This only happens on the models with more than 8 cores.

Probably isn't as bad as it sounds, it's only a part of the chip functions too. Can't say more before actually looking into it or I could become clickbait myself lol

 

It's just looks like one of the super-detailed conspiracy theories about secret services running that report to Microsoft, or Apple. The components in question made it past R+D, Engineering, QA, and other departments at Intel; only to have this one guy figure it all out for them.

 

but, has this any impact on audio work? (vst etc) or is it just a graphics latency problem

 

Damn, the video got private by the time i could see.

Can Anyone please give a gist? i am currently in the process of building a workstation, and if intel has some core issues, id rather go AMD way.

Thank you

 

I had a similar problem, but it turned out to be a signalling (or polling) problem with the mouse/USB port itself.

I've run latencymon for hours (on 13600KF) with no big spikes in latency.

 

You shouldn't worry about it. I'm running AMD and I know that the video is BS.

Imagine every tech YouTuber and journalist who makes a living from putting those chips to the test completely missing out on a very obvious issue during the countless hours they spent analyzing the performance in many synthetic and real world tests. A batch from a couple of years ago of Lian Li AIO water coolers made the news for a very obscure failure, but for some reason flagship intel chips don't. Yeah, sure, the dude must be onto something.

 

Thanks EddieXx, you beat me to it. This seems the best question yet! What is the impact on audio processing?

 

This video is about Intel Gen 11, 12, 13 and possibly the upcoming 14?
I really don't understand the logic of your reply? You're running AMD, but you know this video about Intel is BS?

 

I explained why in the same reply.

 

That's not an explanation! That's just a subjective observation from your side. You're just assuming it's bogus based on your
prejudices about other video's like this? And knowing you're on AMD I don't think you've even seen the video?
Maybe you're right? Maybe it is bullshit? But I would strongly advise you to stick to 'facts' and not claim conclusions based on just your subjective observations based on other videos.

 

I don't make my living from music, despite being in a music-oriented forum. I do, in fact, work in IT, and I have deployed more Intel servers than I can count. From extremely cheap servers to servers that cost in the thousands monthly. Most of my work is to make things work, and sometimes I have to make things work with almost zero processing headroom. It either works, or it doesn't.

Of course, I don't mean to say that I'm an expert, let's assume here that I know as much as the guy from the video. Latency issues are extremely esoteric and what is true for you may not be true for another guy using a different motherboard, mouse, keyboard, graphics card. I can't go and say that Intel has a widespread problem that no one but me is aware of. This is a very bold claim to make, especially when you show proof of it not working, and said proof shows software commonly benchmarked dozens, if not hundreds of times by reviewers all around the world. If there was a clear issue, people would have noticed. Is the issue a lie, though? Probably not, I hardly believe the dude would make a video lying about it when he doesn't really have nothing to gain from it, but I think it's more of a case of the stars aligning just right in regard to his PC parts. Think about the bigger picture, though: Intel owns ~70% of the CPU market share, we're talking about 3 generations of CPU from the course of ~3 years, so what are the chances of Intel really having a DPC latency issue? Are all people doing professional work on Ryzen or Apple M1/M2 chips? I find it very hard to believe.

 

you're right some batches of hardware may suffer from flaws, which may even appear after longer time,

about youtubers missing obivous issue, well it's often not obvious on fresh systems, I have never ever seen a proper real world benchmark after year of use, because nobody cares about reviews of outdated hardware,
on a fresh system, there's much more headroom for issues to not be seen, for ex. almost-empty SSD will have better write speeds and Windows will not obsess about telemetrying shitton of data present on the disk xD

 

