Why does Ryzen 7 1800X performs so poorly in games?

[juanrga]

If Naples is anything, amd primarily made zen to have be better at workstation/server environment. They showed all their demos at 4K. I understand the this thing about 1080p. But who demos at 1080p and calls them enthusiast gaming system? I mean we can go back and forth on it all day. No one company is going to demo titan x pascal at 1080p.

Oh look we are running at 1080p and getting 160fps but the other chip is getting 180fps. It just sounds stupid. I know if I am getting a titan xp. I am sure as hell not running at 1080p. And I don't expect any company to show 1080p as enthusiast system.

Reviews have explained why they tested gaming performance at different resolutions and why the 1080p tests are important to understand the real gaming performance of any chip. TechSpot also wrote an article explaining GPU bottlenecks and the reason for testing gaming performance on CPU-tests. I also explained all of this twice in this thread and give relevant links. I see the reviewers conclusions continue being ignored and the same flawed arguments are being repeated once again.

 

[juanrga]

You talk as if Intel has never had issues with some of their new architectures. You are living in a fantasy world. The scheduler issue will be worked . If only AMD had the clout of Intel, they would have Microsoft kissing their feet.

Lets assume Juanrga is right for once and AMD scheduler fixes things by 5%

I said a better scheduler would improve performance by 2--5%, and not fix the real problem... The important thing is that it seems that AMD itself has killed the dreams that a magic scheduler will fix Ryzen's problems

https://www.pcper.com/reviews/Processors/AMD-Ryzen-and-Windows-10-Scheduler-No-Silver-Bullet

Closing Thoughts

What began as a simple internal discussion about the validity of claims that Windows 10 scheduling might be to blame for some of Ryzen's performance oddities, and that an update from Microsoft and AMD might magically save us all, has turned into a full day with many people chipping in to help put together a great story. The team at PC Perspective believes strongly that the Windows 10 scheduler is not improperly assigning workloads to Ryzen processors because of a lack of architecture knowledge on the structure of the CPU.

In fact, though we are waiting for official comments we can attribute from AMD on the matter, I have been told from high knowledge individuals inside the company that even AMD does not believe the Windows 10 scheduler has anything at all to do with the problems they are investigating on gaming performance.

In the process, we did find a new source of information in our latency testing tool that clearly shows differentiation between Intel's architecture and AMD's Zen architecture for core to core communications. In this way at least, the CCX design of 8-core Ryzen CPUs appears to more closely emulate a 2-socket system. With that, it is possible for Windows to logically split the CCX modules via the Non-Uniform Memory Access (NUMA), but that would force everything not specifically coded to span NUMA nodes (all games, some media encoders, etc) to use only half of Ryzen. How does this new information affect our expectation of something like Naples that will depend on Infinity Fabric even more directly for AMD's enterprise play?

 

[razor1]

The DIY segment is tiny tho.

Mindfactory.de that was one of the 9 preorder sites sold ~1785 Ryzen chips so far.
http://www.mindfactory.de/Hardware/Prozessoren+(CPU)/AMD+Desktop/Sockel+AM4.html

True without the OEM market, AMD can't make any marketshare gains.

 

[razor1]

I said a better scheduler would improve performance by 2--5%, and not fix the real problem... The important thing is that it seems that AMD itself has killed the dreams that a magic scheduler will fix Ryzen's problems

https://www.pcper.com/reviews/Processors/AMD-Ryzen-and-Windows-10-Scheduler-No-Silver-Bullet

oddly enough I told OS2wiz to read up on Numa lol. Not really odd if one knows what they are looking at.

Sigh...... Do you need links to the windows schedular and what it does? It has nothing to do with how the CPU schedules graphics or application needs, it has to do with window's needs so things don't get screwed up when multitasking.

https://msdn.microsoft.com/en-us/library/windows/desktop/ms685096(v=vs.85).aspx

read.

Then tell me what tells the windows scheduler what is priority and what is to be done at what time.

