AMD Socket AM2 has a secret weapon (inq story btw)

[n19htmare]

Originally posted by: OSX
I call BS. Seems like AMD fanboys strike again. Besides, if AMD was releasing conroe, there wouldn't be a damn person out there waiting for 'validated results'. It would be nothing but 'OMG INTEL SUXXORZZZZ I LOVE AMD'.

hahaha so true. Heck the K8L is still on paper and it's already AMD IS GONNA BE ROXXORRRR..... K8L is gonna kick Conroe's ass, K8L has this this this and this and this....look at all the stuff K8L has...


0. Native quad core
1. Hypertransport up to 5.2GT/s
2. Better coherency
3. Private L2, shared L3 cache that scales up.
4. Separate power planes and pstates for north bridge & CPU
5. 128b FPUs - see 14,15
6. 48b virtual/physical addressing and 1GB pages
7. Support for DDR2, eventually DDR3
8. Support for FBD1 and 2 eventually
9. I/O virtualization and nested page tables
10. Memory mirroring, data poisoning, HT retry protocol support
11. 32B instead of 16B ifetch
12. Indirect branch predictors
13. OOO load execution - similar to memory disambiguation
14. 2x 128b SSE units
15. 2x 128b SSE LDs/cycle
16. Several new instructions

Coprocessors:
media processing
JVM/CLR acceleration


LOOK AT ALLL THAT GOOD STUFFF!!!!!!!111 ROXXORRRR Conroe is in trouble.... GO AMD...

:: END FANBOYISM

 

[BassBomb]

inquirer? *dies*

can people stop posting news items seen on inq, it just floods these forums with nothing

 

[RallyMaster]

Originally posted by: BassBomb
inquirer? *dies*

can people stop posting news items seen on inq, it just floods these forums with nothing

QFT.

 

[jedisponge]

I don't get it. If this technology is already here, why hide it? Why not use the strategy that Intel used and let it be as well known as it could to deter consumers from buying Conroe chips?

 

[OcHungry]

Originally posted by: n19htmare

Originally posted by: OSX
I call BS. Seems like AMD fanboys strike again. Besides, if AMD was releasing conroe, there wouldn't be a damn person out there waiting for 'validated results'. It would be nothing but 'OMG INTEL SUXXORZZZZ I LOVE AMD'.

hahaha so true. Heck the K8L is still on paper and it's already AMD IS GONNA BE ROXXORRRR..... K8L is gonna kick Conroe's ass, K8L has this this this and this and this....look at all the stuff K8L has...


0. Native quad core
1. Hypertransport up to 5.2GT/s
2. Better coherency
3. Private L2, shared L3 cache that scales up.
4. Separate power planes and pstates for north bridge & CPU
5. 128b FPUs - see 14,15
6. 48b virtual/physical addressing and 1GB pages
7. Support for DDR2, eventually DDR3
8. Support for FBD1 and 2 eventually
9. I/O virtualization and nested page tables
10. Memory mirroring, data poisoning, HT retry protocol support
11. 32B instead of 16B ifetch
12. Indirect branch predictors
13. OOO load execution - similar to memory disambiguation
14. 2x 128b SSE units
15. 2x 128b SSE LDs/cycle
16. Several new instructions

Coprocessors:
media processing
JVM/CLR acceleration


LOOK AT ALLL THAT GOOD STUFFF!!!!!!!111 ROXXORRRR Conroe is in trouble.... GO AMD...

:: END FANBOYISM

Yap, Im telling you all thses fanboysm like: AMD has 4x4's, Opteron's IMC no match to Kentsfield's FSB bottleneck, conroe busted, benches are all cooked, only E6700, E6800 shows up in benches non about E6400's that crap out against 3800 x2, on and on.

 

[Lord Banshee]

If this is true, it is the only thing that would explain the 4x4 model. Why in the world would AMD release a dual dual core motherboard model for gamers... It makes no sense and amd isn't that stupied to not see 95% of all games are single threaded. For AMD's sake i hope this is true and can i think it would ne usefully if the Driver is SMART, ie know when and when not to use the Reverse-HT. For example if it is App based, i do not think it will be very good, it better be Real-Time. For example in a CAD application i use (Rhino3D) some features are able to use both cores but others can not be done effecitly with SMP so it isn't, games are the same way but have alot alot less SMP code in it. If the Driver only see the application is SMP or single threaded then i can see some bad scores where apps are SMP and Single Threaded (MOST SMP are like this) unless the effiency is >80% for the Reverse-HT which if it is ture and it is AMD needs to get some kind award as thinking about this task is just insane.

Unless the 4x4 is not for another CPU at all but actaully going to have some crazy Physics, GPU, helper co-processors which isn't expensive...

i guess we can only wait and see.

 

[ribbon13]

Well it'd be a great workstation-app boon if it will make 4 cores act as 2 faster ones. A lot of apps I use are 2CPU friendly by 4CPU doesn't do much for em.

