[Don Karnage]

Quote:

We had already noticed that the WinRAR benchmark, but also the application itself, does not benefit from Hyper-Threading Technology from Intel. In fact we had the same values found on the CPU 2600K, both with 8 logical cores and both with 4 logical cores.
Now we will try to shed light on the problems and secrets of the core parking of Windows Seven.

Source

http://www.xtremehardware.it/eng-reviews/eng-reviews/core-parking-on-windows-seven-winrar-performance-with-sandy-bridge--201111226092/

 

[Homeles]

Perhaps this is why there's a performance hit occasionally associated with HT?

 

[tpi2007]

Interesting! Thanks for the link.

It will be interesting to see if Microsoft will be updating the scheduler to also address this problem. Since they are apparently going to patch Windows 7's scheduler to better handle Bulldozer, it makes sense to think so (I hope).

 

[Homeles]

They fix core parking in Win8.

 

[djriful]

You can disable Core Parking yourself here:

http://forum.cakewalk.com/tm.aspx?m=1861804

I'm going to give a shot on my AMD CPU.

 

[importflip]

Quote:

Originally Posted by djriful

You can disable Core Parking yourself here:

http://forum.cakewalk.com/tm.aspx?m=1861804

I'm going to give a shot on my AMD CPU.

I thought this doesn't affect AMD CPUs. At least that's what I gathered from the source link.

 

[Vhox]

Surprised this is news to people >_<. More knowledge the better I suppose!

 

[Starbomba]

I'm wondering why they did not do it with i3 CPU's. Are they too "weak" already to be affected by the setting, or increase is not as significant?

Will try this on my own i3, it was a nice read

 

[djriful]

Quote:

Originally Posted by importflip

Quote:

Originally Posted by djriful

You can disable Core Parking yourself here:

http://forum.cakewalk.com/tm.aspx?m=1861804

I'm going to give a shot on my AMD CPU.

I thought this doesn't affect AMD CPUs. At least that's what I gathered from the source link.

It's base on Core.. not brands...
This is the OS software windows energy power saving control. You're simply forcing the windows not to sleep any cores regardless what brands and etc.

This affects more on Intel side due to HT.

 

[Brutuz]

Quote:

Originally Posted by Homeles

Perhaps this is why there's a performance hit occasionally associated with HT?

No, there is a performance hit in some applications that fill up the CPUs pipeline, HT ends up slowing them down because it's trying to fit more in the pipeline.
Quote:

Originally Posted by importflip

I thought this doesn't affect AMD CPUs. At least that's what I gathered from the source link.

They didn't try it on AMD at all?

 

[MadGoat]

Quote:

We verified that with CPU without Hyper Threading or with Hyper-Threading disabled, the "failure" does not occur.
It does not occur even with AMD Bulldozer, which could be affected by a similar problem because of the modules architecture. The hypothesis we did is that Windows Seven still did not use the core parking on bulldozers CPU. We will investigate this problem as soon as possible in a future article.

Seems to be a intel hyperthreading investigation / issue only...

I personally always see all my cores in use... (Seems AMD need not apply here)

 

[G3RG]

I wish they did this more indepth...
why only test winrar? need more benchs with more cpus

 

[DuckieHo]

Quote:

Originally Posted by Brutuz

No, there is a performance hit in some applications that fill up the CPUs pipeline, HT ends up slowing them down because it's trying to fit more in the pipeline.

There's some overhead to manage the additional threads as well.

 

[Rookie1337]

According to that link the problem exists with anything beyond 2 cores...but that was 2 years ago. Have they updated something in Windows since then to address the problem and how do you know if you have a "parking" issue?

 

[EfemaN]

Would this show any improvement in applications that aren't heavily threaded like WinRAR? And does this mess with any powersaving features like speedstep, c-states, etc.?

 

[Woundman]

Edit: Original source was incorrect. Updated post with correct information.
Quote:

How to Disable Core Parking (the correct way)

1) Press the Windows + R keys > type regedit

2) Press CTRL + F keys > copy / paste / search " 0cc5b647-c1df-4637-891a-dec35c31858 " (no quotes). Look for data entries with Attribute, ValueMin, ValueMax, ValueIncrement.

(NOTE/REQUIRED) If you modified ValueMax to 0, change it back to 64 hex(100 decimal) for all the 0cc5b647-c1df-4637-891a-dec35c31858 search results. Otherwise, ignore this and continue with the rest of the guide.


3) Right-click and modify all Attribute data entries found to 0. (I found two or three Attribute data entries searching for " 0cc5b647-c1df-4637-891a-dec35c31858 " (no quotes))

5) Press the Windows + R keys > type powercfg.cpl

7) Click "Change plan settings"

8) Click "Change advanced power settings"

9) Expand "Processor performance core parking min cores". Change value to 100%.

10) Restart

11) Press the Windows + R keys > type resmon.exe

12) Click the CPU tab and ensure your cores aren't parked.


Enjoy.

Edit: Clarity

 

[GanjaSMK]

Hrm I suppose I'm no so savvy with core parking. Does this somehow help AMD systems, is there data behind it, relevance, or does this only affect CPU's with CMT/SMT?

 

[Strat79]

Quote:

Originally Posted by GanjaSMK

Hrm I suppose I'm no so savvy with core parking. Does this somehow help AMD systems, is there data behind it, relevance, or does this only affect CPU's with CMT/SMT?

Apparently it is an HT only thing as far as we know. At least that is what was said earlier in some quote. Non HT or HT disabled CPU's are not being affected by parking.

 

[Z Overlord]

So wait it my i7 920 losing performance in games due to hyper threading? I'm too tired to read

 

[Blameless]

EDIT: Testing the alternate method Woundman posted now....

EDIT 2: No difference between 10% and 100% on "processor performance core parking min cores" using the alternate method.

I cannot replicate this issue on any of my Hyperthreaded systems.

I currently have:

i5 430M (2C/4T)
i7 920 (4C/8T)
Xeon W3540 (4C/8T)
i7 2600k (4C/8T)
i7 970 (6C/8T)

The Xeon and i7 970 are running Server 2008 R2 SP1, all the other are running Windows 7 SP1. Power profile is high performance on all of them, and min processor state (while plugged in) is set to 100% in the control panel power options.

None of these systems show any hint of core parking, and the supposed fix for it does not produce any changes from the default value.
Quote:

Originally Posted by Brutuz

No, there is a performance hit in some applications that fill up the CPUs pipeline, HT ends up slowing them down because it's trying to fit more in the pipeline.
They didn't try it on AMD at all?

The original Czech guys did.
Quote:

Originally Posted by G3RG

I wish they did this more indepth...
why only test winrar? need more benchs with more cpus

Because WinRAR was the only one that showed any real difference.

http://www.diit.cz/clanek/analyzing-core-parking-in-windows-7-and-windows-8-in-relation-to-amd-bulldozer-performance-in-english-winrar-vs-core-parking-vs-core-i7-in-windows-7-and-windows-8/39100/37375
Quote:

Originally Posted by Z Overlord

So wait it my i7 920 losing performance in games due to hyper threading? I'm too tired to read

No.