Then if you want further reading, check out NUMA support, that is where the problem is most likely occurring. Its all DEVELOPER controlled.

So if Windows scheduler is where the problem is, it will happen in all programs not specifically games. That is not where the problem resides!

Please separate Windows scheduler from the CPU instruction scheduler in your mind, they are not one and the same. One is a pass through that checks for its own needs vs the other (where the problem is actually happening) feeds the ALU's of the CPU.

I wish you guys keep your eyes open, you don't even need to read the review sites to figure out what is going on lol, I answered this here yet the same old boomerang and deflections. This problem can't be deflected, there is no way around it so just accept its there and move on.

 

[juanrga]

oddly enough I told OS2wiz to read up on Numa lol. Not really odd if one knows what they are looking at.

Before launch, he was negating the existence of any latency issue

and he pretended in other forums that it was all an invention of mine: "I found a tweet from Juanrga stating that Ryzen has latency issues. He seems to invent a new theory every week to discredit Ryzen and the AMD engineering team."



I think we can close this thread because we know which is the answer to the title of this thread. RyZen has a fundamental design flaw and no magic scheduler will fix this. RyZen is good at throughput workloads like rendering/encoding. On gaming, it is about 20% behind Broadwell-E clock-for-clock, and play games like an i5 at stock.

 

[razor1]

well can't really blame him either, not many people understand the difference between distributed computing over multi CPU's vs multithreaded, even programmers whom don't have that type of experience.

 

[-Sweeper_]

Latest from PCLab

Spoiler

If you're CPU limited, sometimes even at high resolutions, 7700K is faster. Days after the most hyped launch in CPU history and the best perf/$ CPU in Ryzen's lineup (1700) is at #9 in Amazon US best sellers. Looks like what really drives DIY sales are gamers.

 

[ZeroBarrier]

Latest from PCLab

Spoiler

If you're CPU limited, sometimes even at high resolutions, 7700K is faster. Days after the most hyped launch in CPU history and the best perf/$ CPU in Ryzen's lineup (1700) is at #9 in Amazon US best sellers. Looks like what really drives DIY sales are gamers.

No surprise there, the gaming industry is just shy of 100 billion dollar market. It's actually the reason PC has stayed afloat and not been snuffed out.

 

[Presbytier]

Latest from PCLab

Spoiler

If you're CPU limited, sometimes even at high resolutions, 7700K is faster. Days after the most hyped launch in CPU history and the best perf/$ CPU in Ryzen's lineup (1700) is at #9 in Amazon US best sellers. Looks like what really drives DIY sales are gamers.

Now hold on here there is that one benchmark from a site no one has ever heard of that clearly shows that the 1080Ti favors Ryzen like a new calf favors his mommies tits, so clearly these tests are wrong

 

[juanrga]

well can't really blame him either, not many people understand the difference between distributed computing over multi CPU's vs multithreaded, even programmers whom don't have that type of experience.

His ignorance is not excuse for him calling names to people was only providing accurate information about RyZen and its gaming performance.

Let us see if he will take the new information and the last PcPer analysis and will write about the latency issues of RyZen in the same forums where he pretended it was all an invent to "discredit Ryzen and the AMD engineering team"...

 

[OrangeKhrush]

Will still wait and see on outcomes of windows fixes, Microsoft themselves have stated there is comflict, but more in how the CCX and cache allocations work, it is probably bits and pieces to fix by all parties concerned and I will just wait until more issues are systematically worked through.

 

[cybereality]

Now hold on here there is that one benchmark from a site no one has ever heard of that clearly shows that the 1080Ti favors Ryzen like a new calf favors his mommies tits, so clearly these tests are wrong

 

[Dreamerbydesign]

Latest from PCLab

Spoiler

If you're CPU limited, sometimes even at high resolutions, 7700K is faster. Days after the most hyped launch in CPU history and the best perf/$ CPU in Ryzen's lineup (1700) is at #9 in Amazon US best sellers. Looks like what really drives DIY sales are gamers.