 

[0roo0roo]

it only has to be like 10% faster in single threaded to make it a deal for many people.

 

[Avalon]

If all AM2 CPUs were outfitted with this, why is this feature turned off? Doesn't make any sense, because if this were a feature of AM2 CPUs, people would actually be buying AMD's new platform. This is one reason why I don't believe this.

 

[Newf]

The reference to "reverse hyperthreading" that I've seen looks at AMD's long-term plans, not near term. I think that if this was available today it would require OS and also probably BIOS support. For a CPU to allocate its own resources like this on the fly would be pretty amazing! I think this one is a ways off...
BTW:
Syncronizing dual cores today only requires an entry "usepttimer" in XP's boot.ini, (works for either Intel or AMD). This is a current WinXP bug fix.

 

[Lord Banshee]

Avalon,

If it is true, maybe they are still working on the software side of things, OS update and Processor Driver...

 

[aka1nas]

IIRC correctly, Intel did the same thing with the P4 and Hyperthreading. I.E. they released the chips for a bit before announcing the HT feature. It also did and does require BIOS support and in some cases OS updates to make it happen.

 

[Duvie]

Originally posted by: ribbon13
Well it'd be a great workstation-app boon if it will make 4 cores act as 2 faster ones. A lot of apps I use are 2CPU friendly by 4CPU doesn't do much for em.

right on bro!!! i saw the same thing in my testing....if each of the dual cores could be seen as one faster core then 2 of them I could actually realize the other 2 real cores advantage...2 cores to 4 cores was usually in the 5-13% range...only using radiosity did I see effective cpu usage when compiling it...

 

[Avalon]

Originally posted by: Lord Banshee
Avalon,

If it is true, maybe they are still working on the software side of things, OS update and Processor Driver...

Maybe, but if that were true and it existed, wouldn't AMD still want to pimp it early to prevent too many people from jumping to Conroe?

I still don't believe it's here in the present. If AMD can surprise me, I won't complain, though. Any feature that makes any CPU faster for the dollar is good for us.

 

[Duvie]

Originally posted by: Avalon

Originally posted by: Lord Banshee
Avalon,

If it is true, maybe they are still working on the software side of things, OS update and Processor Driver...

Maybe, but if that were true and it existed, wouldn't AMD still want to pimp it early to prevent too many people from jumping to Conroe?

I still don't believe it's here in the present. If AMD can surprise me, I won't complain, though. Any feature that makes any CPU faster for the dollar is good for us.

We will still have to see....As far as i know I cannot buy a Conroe yet...So if the was an argument against the validity of this (which I am not saying one way or another is correct) rumor/ speculation then it doesn't fly, becuase no one can "jump ship" yet with no cpus available.....

It may be take one out of Intel' book with a spoiler announcement right at conroe official launch.....


I dont know if most ppl listen to Jim Kramer the flambouyant stock guy, but he is baerish on Intel and his reason....."At one time Intel was ran by engineers, and now it is ran by marketing people"....from Friday's radio show...

 

[Furen]

While it is true that no one can jump ship quite yet having Conroe as the performance-leader-apparent is certainly not good for AMD. A launch-day announcement can only blunt 4+ months of momentum so far. I do think that the FX-64 will help out things a bit, though, even if we dont get anything besides that and the price drops.

 

[DrMrLordX]

Originally posted by: Duvie

I dont know if most ppl listen to Jim Kramer the flambouyant stock guy, but he is baerish on Intel and his reason....."At one time Intel was ran by engineers, and now it is ran by marketing people"....from Friday's radio show...

Maybe he's right, though I got the impression that the "marketeering" culture at Intel sort of went up in flames after the Prescott/Smithfield debacle. You have to admit, even at stock speeds, Conroe is a marvellous CPU that outshines their previous desktop, workstations, and small server CPUs considerably. Its mere existance speaks well of Intel's R&D. It seems to me that Kramer is commenting on Intel's status as of about six to nine months ago.

A stock analysis guy like Kramer is going to be looking at future profits for Intel. Intel is obviously going to be low-balling on the processor prices for awhile to steal market share back from AMD. This will result is lower profits for Intel in the short term, but it will also make them more competitive in the long term. If people are hell-bent on dumping Intel stocks just because the firm is trying to get back into its A game, then so be it. More fodder for the bargain-hunters.

Gone are the days of Intel outselling AMD with inferior hardware.

As far as this "reverse HT" goes, all I can add is that, if AMD plans to make its move on Intel using reverse HT, their advantage won't last for too long. Intel *does* have Mitosis in development. Mitosis on Kentsfield or its descendants would be stiff competition for K8L using "reverse HT".

 

[the Chase]

Well supposedly Microsoft is working on the OS patch for this right now. So AMD may be waiting to see what real performance gains come from this (or even if it works at all) before shouting it from the rooftops.

 

[sandorski]

Wait and see.