AMD hot fix just boosted my sandy in cinebench

[formulav8]

Hopefully Anand does a nice review of the patch, as he did with the TLB patch and eventual stepping fix on Phenom.

Well don't hold your breath for him to do much of anything AMD. :whiste:

But hopefully he will.

 

[Phynaz]

What do you mean "official" there are people on this forum which produce more accurate and authentic results than any hardware site, AT included.

What I mean by official is MS stating this does something for Intel chips.

A few random posters on the net doesn't mean it does anything. As far as I'm aware, according to MS this patch only works on BD chips. Which would indicate they ID the CPU.

 

[Diceman2037]

What I mean by official is MS stating this does something for Intel chips.

A few random posters on the net doesn't mean it does anything. As far as I'm aware, according to MS this patch only works on BD chips. Which would indicate they ID the CPU.

This is a backport of the windows 8 scheduler, changes that affect all processors were likely left in place.

 

[tweakboy]

What about me? Can use this hotfix,

BTW I can't find the hot fix onlien anywhere, google all sites micrsofot seems to have removed the hotfix.


If this hotfix can give me 3 more fps a second Ill do it.

Can I have a working link for this, or you guys things its waste of time for me.

thx gb gl

 

[grkM3]

here ya go guys,this is the original file that MS put up and then took down with an updated 2 file install.

use at your own risk and this is for windows 7 64bit

 

[tweakboy]

Wow thanks big time, grkM3 .. thank you my brother.. Oh wait since they took it off ,,,,, its a really big risk uh ? thx gl

 

[grkM3]

Wow thanks big time, grkM3 .. thank you my brother.. Oh wait since they took it off ,,,,, its a really big risk uh ? thx gl

thats the file I have been running for a little over a months now.

before you install run cinebench and post before and after gains.

edit I saw the gains on my 2600k,I dont know what it will do with other intel chips.I thinks its just a better way for the OS to handle more threads like others have said and it will install no problem without an amd cpu installed

 

[Kenmitch]

http://www.mediafire.com/?y3ba4xcs50w6ki7

I tried this above patch for my somewhat unscientific testing.

thats the file I have been running for a little over a months now.

before you install run cinebench and post before and after gains.

edit I saw the gains on my 2600k,I dont know what it will do with other intel chips.I thinks its just a better way for the OS to handle more threads like others have said and it will install no problem without an amd cpu installed

It does seem to have some effect on my 2700k. My board in sig is in the rma process with Asus. I'm currently using a ASRock P67 Ext4 Gen3 I picked up as a temp board with a 2500k for future testing....Microcenter bundle

I'm still trying to dial in my 2700k at 4.8ghz as this uEFI is primitive compared to the Asus implementation....It's doable but takes some more tweaking to get there. Taking some effort as I'm testing with Intel Burn Test using the (ALL) threads setting which HT doesn't like I hear.

Initial testing shows the GFlop output is alot more consistent. The GFlop output did somewhat scale a little better the more memory I used. The system seems more responsive when IBT is running and I try to surf the web or view screenshots of GFlop output for referance.

The downside is I need to tweak this overclock a little more and got a bsod on the 10th run of 6000mb of IBT. I was using the default of 1024,2048,4096, and picked 6000mb custom to keep the results more consistent. Didn't get the after screenshot

I'm gonna go for another round without tweaking any settings and start with 6000mb as it was most likely a random bsod due to HT and using ALL cores instead of changing to 4 threads like most suggest using IBT with HT chips.

 

[Kenmitch]

here ya go guys,this is the original file that MS put up and then took down with an updated 2 file install.

use at your own risk and this is for windows 7 64bit

http://www.mediafire.com/?bz1ypn53n4lf6x1

??? is linked to 7-Zip file manager installer. Wrong link or is it something I did wrong?

 

[grkM3]

??? is linked to 7-Zip file manager installer. Wrong link or is it something I did wrong?

IM 99% sure this is the right file.