Ouch that dx12 hit an score hurts. I didn't realize there was that much of a difference in any game.

 

[OrangeKhrush]

From the comments in PCPer

This article has good initial data collection but questionable analysis and a premature conclusion, since SMT core allocation and inter-core/CCX latencies are only part of the picture.

Something that could murder Zen performance but will not affect smaller Haswells/Broadwells much is process migration across arbitrary cores, that is, back and forth across CCXs. Smaller Xeons (and their -E workstation counterparts) have every core sitting on a single ring bus with cache lines hashed or otherwise interleaved by low-order address bits across the set. At worst, an i7-5960X needs to rewarm its L2 cache, which in 256 kiB and has a 64B pipe to L3.

On the other hand, Zen has twice as large L2s at 512 kiB, only 32B lanes to L3, and higher latencies to the remote CCX, so full cache rewarming will take at least 4x more time even in the case that all the needed lines are still in the remote L3 cluster.

Before claiming that scheduling has nothing to do with the issue, it needs to be measured and provable that the scheduler does not lightly bounce given threads across CCXs on a moderately to heavy system load.

As a final note, it seems disappointing that the scheduler put 4 moderately busy threads on the same CCX, since most workloads would benefit more from bigger shares of local L3 than from lower inter-core latency.

That is the issue we see, the same issue is less pertinant on Windows 7 which tends to allocate heavy loads on each CCX better and it is why gaming improves significantly on Windows 7, that may also be why cutting off one unit of the CCX shows marked improvements, windows is treating it then as a single unit without overspilling and forcing latency between transporting loads.

 

[Gideon]

It's like a echo chamber in here, things just keep repeating and repeating and repeating.............

 

[OrangeKhrush]

It's like a echo chamber in here, things just keep repeating and repeating and repeating.............

AMD threads get all the internet traffic.

Anyways recognizing cores and threads is one thing, the ability to utilize them effectively is another and the schedular while I don't expect 20% type gains, 5% or up to possibly 10 or thereabout is possible. The impact on how Ryzen's load is split has a major effect on gaming which is less static in nature. Like I said I am just going to wait on this, the growing pains are there but it has to start somewhere.

 

[lolfail9001]

From the comments in PCPer

Ironically, Allyn made a comment that did establish that scheduler bounces threads between CCXs. Also, his final note is wrong, in absolute majority of workloads 2+2 is inferior to 4+0 from what we know. Even with Linux scheduler, that is supposedly aware of CCX concept.

 

[Presbytier]

From one of the comments in the PCPER article

Yes, and NO!

You just proved all the sitre that show that windows10 sheduler is not working correctly.

Why? because we see a repeated 20% win in games by using windows7 VS windows10. (min frame)

The issue is not SMT... its cross complex switching and erroneous cache estimation.

Since cross complex talk at 1/2 DRam speed, thats why dram speed effect gaming that much on windows 10.
Because when windows put a thread from core3 to core4, all the L3 data is non local. and its going over the fabric at 1/2 dram speed. (but on a wider bus)

To prove this. run the same games under windows7 VS windows 10.
massive boost in frame rate consistency.

Also. look at the FPS for a simulated 4 core Ryzen. its like 1:1 to a i7-7700K at equal clock.

Like I suggested, juts try to review the 1800x with half the core (from the same CCX) disabled.
This will eliminated the cross CCX task switching of windows10

This is also why 2166 VS 3200 dram makes a huge difference.

Anyways, you need to think of this problem as cache locality, CCX organization, fabric performance (ram clock).

Not at a 10,000 feet view from task manager graph.
.

 

[razor1]

From one of the comments in the PCPER article

actually who ever posted that is right, that is the exact same problem that 2 CPU's communicating have, and why without a high speed interconnect between them they can't function at all in instances where data needs to be shared between the two or more CPU's.

 

[Presbytier]

actually who ever posted that is right, that is the exact same problem that 2 CPU's communicating have, and why without a high speed interconnect between them they can't function at all in instances where data needs to be shared between the two or more CPU's.