DPC is measurable and comparable, dismissing it as "esoteric" says a lot. Market share is also not a convincing argument. Many people have already overlooked a lot, and only a few have discovered it. For example, all the vulnerabilities that have been found, most people (even the vendors) haven't noticed. So the argumentation is seriously flawed. sry shinjiya
The impact of the vulnerabilities on the performance of Intel CPUs varies depending on the specific vulnerability and the mitigations implemented. For example, the mitigations for the Meltdown and Spectre vulnerabilities can decrease the performance of affected Intel CPUs by up to 39%, while AMD CPUs may lose up to 14% . Similarly, the mitigations for the BHI-variant of Spectre V2 can cause a 35% decrease in performance. The MDS attacks also affect the performance of modern Intel CPUs in servers, desktops, and laptops, including the latest 9th-generation processors, despite their in-silicon mitigations for Meltdown. Ironically, 9th-generation CPUs are more vulnerable to some of the attacks compared to older generation hardware. Intel has released redesigned CPUs and added hardware and firmware mitigations regarding these vulnerabilities to its latest processors.

Intel has had a higher number of publicly disclosed vulnerabilities compared to AMD. However, it is important to consider that the number of vulnerabilities alone may not be the sole factor in determining the overall security of a processor. Both Intel and AMD continuously work to address and mitigate vulnerabilities to ensure the security of their processors.


[1] https://wccftech.com/intel-mitigations-spectre-v2-exploit-35-percent-drop-in-cpu-performance/
[2] https://mdsattacks.com
[3] https://en.wikipedia.org/wiki/Meltdown_(security_vulnerability)

 

Yes, I suspected you work in IT and claim to know all about how computers work and in the process also know how all intel processors are build and exactly know the flaws of each generation. That's obvious when you work in IT and have deployed more servers than you can count? Come on...

There we go again! You're assuming without any proof!

Again! You're assuming! The video clearly stated that it only noticeably happens when dragging content. And also states that it might not impact gaming or other workflows?

The only thing we're after here is to find out if this 'bug' is of any consequence to the workflow of audio/multimedia systems based on these Intel gens?

We're not looking for "I'm an IT professional and I think this video is BS. And btw I'm on AMD".

It doesn't help in any way!

If you want to help, post facts about the matter, not this BS like "I know everything, because I'm in IT"!

 

I finally watched the videos. And I think the latency is there, although I don't know how much or for how much people (hardware combo).
But that latency is basically with the Windows and its Applications GUIs when doing something relatively demanding (drag and drop for instance).

So yes, GUI latency. The other claims, especially the "Intel 12th-13th CPUs have serious problems" guy, seem to be very exaggerated. Surprise, he hid the video.

PS. I downloaded it before so I could see it

 

Cool. Can you upload it for me and maybe others somewhere?
Because I didn't saw.

 

This the full story. Not the actual video that was posted here. But he was just quoting based on this one.

Part 1:

Part 2 is coming I presume?

 

Disclaimer: I'm not a Windows kernel specialist nor am I a driver developer. But I've dealt with these kind of things at work.

So, to add to this discussion while it's a driver issue in the end the kernel has very much to do with this, because we get one DPC queue (two actually since there's also an additional threaded queue per "real" queue) per CPU. As I understand it from the docs*, the driver can use either and it can request a specific CPU to run the DP call on. While we as users can set the preferred CPUs via process affinity, we have no control over how the queues get distributed for drivers. Also processes don't normally have control if the task is going to be run on a "real core" or a thread, which is the case for SMT systems (most of them these days), then there's architectures where we have asymmetric cores, like Intel's P and E cores. This is all handled by the task scheduler which we can't control either. Sure, there's task priority, but that doesn't change a) the DPC queue distribution or drivers, at all.

So it very much dependent on the CPU used. But since the DPC kernel code is the same for all drivers on all architectures I'm guessing it's down to how the drivers are written and the task scheduler for different architectures. Sometimes there's probably a mismatch somewhere which makes this tricky as hell to figure out.

*) Organization of DPC Queues
and
Introduction to Threaded DPCs

 

you're right!
part 2 was shared by me in comments of this thread

 

just a sidenote, im having huge graphics lagging/stuttering since the latest Windows update, Ive never had this problem before. Goes to show how intricate these issues can be. And im not using any of these latest architectures