Still hope for AMD bulldozer!!??

[KING_OF_SAND]

So I was VERY disappointing with the outcome of Bulldozer, and was thinking of not getting one and just switching back to intel. But it appears there was a registry problem! AMD has not officially stated but it's rumored that there will be a registry fix and will give between a 40% to 70% gain in performance with Windows 7.

It's useless to post a link, just google "AMD Bulldozer registry fix" TONS of info on it.

 

[Azurite]

I sure hope your right about the registry fix. We need someone to compete with intel and to push prizes down. Though, i want to see the performance boost first, and i wouldnt wanna buy it before they actually can prove it.

 

[KING_OF_SAND]

I am still debating whether or not to keep my sabertooth 990fx. If this is true I will. If not, I am selling.

 

[S_I_N]

sry sounds like a load of shite to me. I doubt a registry fix would give that much of a boost

 

[JezEngland]

I sure hope your right about the registry fix. We need someone to compete with intel and to push prizes down. Though, i want to see the performance boost first, and i wouldnt wanna buy it before they actually can prove it.

I hope they succeed.

Intel prices are good for 1155, & i rekon Intel are giving better value for the hardware they give, because they're a hugely bigger manufacturer, have their own factory etc so can afford to give better value, and they're more expensive than AMD because the manufacturing is better, better materials etc. I doubt Bulldozer will nock many pennies if any from the Intel range. 1155 2600k under £200 is great and I think SE-E is worth the extra, it's just good value in understanding it goes alot further.

A perspective of Intel's market with 1155, comparing it to 1156 the i7 2700k is £250ish while 1156 i7 880 was about £450

 

[UFN]

I really doubt that a registry error would cripple the CPU that much

 

[dugdiamond]

i think it is BS - i mean... what OSs did AMD test BD on?

surely they tested them on all that were compatible with the architecture - and even then if there were issues they should have stated so on the INSTALLATION INSTRUCTIONS that an update was needed and supplied a link to it.

cos none of this happened, i find it hard to believe that a registry issue is the cause of the massive FAIL in performance.

i think it is just more hype to dupe potential customers.

i warned everyone on here about the 990FX chipsets (didn't i?). even i maybe am jumping ship. my i7 arrives next week

 

[Dandaman91]

I find it hard to believe this could be true.... I somehow doubt a registry fix would increase performance that much

I deffo wont be upgrading my 1100T to a BD cpu anyways unless I got it dirt cheap or something cos otherwise it would be rather pointless imo

 

[KING_OF_SAND]

This is how I see it, and why I think it may be true.

AMD BD is a whole new architecture with eight cores and while testing their BD on Windows 7, their machines had the correct registry files. Also I have a hard time believing that even AMD can make a new architecture ship that is not even better then their phenom. Also it's not uncommon to see 40% to 70% increase in performance on stuff. I see it frequently on GPUs.

But again if not, I will be switching to intel. Also dealing with registry files in my networking class, seeing stuff like this is also not uncommon with servers.

 

[nothingspecial]

I'd be suprised if AMD couldn't put together a fix for this before/at release, if it really was a registry problem.

The fact there's still next to zero stock at suppliers lead me to think there's definately something behind the B3 stepping rumours though. Wouldn't be at all suprised if they show up within the next fortnight, ie. just before SB-E.

Edit: 8120/8150 listed for pre-order at Aria and Scan, Dabs have 5 due in 3-4 days, not listed on Ebuyer.

 

[KING_OF_SAND]

Or their could have been something wrong with all the chips. The claims AMD made about performance are so far off from reality you can sue them for false advertising. But if AMD can prove the claims that it is as good as they say on THEIR machine you have no case. So I am leaning towards AMD had a goof, or anyone who has a BD chip can sue AMD and WIN!

 

[jeblesh]

if that reg fix happen 2 b true Linux bench test's looked a lot different but although no overclock bench available under Linux

stalk speed's show same differences as windows more or less

maybe wen updates for compiler's well support on or more twik's for the new things

but it still so far off 2600 even without considering OC performance difference i doubt it well compete

and again haw long until the app's u using well actually support these ?

epic fail for one graphic card

maybe with the PCIE X16 X2 it well b a nice budget system for these setup's but probably not a real competition

for x79 upcoming ones nothing more as far as i c ...