I will install it on my other os and before you install this patch run cinebench with 8 threads and then run it after the patch to see if it gained performance

EDIT sorry worng file.

this is the hotfix

http://www.mediafire.com/?y3ba4xcs50w6ki7

 

[Diceman2037]

here ya go guys,this is the original file that MS put up and then took down with an updated 2 file install.

use at your own risk and this is for windows 7 64bit

you can use the 1st of the 2 file fix, it does the same thing except it also includes a newer version of powercfg

 

[exar333]

Toms did a write-up comparing the 'gains' from the patch for the 8150. Pretty-much all within the margin of error, IMHO. 1-2% AT BEST. Definitely not the silver bullet folks were looking for. More AMD marketing lies.

http://www.tomshardware.com/reviews/windows-7-hotfix-bulldozer-performance,3119.html

 

[grkM3]

ok guys I just did a test on a clean windows 7 64 with sp1 install.

system running at 4.8ghz with a 2600k and with 8gb ram at 2133 9-10-10-2t
ran cinbench 3 times with out the patch

first run
9.27

second run
9.32

third run
9.38

installed hotfix that I posted above and changed nothing else in the setting

first pass
9.47

second pass
9.47

third pass
9.47

this hotfix is awesome and windows spreads the load on the cores a lot better and the results are real.

the new hotfix windows put out wont let you install with an intel chip but the one I posted above works and it came right from microsoft.

and to anyone that benches with cinebench going from 9.38 to 9.47 would need to overclock the cpu about 150-200mhz to get that increase in cinebench

 

[red454]

I saw a similar improvement with that patch (Windows6.1-KB2592546-x64).

Not a huge gain, but I'll take it. And thanks grkM3 for posting it...

 

[Diceman2037]

the new hotfix windows put out wont let you install with an intel chip but the one I posted above works and it came right from microsoft.

False

 

[grkM3]

False

I have the hotfix on my desktop right now and it wont let me install

 

[red454]

False

Do you know if the two new hot fixes add even more improvement? Is it cumulative?


Edit: Post 61 sort of answered my question...

 

[Diceman2037]

Do you know if the two new hot fixes add even more improvement? Is it cumulative?


Edit: Post 61 sort of answered my question...

the new patch completely replaces all files from the old patch.

 

[Diceman2037]

I have the hotfix on my desktop right now and it wont let me install

it relies on files in sp1

 

[greenhawk]

What about me? Can use this hotfix,

really? after this previous post?

Any chance you can be consistant in your outlook on things. It has only been a little while between those posts (and both in this thread).

Benchmarks are getting old. A CPU is for productive work. not benchmarking and getting happy or stability testing.....

 

[grkM3]

just did 10.20@5.2ghz

that is almost on par with a 16 thread xeon x5570

 

[Diceman2037]

I have a feeling that the new update that enables Force ALSR in the win 7 kernel will also bring the same performance improvements as the bulldozer kernel hotfix

 

[NTMBK]

I have a feeling that the new update that enables Force ALSR in the win 7 kernel will also bring the same performance improvements as the bulldozer kernel hotfix

Why would a change to how objects are laid out within an application's address space have anything to do with a change to how threads are scheduled? :\

 

[Diceman2037]

why would firefox 12 include the fixes of firefox 11?

the file versions inside (where it matters) are within 3 revisions of the second version of the scheduler hotfix.
the first version of the scheduler patch has an older version string then the force ASLR hotfix

do the math.

 

[Lonbjerg]

ok guys I just did a test on a clean windows 7 64 with sp1 install.

system running at 4.8ghz with a 2600k and with 8gb ram at 2133 9-10-10-2t
ran cinbench 3 times with out the patch

first run
9.27

second run
9.32

third run
9.38

installed hotfix that I posted above and changed nothing else in the setting

first pass
9.47

second pass
9.47

third pass
9.47

this hotfix is awesome and windows spreads the load on the cores a lot better and the results are real.

the new hotfix windows put out wont let you install with an intel chip but the one I posted above works and it came right from microsoft.

and to anyone that benches with cinebench going from 9.38 to 9.47 would need to overclock the cpu about 150-200mhz to get that increase in cinebench

So an "improvement" from ~9.3ish to 9.47 is "awesome"?


lulz