I don't really understand why AMD would split the CCX function into two, but then again this is clearly a Server chip downscaled so shouldn't realy surprise me.

 

[razor1]

I don't really understand why AMD would split the CCX function into two, but then again this is clearly a Server chip downscaled so shouldn't realy surprise me.

I don't think they did it on purpose, something is going or went wrong. If we look at how Intel's chips and Ryzen are set up, they look fairly close, and it looks like Ryzen's caching should be better with increased L2 cache amounts, but it all falls apart once ya get to L3 and its latency issues.

 

[Presbytier]

I don't think they did it on purpose, something is going or went wrong. If we look at how Intel's chips and Ryzen are set up, they look fairly close, and it looks like Ryzen's caching should be better with increased L2 cache amounts, but it all falls apart once ya get to L3 and its latency issues.

The cache system is is a dumping system where it starts in L1 and works its way to the RAM based on size, so considering games are probably getting dumped to the RAM due to size that is why we see frametimes improved by faster RAM. What is odd is the L3 on Ryzen is 16mb I don't know why it would be dumping data in the RAM on games as an 8mb i7 7700k doesn't have the same issues. I am no engineer so I can only speculate.

 

[OrangeKhrush]

It is the expected bugs in a new arch, these take time, it is slso necessary for all things considered to integrate the uarch with ryzen, so that pinnacle ridge can benefit from pre adoption

 

[lolfail9001]

It is the expected bugs in a new arch, these take time, it is slso necessary for all things considered to integrate the uarch with ryzen, so that pinnacle ridge can benefit from pre adoption

Frankly, the data fabric looks like more of a design oversight than a bug. So, yeah, pinnacle ridge it is, for now one would have to settle for it as a trade-off. I guess AMD knew what they were doing when they tried to undercut Intel like that.

 

[Presbytier]

Frankly, the data fabric looks like more of a design oversight than a bug. So, yeah, pinnacle ridge it is, for now one would have to settle for it as a trade-off. I guess AMD knew what they were doing when they tried to undercut Intel like that.

PREORDERS! I mean seriously they had to know this was an issue ahead of time. If it was simply a schedular fix you would think they would have made sure that was fixed before launch.

 

[FrozenSun]

I find some of the discrepancies from situation to situation to be interesting, based on config or exact testing. For example, look at these GTA V results:




The first one is DX12, with a Titan X at 1080p. The second is DX11 with a gtx 1070 at 1080p. The 7700k murders the Ryzen chips in the first. In the second one, the tables flip and Ryzen is significantly ahead in both min and avg frames. Obviously the second (with a 1070) is a bit more GPU bound, but the results are far enough apart that it's clearly not totally GPU bound. The minimum frames in particular are heavily in favor of Ryzen.

That's just...weird...am I missing something? Why such a variance from DX11 to 12.

Edit: I think they're both dx11 and I just dreamed up the dx12 part. That makes the results even more odd...

 

[Presbytier]

I find some of the discrepancies from situation to situation to be interesting, based on config or exact testing. For example, look at these GTA V results:




The first one is DX12, with a Titan X at 1080p. The second is DX11 with a gtx 1070 at 1080p. The 7700k murders the Ryzen chips in the first. In the second one, the tables flip and Ryzen is significantly ahead in both min and avg frames. Obviously the second (with a 1070) is a bit more GPU bound, but the results are far enough apart that it's clearly not totally GPU bound. The minimum frames in particular are heavily in favor of Ryzen.

That's just...weird...am I missing something? Why such a variance from DX11 to 12.

GTAV is a DX11 only game. Settings are also not the same one is a 1080p test one is a 720p test

 

[FrozenSun]

GTAV is a DX11 only game. Settings are also not the same one is a 1080p test one is a 720p test

For some reason I remembered one being dx12, I must have dreamt that up. That makes it even weirder if they're both dx11.

They're definitely both 1080p though. I had the timestamp on the second one messed up a bit and it flips to the 720p test almost immediately, but if you back it up there 1080p test is there.

