AMD 'Bulldozer' gets an Update from Microsoft

[Gikaseixas]

Today Windows updater may have brought "Bulldozer" users a little surprise. A hotfix that increases the AMD flagship processors performance. As this "hotfix" is bleeding edge news any benchmarks have yet to be seen but this confirms Windows 7 was in fact hampering “Bulldozer” from performing at 100% in all prior benches. What percentage it was previously performing at has yet to be determined. Here is a small snippet from the Hotfix release notes.
This article introduces an update that optimizes the performance of AMD Bulldozer CPUs that are used by Windows 7-based or Windows Server 2008 R2-based computers. Currently, the performance of AMD Bulldozer CPUs is slower than expected. This behavior occurs because the threading logic in Windows 7 and in Windows Server 2008 R2 is not optimized to use the Simultaneous Multithreading (SMT) scheduling feature. This feature was introduced in the Bulldozer family of AMD CPUs.

What are your expectations? Will it bring 5%, 10% boost overall???

Link: http://www.techpowerup.com/156844/AMD-Bulldozer-gets-an-Update-from-Microsoft..html

Download it here: http://support.microsoft.com/kb/2592546/

 

[Arachnotronic]

Interesting...can't wait to see the benches.

 

[eternalone]

Hmmmm could it be a comeback??

 

[LoneNinja]

http://www.tomshardware.com/reviews/fx-8150-zambezi-bulldozer-990fx,3043-23.html

Toms did a preview on Win8 developer, only 2 applications tested but performance did improve over Win7. It seems they actually put more threads on the same module when possible so other modules and deactivate allowing higher turbo speeds. If the updates to Win7 scheduler function the same way, I don't expect to see much if any improvement for those who disable turbo core and overclock. It is interested FX gets slightly improved performance in Win8 while Sandy Bridge lost performance.

 

[Kenmitch]

Hmmmm could it be a comeback??

Guess will have to wait for the couple of members who bought one to report back on it.

As for a guess on performance ???

 

[Cerb]

Hmmmm could it be a comeback??

To beat a dead horse, no. A remedy, which I would hope improves FPS varaibility seen in some games. But no, the comeback is going to need improved CPUs.

 

[Ferzerp]

Yeah, there are competing threading needs with this poor processor.

If you load up on fewer modules, you can clock higher, but you suffer a pretty big penalty from the shared resources.

If you load up more modules keeping half of each module unloaded, you clock slower, but get better performance due to fewer issues with resource sharing.

You can't have it both ways, but from some of the benchmarks we saw, it's better to load half of each module than it is to try to stack on the fewest number of modules... so....?

 

[podspi]

Ouch, they call it SMT...

 

[tweakboy]

Great OP, wow, now we have software hotfixes that speeds up our CPU's. lololol

out of topic,
How much is a 8 core bulldozer ? If its 300 or less then its a good deal and AMD is doing good job competing with Intel. Sandy is sorta a 8 core chip itself, with HT ,,, and the right app , you have 8 cores,,, soo I think bull and sandy are about equal alto I know in benchies the 2600k or 2700k reign supreme.. thx gl

 

[pelov]

I really hope AT does benchmark this and provide a snippet of other benchmarks for comparisons sake. I wonder if AMD/MSFT chose to throw <4 threads onto separate modules or if they're going with a more complex approach that also takes into account the task rather than just "a thread is a thread."

If they split the threads to separate modules when under 4 then you will see a decent improvement over the current random nature they're assigned. Of course this approach isn't ideal for anything with more than 4 threads.

Will be interesting, but I'd suspect that we'll see some benchmarks decrease and others increase by a decent portion (10-20&#37. Overall it'll probably be a mediocre (1-10%) increase. I'm just taking a shot in the dark based on benchies I've seen with win8 and where thread affinity was manually optimized.

 

[superccs]

Would this hotfix do anything for people with Thuban or Sandy cpus?

I regularly see 10-15&#37; core activity across all my PIIX6 cores during during browsing and listening to music, which keeps the core states active when they should be parked.

 

[soccerballtux]

Guess will have to wait for the couple of members who bought one to report back on it.

As for a guess on performance ???

what members that bought it. Lol.

performance -- 10-15% when running 4 or fewer threads.

 

[Gikaseixas]

Would this hotfix do anything for people with Thuban or Sandy cpus?

don't think so since the architecture are very different

 

[Rvenger]

don't think so since the architecture are very different

I hope it doesn't hinder the performance of Sandy... then Intel will be up Microsofts you know what

 

[sm625]

I dont expect more than 8&#37; improvement. A drop in the bucket really. The problem is that we need a scheduler that puts one thread onto each of the 4 modules before loading a 5th simultaneous thread onto any module. Then you'll get much better performance, especially if you are overclocking with turbo disabled.

 

[notty22]

As this "hotfix" is bleeding edge news any benchmarks have yet to be seen but this confirms Windows 7 was in fact hampering &#8220;Bulldozer&#8221; from performing at 100&#37; in all prior benches.

I can't agree with that logic, the writer there arrived at. It's juicy sounding though. Widows 7 would be what AMD engineers tested their product with ? Among other software.

 

[Chiropteran]

I can't agree with that logic, the writer there arrived at. It's juicy sounding though. Widows 7 would be what AMD engineers tested their product with ? Among other software.

What does that have to do with the price of tea in china?

If Windows 7 scheduler is causing bulldozer to perform worse than it otherwise could perform, it is hampering bulldozer. Whether or not windows 7 was used for testing during the development of bulldozer is completely irrelevant.

 

[gevorg]

So is it finally going to beat Phenom or what?

 

[notty22]

Only if you are using windows 7. They will have to put that requirement on the box.
All other versions of Windows will hamper bulldozer performance.

 

[Rvenger]

So is it finally going to beat Phenom or what?

Thats what we are at least hoping for.:hmm:

 

[Edrick]

If Windows 7 scheduler is causing bulldozer to perform worse than it otherwise could perform, it is hampering bulldozer. Whether or not windows 7 was used for testing during the development of bulldozer is completely irrelevant.

So AMD makes a processor that doesnt perform at 100&#37; on the most popular OS out there? And that seems ok to you?

 

[bryanW1995]

No, no, aren't you reading? After being 2 yrs late, it only took another 6 months for them to achieve a ginormous 5-10&#37; increase in performance through a software update that they probably had to write themselves. In fact, this update is so super awesome that BD is now actually possibly going to be a tiny bit faster per clock than a 3 yr old phenom II! Good times, good times.

 

[frostedflakes]

Looks like the patch makes BD show up as 4C/8T in Windows now instead of 8C/8T. The marketing guys at AMD just cringed.

 

[skipsneeky2]

Is this miracle of a update gonna drop that out of this world power consumption too?

Nevermind performance how about power consumption .

Maybe microsoft can ship the update with the processors and send everyone also a free phase cooling unit LOL.

Maybe i'll buy into the bs and i will buy one of those magical intel cards that updates your processor specs maybe for $50 more i could turn my i3 2100 into a i7 2600k and see what sounds more stupid the ms update or the intel card LOL.

The damage has already been done it still needs a second revision you don't see our intel processors and phenom 2 processors requiring special microsoft updates.

 

[Jovec]

you don't see our intel processors and phenom 2 processors requiring special microsoft updates.

Umm, Intel did need "special microsoft updates" when they released their original HT CPUs.

The only issue here is why AMD didn't work with Microsoft to get this released prior BD's launch. The scheduler is one of the fundamental functions of a modern OS, and while the changes needed for BD should have been relatively minor, AMD should have known that MS wasn't going to ship a hotfix to the scheduler without serious validation.