Bulldozer performance hotfix for windows

[jonnyp11]

Awaiting test results...

if it got pulled for decreasing performance in some cases, i don't think banches really matter unless it improve one or 2 and gives us a peak, but i doubt that. i sure hope it is an easy thing to remove otherwise i'm sure many people will be/are pissed

Don't they test this stuff before software patches are released? AMD cannot seem to get Zambezi right thus far. Hopefully they can work something out.

Also this incident was all on microsoft. the article i linked said that amd said they and intel were both releasing the patches Q'1 '12, and they didn't know why intel had released their side yet and it would not yield any improvements without the rest of the patch.

 

[alanalian]

Ive heard of nearly 40% increase in performance. Is that true

No. The low performance of AMD is not because of the windows' problem but the software don't support CMT
The changing of 8C8T into 4C8T could not solve problems unless the software support CMT

 

[linkin]

No. The low performance of AMD is not because of the windows' problem but the software don't support CMT
The changing of 8C8T into 4C8T could not solve problems unless the software support CMT

Well let's hope microsoft give us a nice little christmas present then

 

[Aastii]

Well let's hope microsoft give us a nice little christmas present then

 

[StrangleHold]

No. The low performance of AMD is not because of the windows' problem but the software don't support CMT
The changing of 8C8T into 4C8T could not solve problems unless the software support CMT

Wrong. Of course the software has to support it. Since AMD has Modules instead of full cores the OS doesnt know which core is the best to run each thread on. I could go on with more details but most here already knows what the deal is.