Read more.Quote:
Microsoft lets slip with premature release of hotfix for Windows 7 and Server 2008 R2.
Printable View
Read more.Quote:
Microsoft lets slip with premature release of hotfix for Windows 7 and Server 2008 R2.
2-7% of "free" CPU performance from a hotfix is nothing major? 0_oQuote:
initial suggestions claim a two to seven per cent increase in performance, nothing too major
Hexus not easily impressed today it would seem!!
Well 2% certainly isn't anything to write home about.. 7% a little more worthy but it's unlikely to change anyone's buying decision. It's only a minute off every 14 minutes of x task.
i think the main thing is, the hotfix was half of one so having a 2% - 7% increase is significant as having a full hotfix means it could possibly hitting around 15% we have no idea but any increase is a good increase as BD isnt THAT bad anyways and a nice cashback scheme there putting the fx8120 £2 cheaper than a i5 2320!
How long has Bulldozer been in development? Definitely a long time and much longer than it should have been, yet despite extra time nobody at AMD thought to contact Microsoft and work with them to do a hotfix to be ready at launch, hell it probably could have been in Windows 7/2008R2 SP1 if they'd planned ahead. And if AMD didn't even think of this until someone pointed it out after launch then that's just plain pathetic, didn't they do any testing/validation?
I doubt they have contacted Linux kernel devs either then? What a bloomin shambles. If you're making a new processor architecture that's mildly disagreeable with current OS then you need to do something about it BEFORE launch! :wallbash:
Ask MS for an explanation.....they managed to get the code in Windows 8 early enough!
Short memory,eh??
HT took a while to be supported properly under Windows XP. In fact it has taken until Windows 7 for HT to have optimal support from MS. Native dual cores also caused Windows XP to have a hissy fit too. Even the Pentium Pro was not optimally supported by Windows 95. Everytime there is a big change to CPUs,MS has taken time to patch Windows or expects you to get the next version.
Look at Windows XP 64 bit - Vista 64 bit and Windows 7 64 bit were far better. It took three years for MS to introduce a relatively solid 64 bit OS with Vista 64 bit.
AFAIK,Linux kernal patches are already out. This is not surprising as Linux tends to be quicker to accommdate such changes unlike the lumbering brute which is MS.
even at the best XP64 was aweful supported (i know i had it , still got the dvd somewhere)
the patch was released last week - some sites actually snagged it before MS pulled it - and it showed a DROP in performance, so they needed to rewrite it:
some numbers;
http://i1-news.softpedia-static.com/...sappoint-3.jpg
http://www.sweclockers.com/artikel/1...re-i-windows-7
http://ht4u.net/news/24857_patch_sol..._enttaeuschen/
hotfixed ID`d as
KB2592546
now taken down though
Microsoft aren't exactly in AMD's pockets though, are they? It's only been 2 months since launch, and I doubt this was a trivial change to scheduling. I also doubt Microsoft were willing to start working on a scheduler patch without proof that it would be beneficial, i.e. they would want to test on something approaching shipping silicon to ensure that the time developing a new scheduler would be well spent.
OTOH, Windows 8 is still some distance before pre-beta, so patching in a slight modification to its scheduler during development wouldn't be seen as such a big issue. So I don't really see that it's much of a surprise that the developer preview of Windows 8 had a more bulldozer-optimised scheduler.
I was thinking that, I've seen less being praised for expensive memory upgrades. Really isn't bad for nothing...
I think the Linux kernel patches were out before Bulldozer was even released TBH. And yes, features do tend to make it to Linux way before Windows - look at EFI/GPT for another recent example.
It's out already. And http://forums.hexus.net/cpus/200143-...ml#post2266377