As mentioned, test setups are not identical, but it's odd just how different the results are. It's pretty much inconceivable that you can have results which (comparing the relative results of ryzen vs 7700k) flip flop that hard.

 

[OrangeKhrush]

PREORDERS! I mean seriously they had to know this was an issue ahead of time. If it was simply a schedular fix you would think they would have made sure that was fixed before launch.

The issues are well covered by the Stilt and others on Anandtech, they are workable problems as not all loads are treated the same and that is why there is the odd anomaly. You can be a negative nancy along with a few others, but it doesn't mean it gets menotonous reading the same crap over and over.

 

[Presbytier]

The issues are well covered by the Stilt and others on Anandtech, they are workable problems as not all loads are treated the same and that is why there is the odd anomaly. You can be a negative nancy along with a few others, but it doesn't mean it gets menotonous reading the same crap over and over.

I think you misunderstand me I would love to see Ryzen succeed, but I have my doubts and they are growing daily. CCX being split in two across 4 cores on one and 4 on the other split loads having high latency the issues with the cache all these things add up to real questions about the architecture for every day users and if the scaling down server chips is the best use for consumers. I look forward to seeing how this plays out over the next few months and hope to see some solid improvements by time the R5 launches as that 6/12 CPU has my eye. I am running a 3770k and it is growing long in the tooth I was looking at upgrading to a 7700k and probably still will but I am not in a hurry just yet and can wait to see if any of these issues can be ironed out.

 

[Presbytier]

For some reason I remembered one being dx12, I must have dreamt that up. That makes it even weirder if they're both dx11.

They're definitely both 1080p though. I had the timestamp on the second one messed up a bit and it flips to the 720p test almost immediately, but if you back it up there 1080p test is there.

As mentioned, test setups are not identical, but it's odd just how different the results are. It's pretty much inconceivable that you can have results which (comparing the relative results of ryzen vs 7700k) flip flop that hard.

The average results show the 7700k leading Ryzen in nearly every test sometimes by significant margins others not so much. I generally tend to look at large outliers with a bit of suspicion and yeah I notice now the second was 1080 as well after backing it up.

 

[Shintai]

I said a better scheduler would improve performance by 2--5%, and not fix the real problem... The important thing is that it seems that AMD itself has killed the dreams that a magic scheduler will fix Ryzen's problems

https://www.pcper.com/reviews/Processors/AMD-Ryzen-and-Windows-10-Scheduler-No-Silver-Bullet

Yep, that excuse died fast.

 

[Shintai]

Latest from PCLab

If you're CPU limited, sometimes even at high resolutions, 7700K is faster. Days after the most hyped launch in CPU history and the best perf/$ CPU in Ryzen's lineup (1700) is at #9 in Amazon US best sellers. Looks like what really drives DIY sales are gamers.[/spoiler]

And Hitman is an AMD game so no excuse there.

 

[NKD]

And Hitman is an AMD game so no excuse there.

By now everyone and their mother knows if you wanna game 24/7 go with 7700k. Everyone and their mother knew 7700k was best for gaming even against intels own $1000 dollar chips. Anything else About and you would like to get of your chest.

 

[juanrga]

From the comments in PCPer

If we start quoting comments then we must quote this:

We've been communicating with AMD throughout our testing process and they agree with our results, and that 'few lines of code' is doing a lot better than your armchair commenting.

As demonstrated, the W10 scheduler assign threads to the same CCX on slightly threaded workloads. In heavily threaded workloads, the scheduler assigns them to both CCX because there is no enough cores in a single CCX.

There is no silver bullet and even AMD agrees on this. Their public explanations of the latency issues with RyZen involve claims of close collaboration with game developers for further optimization of games for RyZen. Evidently game developers will not touch the OS scheduler, game developers will optimize individual titles to extract more performance in an attempt to hide the chip deficits. It remains to be seen how many of those optimizations will translate to non-Ryzen chips.

