AMD Ryzen 7 9700X

[THU31]

Even more details above. Intel makes the scheduler in the CPU, collaborates with MS to help them make their part work better with their CPUs. They don't mention to even touch the OS though.

“Early on in Windows 11 development, Intel approached us with a proposal for an interface that would allow the CPU to guide the operating system scheduler by providing information about how much a specific workload would benefit from being scheduled to a performance core instead of an efficiency core.”
"According to him, the proposal kicked off a deep collaboration during the early phases of Windows 11’s development, incorporating Thread Director feedback right into the thread scheduler."

It seems like AMD forgot about the collaboration part. I rest my case.

 

[Zach_01]

I wonder if this windows "optimization" was held in the bag just for the right time....

 

[HD64G]

“Early on in Windows 11 development, Intel approached us with a proposal for an interface that would allow the CPU to guide the operating system scheduler by providing information about how much a specific workload would benefit from being scheduled to a performance core instead of an efficiency core.”
"According to him, the proposal kicked off a deep collaboration during the early phases of Windows 11’s development, incorporating Thread Director feedback right into the thread scheduler."

It seems like AMD forgot about the collaboration part. I rest my case.

You read but do not understand and there is the cause of that discussion going on for so long: " incorporating Thread Director feedback right into the thread scheduler" means that the one is made by Intel, the other by MS and they communicated to help the software side (MS) get the best performance out of the new and harder to program hybrid arch (Intel). Where it says that Intel messed with the OS there?

Another argument: AMD didn't know any such performance hidrance of their CPUs from windows until MS made it work better a few months ago with the insider version of 24H2. That fix wasn't the result of any collaboration. Until MS speaks publicly about how they found out and why it wasn't found for years now, how anyone can blame AMD since they couldn't know the existence of this problem until MS solved it? WIth what could they compare the performance to discover it? And their arch wasn't any difficult to work with from Zen 1st gen. Only the 12 and 16-core had 2 CCDs that caused latency problems and that showed in specific apps.

Prove me wrong but there is no chance that a bad optimized code that exists for years in a closed-kernel windows scheduler could be found from engineers that are outside the os company and not from the ones onwing that code and kernel. So, by your logic alone, if AMD has a responsiblity for that performance problem in windows, MS has at least 10X that responsiblity.

 

[529th]

I wonder if this windows "optimization" was held in the bag just for the right time....

Yeah like some Lisa Su Kung Fu

Onna fo real

 

[THU31]

You read but do not understand and there is the cause of that discussion going on for so long: " incorporating Thread Director feedback right into the thread scheduler" means that the one is made by Intel, the other by MS and they communicated to help the software side (MS) get the best performance out of the new and harder to program hybrid arch (Intel). Where it says that Intel messed with the OS there?

But did you understand anything I said? Not once did I mention anything about messing with the OS. All you keep saying over and over is about "touching the OS".

I will say it for the 1500th time - hardware maker is supposed to provide all the necessary information, the OS maker is supposed to implement it.

I'm done. Unwatching this thread now.

 

[chrcoluk]

https://www.intel.com/content/www/us/en/support/articles/000091284/processors.html

Official word from Intel says that the AL thread director is embedded in the CPU. So, this is a thing MS has nothing to do with.

https://www.digitaltrends.com/computing/how-intel-thread-director-marries-alder-lake-windows-11/

Even more details above. Intel makes the scheduler in the CPU, collaborates with MS to help them make their part work better with their CPUs. They don't mention to even touch the OS though.

And why MS needed to reverse engineer anything? No change in the basic cpu arch happened since Zen3. They are supposed to try improving performance by altering parameters of the OS in the way the CPU threads work with the software's demands. Just needs some time and effort. They had years of time for sure since Zen3. Did they try hard though? If they did, the increase in performance just by one update wouldn't be as huge as this!

Ok but it still required them to implement recognition of the thread director in the thread scheduler and the article clearly states Intel provided the input for this.

 

[HD64G]

Ok but it still required them to implement recognition of the thread director in the thread scheduler and the article clearly states Intel provided the input for this.

Input means results of testing and suggestions for further improvement. Agreed.

But did you understand anything I said? Not once did I mention anything about messing with the OS. All you keep saying over and over is about "touching the OS".

I will say it for the 1500th time - hardware maker is supposed to provide all the necessary information, the OS maker is supposed to implement it.

I'm done. Unwatching this thread now.

If so, how can a CPU manufacturer know if the OS is optimized enough for their hardware in order to help improve it further down the road?

You are out of solid arguments since the start and keep saying about AMD being most responsible without once mentioning the MS being at fault.

 

[chrcoluk]

This will be my final reply to you, as I feel its hard to get my point across.

The original point I replied to suggested some kind of theory that Microsoft is solely responsible for implementing hardware technical features into its software, but my point is this is not how things work, there is no motivation from Microsoft to hinder how well hardware works on its software.

When it comes to hardware and the way it interacts with the operating system, there is collaboration between vendors and the software developers of the operating system. That could be in the form of a direct contribution of code or drivers, or it could be technical guidance on changes that should be made to allow the OS to interact properly with the hardware. There is is some form of collaboration. Intel's input would have been more than testing and suggestions, it would have at the very least been technical guidance, likewise it will be the same for AMD, and we have now seen Microsoft move quickly to actually fast track the updates into the current release build of Windows 11.

 

[rcarlos243]

I just read the Ryzen 7 7700 review and there is a result with "PBO Max", but it does not have any explanation on what it is.

This review also has "PBO Maximum", but it has no explanation on what it is.

Can someone clarify what does "PBO Max" mean?
Is it just going into BIOS and setting PBO to enable?

 

[95Viper]

Understanding Precision Boost Overdrive in Three Easy Steps

More info that may be of help

 

[Zach_01]

I just read the Ryzen 7 7700 review and there is a result with "PBO Max", but it does not have any explanation on what it is.

This review also has "PBO Maximum", but it has no explanation on what it is.

Can someone clarify what does "PBO Max" mean?
Is it just going into BIOS and setting PBO to enable?

Most likely it was PBO enabled with high limits.
If you just enable PBO or set it to advanced with board limits its practically unlimited.
In this situation the limiting factor(s) will be the CPU operating temperature limit if board VRM can supply the power.

I'm not sure if they use curve optimizer.

 

[HD64G]

New data prove the #11 being Wintel's mess of an OS hurting AMD CPU's performance while win10 is fine in most occations.

 

[AVATARAT]

New data prove the #11 being Wintel's mess of an OS hurting AMD CPU's performance while win10 is fine in most occations.

It was expected.

 

[TumbleGeorge]

Must be retested again with latest agesa and lower between chiplet latencies because problem was fixed.

 

[HD64G]

 

[Ayhamb99]

The video demonstrates that while the gains from 23H2 to 24H2 for Intel were not as much as AMD had, there was still improvement. Hopefully the conspiracy theorists now calm down and understand that it was Microsoft's incompetence that was to blame for W11 performance gimping and not out of malice towards AMD, since Intel was also being performance gimped. Any chance there's a video of 23H2 vs 24H2 but with applications performance being tested? Or was this whole 23H2/24H2 ordeal only affecting gaming?

Still would choose the 9700X anytime of the day though over the 14700K. Intel has a lot of confidence/trust to regain due to how they blundered handling the 13th/14th gen burnouts. If they just had handled the situation better and not act so defensive and reluctant, they wouldn't be this despised now.

 

[Footman]

Not sure if this has already been spoken about, but any thoughts on why the 9600x is faster in gaming at lower resolution than the 9700x?