece by one guy and it was a counter article to Derrick's "wait" opinion piece. But anyway yeah I still see people talking about it, and not just here. PCGamesN's original coverage (they took the results on blind faith) of this Intel-paid controversy was drawing comparisons to that incident in their comments. Anyway yes, the just buy it article did a lot of damage, which is a shame because the actual review was both fair and well-written IMHO, and yet many randos will only recall the one piece and assume coverage must have been biased.

 

[jpe1701]

It really is a shame that more people didn't notice the fact that it was a counter article to the wait article. I took it as the Tom's crew just having some fun with it and it's too bad that so many took offense. The review, the actual serious tech press stuff was pretty fair I thought. But I'm off topic so I'll shut up.

 

[Krazie_Ivan]

lil history helps to add context

https://www.youtube.com/watch?v=3V8pEsjNa4Q

 

[cream]

money buys you the right to cheat oh dear

 

[paul prochnow]

Lame tweak time.
Bad RAM and bad Ryzen Master choice.
My RAM is 3333Hhz and you need 3000MHz at least to get Ryzen into
that nice Infinity Fabric Zone.

 

[kalmquist]

The bottom line is that Intel doesn't believe it can make an honest case for the i9-9900K. I was expecting the i9-9900K to be a serious problem for AMD, because Intel is matching AMD on core count, and AMD won't match Intel on IPC until Zen 2 at the earliest. I suppose that Intel is commissioning and "standing by" misleading benchmarks because even they think that the i9-9900K is overpriced.

 

[OFF3ND3R]

It’s 1998 all over again Athlon vs PIII. Intel benchmarks were biased. Clock for Clock the PIII couldn’t hold a candle to the Athlon until the coppermine refresh hit.

 

[justin.m.beauvais]

Honestly, I'm not sure what Intel or PT was trying to do here. Intel has the IPC and clock speed advantage. All they had to do was put some cores on a chip, clock it at 5 GHz and win at gaming. People would gobble up chips and all would be happy and good. Instead they pull these shenanigans. Intel has just been on a roll of bad ideas and bad decisions that have basically let AMD close the gap. The good news for AMD is that NVidia is starting to make bad decisions as well... so maybe AMD will catch up with them in a couple years.