 

[hmmblah]

If there is a fault in the OS it would be in the Kernel, not the registry. Linux sees the same lackluster performance with BD as Windows and Linux doesn't have a registry.

This isn't some quick reg fix, this (if true) will take some work on the kernels of the OS.

 

[Rastalovich]

AMD have produced fixes before now for previous generations. Nothing as simple as a regfix. Windows 8, which could be the test bed, does - as a matter of fact - work microsoft magic with the 8150.

40% - 70% efficiency improvement on core allocation, processing tasks - maybe. A difference in how 8 handles tasks over 7.

40% - 70% performance gain - no chance. No regfix can do that. An overclock may well.

A "patch", which is their previous way of getting x2 Athlons working nicely on XP prior to.... sp2 I *think*.. is possible.

If there is a Windows 7 patch - it might end up as a KB.

 

[dugdiamond]

the reason that win8 may improve BD's performance is that win8 has no server edition. meaning less overheads.

win7 has server threads running that cannot be disabled.

EDIT: windows server 2013 professional (beta) = ETA feb 2012

 

[Chaney579]

Its a load as i read it, a registry error will not happen. And even so a 40-70% gain is so ridiculous of a number i cant even fathom a dang joke for it. Like a post above, Linux had the same performance and does not have a registry. Its a nice marketing try but face it. AMD flopped, not until they release a new set of Mobo's (if they do) Will i even consider the fact that BD might have some "untapped" potential. False advertising and hipe, now i have to listen to idiotic AMD fanboys on youtube post like no tomorrow about how amazing it "Will" be. Its starting to really frustrate me to the point of calling people.................

 

[Rastalovich]

the reason that win8 may improve BD's performance is that win8 has no server edition. meaning less overheads.

win7 has server threads running that cannot be disabled.

EDIT: windows server 2013 professional (beta) = ETA feb 2012

Eh ?

Here's a very rough explaination of the way Windows 8 handles threads over 7:

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

Performance here, even in beta, is quite traumatic.

 

[KING_OF_SAND]

Its a load as i read it, a registry error will not happen. And even so a 40-70% gain is so ridiculous of a number i cant even fathom a dang joke for it. Like a post above, Linux had the same performance and does not have a registry. Its a nice marketing try but face it. AMD flopped, not until they release a new set of Mobo's (if they do) Will i even consider the fact that BD might have some "untapped" potential. False advertising and hipe, now i have to listen to idiotic AMD fanboys on youtube post like no tomorrow about how amazing it "Will" be. Its starting to really frustrate me to the point of calling people.................

40% - to 70% is not uncommon. GPUs see this quite a bit with newer drivers over time.

 

[hmmblah]

AMD have produced fixes before now for previous generations. Nothing as simple as a regfix. Windows 8, which could be the test bed, does - as a matter of fact - work microsoft magic with the 8150.

40% - 70% efficiency improvement on core allocation, processing tasks - maybe. A difference in how 8 handles tasks over 7.

40% - 70% performance gain - no chance. No regfix can do that. An overclock may well.

A "patch", which is their previous way of getting x2 Athlons working nicely on XP prior to.... sp2 I *think*.. is possible.

If there is a Windows 7 patch - it might end up as a KB.

Even in Windows 8, the 8150 is behind the 2500k. Sure it may perform better than on Windows 7, but it's not enough.

 

[Rastalovich]

Even in Windows 8, the 8150 is behind the 2500k. Sure it may perform better than on Windows 7, but it's not enough.

That's the thing isn't it.

Whatever magical fixes they apply to 7 - even if it is in an attempt to massage it inline with 8 (if that's the magic bean theory), it may well be benefitting the Intel variant in anycase.

This 40-70% increase, can-not be in flat out performance. They have to be using numbers (if it comes to fruition) like nVidia use similar numbers with new graphic card drivers.

e.g. 40% increase in Batman.... erm that mean your 100fps now become 140fps - which doesn't happen. You probably get cores working 40% more efficientlty, which ends up as probably a 5% increase in overall performance.