 

[juanrga]

By now everyone and their mother knows if you wanna game 24/7 go with 7700k. Everyone and their mother knew 7700k was best for gaming even against intels own $1000 dollar chips. Anything else About and you would like to get of your chest.

And still certain people pretends that current latency issues are only temporal 'bugs' and that BIOS updates, scheduler updates, <insert-any-other-lame-excuse-here> will fix everything soon. That same people has been pretending that mainstream reviews are all wrong or biased and then providing us links to ridiculous and unprofessional reviews and youtube videos were Ryzen matches a 7700k...

 

[JustReason]

And still certain people pretends that current latency issues are only temporal 'bugs' and that BIOS updates, scheduler updates, <insert-any-other-lame-excuse-here> will fix everything soon. That same people has been pretending that mainstream reviews are all wrong or biased and then providing us links to ridiculous and unprofessional reviews and youtube videos were Ryzen matches a 7700k...

Even most of all reliable reviews point to higher resolution gaming will net parity amongst most all chips. They even comment that in those situations the FX8370 shows parity. So again it is far more likely you are obfuscating the issue for your own means than being the voice of reason. Of course the general consensus is that these "fixes" are more like patches and not likely to overcome all circumstances of issues, but again when using context it probably doesn't matter a great deal. I think in these type of cases they will in fact fix the extreme issues but an underlying limit will remain. Even looking at some of the graphs the issue wasn't as huge as some are pretending it is, hence why you see so many already with the CPU speaking so highly to its real world scenario as being Great not just ok, many coming from Haswell Intel CPUs.

Like I have stated before it is the likes of you and Shintai that create ( extreme emphasis on CREATE) these issues well outside their original context. AMD shows high resolution gaming and the performance of such a setup. Why? Well from a marketing stand point we know it will level the field and give a far better showing than what can be had with 720p, as we have seen with reviews. Also considering AMD was far more likely after those that would like 8c/16t but cant afford or refuse to pay the astronomical prices Intel charges than the standard 4c/8t purchaser, then they resolution testing makes even more sense. Most rational beings expected moderate performance and instead got good performance just shy of great. Of course with any new arch and release there will be issues, some fixable some not till respin/second release ie: ZEN+. The real question is when and to what degree does that issue become a concern. Seems thus far for gamers in general it does not. A lot of the issue will be decreased when higher ram speeds are available to be used with the current crop of MoBos, being the fabric is tied to it.

Also what might help you a bit is word usage, also known as connotation. Maybe if you went with perspective and used far less negative connotation you would get far better response when others read you posts. It is hard to trust the words of one that appears to have disdain for the product and therefore hard to trust they are giving accurate and rational information within the context of reason.

 

[Osjur]

From Ryzen: Strictly Technical thread @ Anand:





Clearly shows that there is a performance penalty when threads are on different CCX. Second video shows that Windows scheduler just assigns the threads "randomly" and why we are seeing such a varied results from game benchmarks.

I'm 100% sure this doesn't happen on intel 6-8 core processors because they don't have this CCX interconnect bw limitation as they are one monolithic die. It's no wonder we are seeing quite big gains with faster ram on ryzen because this "Infinity Fabric" (HyperTransport with a fancy name) which communicates between CCX is tied to ram speed.

The real question is that can it be fixed so that windows scheduler tries to keep threads which have dependency on each other on the same CCX. Maybe just make it behave like NUMA

 

[sover]

From Ryzen: Strictly Technical thread @ Anand:





Clearly shows that there is a performance penalty when threads are on different CCX. Second video shows that Windows scheduler just assigns the threads "randomly" and why we are seeing such a varied results from game benchmarks.

I'm 100% sure this doesn't happen on intel 6-8 core processors because they don't have this CCX interconnect bw limitation as they are one monolithic die. It's no wonder we are seeing quite big gains with faster ram on ryzen because this data fabric (HyperTransport with a fancy name) which communicates between CCX is tied to ram speed.

The real question is that can it be fixed so that windows scheduler tries to keep threads which have dependency on each other on the same CCX. Maybe just make it behave like NUMA

That's not an insignificant difference in performance, just having two threads split between CCXs created a 15-20% performance decline. Makes me wonder if it becomes exacerbated as the thread count increases (might account for the massive disparities in some strategy type games that seem like they should benefit the most from having so many threads available). Having a change in the windows scheduler would help, but it seems that AMD is probably right when they say that developers will also need to code to this to make windows aware of which threads are safe to split and which should always be grouped.

 

[juanrga]

Even most of all reliable reviews point to higher resolution gaming will net parity amongst most all chips. They even comment that in those situations the FX8370 shows parity

You don't need to say me stuff I know and have mentioned before. In post #687 I gave a BF1 bench showing a FX-8370 on pair with a R7-1800X at 4K. In the same post I demonstrated how parity disappear when the GPU bottleneck is removed, and I have explained in multiple other posts what this mean for RyZen.

Like I have stated before it is the likes of you and Shintai that create ( extreme emphasis on CREATE) these issues well outside their original context. AMD shows high resolution gaming and the performance of such a setup. Why? Well from a marketing stand point we know it will level the field and give a far better showing than what can be had with 720p, as we have seen with reviews. Also considering AMD was far more likely after those that would like 8c/16t but cant afford or refuse to pay the astronomical prices Intel charges than the standard 4c/8t purchaser, then they resolution testing makes even more sense. Most rational beings expected moderate performance and instead got good performance just shy of great. Of course with any new arch and release there will be issues, some fixable some not till respin/second release ie: ZEN+. The real question is when and to what degree does that issue become a concern. Seems thus far for gamers in general it does not. A lot of the issue will be decreased when higher ram speeds are available to be used with the current crop of MoBos, being the fabric is tied to it.

Also what might help you a bit is word usage, also known as connotation. Maybe if you went with perspective and used far less negative connotation you would get far better response when others read you posts. It is hard to trust the words of one that appears to have disdain for the product and therefore hard to trust they are giving accurate and rational information within the context of reason.

Before launch we had official claims, demos, and leaks promising us that AMD would be a beast on gaming. On launch an overwhelmed consensus was achieved among reviews:

Ars: "an excellent workstation CPU, but it doesn't game as hard as we hoped."

PCWorld: "multithreaded monster with one glaring weakness"

PCGamer: "plenty of power, but underwhelming gaming performance"

Extremetech: "amazing workstation chip with a 1080p gaming Achilles heel"

...

Instead admitting that "Ryzen Doesn't Quite Live Up To The Hype", a vocal group of people in forums rejected the conclusions of reviewers, did make unfounded claims of bias, started to submit links to alternative 'reviews' and youtube videos (from people that pretends that 99% load on a GPU is not a bottleneck), started to post lame excuses about (non-existent) bugs affecting RyZen performance, and accompanied those excuses with promises about how RyZen would start to game once the "bugs" are fixed. First a "BIOS update" was going to fix everything. Then it changed to a "scheduler patch" will fix everything. Now the last-minute try coming from overclocker forums is that faster memory fixes gaming performance... Their 'proof' consists on giving GB3 scores and ARMA III scores for RyZen using 3400 memory. What they don't mention is that non-RyZen chips also see huge performance gains in the same benches when using faster memory.

Meanwhile the Ryzen issue has been identified (even AMD agrees). It is a fundamental design flaw and it will be not solved until Zen+, if it is solved finally.

I said before launch that Ryzen has a latency problem and it was going to affect games. People in forums and twitter attacked and insulted me.

Reviews confirmed the existence of the problem and how it is affecting games. The same people attacked and insulted the reviewers,

Reviewers have continued analyzing this issue. PcPer has confirmed no scheduler will solve this. If you check the comment section of their article they are receiving insults and attacks: "You're literally showing to everyone around the world your true face, an ugly mug of a completely biased and undoubtedly bribed Intel shekel mongler."

Anyone that doesn't say what they want to heard is being attacked and/or insulted.