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

[Sable]

Now don't shoot the messenger, I know it's from theinq.

http://www.theinquirer.net/?article=32589

AMD Socket AM2 has a secret weapon
Expect wonder results with Reverse HT

It seems that all AM2 CPUs were outfitted with a support for Reverse-HyperThreading, an architectural change which enables software to think that it is working on a single-core alone. By combining two cores, the company has been able to produce the six IPC "core" that will go head to head against four IPC "core" from Conroe/Merom/WoodCrest combo.

It seems that in certain cases, even an old AMD Athlon 64 3800+ can wipe the floor with Core 2 Duo E6300 CPU.

In single-treaded apps, Core 2 Duo is expected to struggle against Reverse HyperThreading CPUs, which work at higher clock frequencies and produce higher instruction per clock ratios (IPC).

AMDs Reverse-HT is a dynamic technology, and with Microsoft's Windows update and a new processor driver, the driver will copy the graphics drivers of today's 3D accelerators. The driver will detect the app, see if it is multithreaded or not and turn the ReverseHT on, or leave it off.
------------------------------------------------------------------------------------------------------

This sounds :shocked: but it could be that theinq just read that xtremesys thread and rehashed it into a story.

The reason I say this is because if RHT is as good as people reckon surely AMD would have shouted about it to combat all the Conroe hype.

 

[DrMrLordX]

Haven't we already been discussing this?

 

[Sable]

I haven't, you might well have been. Didn't see mention of the inq article anywhere though.

 

[Fraggable]

I don't understand much of what it's saying, they don't give much detail about how it actually works. Any improvement in running single-threaded apps (games) is good though. It seems it'll be a while before dual cores are really a benefit for gaming

 

[StrangerGuy]

The big question is: How much faster running single-threaded apps with RHT? (If it exists)

I predict ~50%.

 

[Vegitto]

Seriously, I believe it's true. XS' Cooler said so, and he's always been right. AL-fsckin'-WAYS.

 

[dorion]

Reverse HT will probably show some speeds ups, but it wont be the all miraculous combine two chips to be one chip everyones talking about, the closest the chips get to communication is the crossbar. And if one core needs Data the other has its going to slow down the process. But from what I can get from the article AMD might have a real time optimizer to keep the cores as synced as possible. Just wonder how they handle it, seems like a monster of a task.

 

[Ika]

uh, there's already a pretty big discussion about this one thread below yours...

Supposedly Intel is developing something like this too.

 

[classy]

If you stop and think about it for a second it would explain why AMD has done all the things it has lately by canceling almost 50-60% of its product line. And consolidating others. As succesful as they have been, there is no way they would just throw in the towel. I would also imagine that they had this planned for quite some time, considering they have embedded hardware support for this. Its really not surprising either, we have been using raid 0 with harddrives for ages. I bet in some games you'll probably see some huge gains, if it works properly. Thats the key, how well will it work?

 

[Ika]

Originally posted by: classy
If you stop and think about it for a second it would explain why AMD has done all the things it has lately by canceling almost 50-60% of its product line. And consolidating others. As succesful as they have been, there is no way they would just throw in the towel. I would also imagine that they had this planned for quite some time, considering they have embedded hardware support for this. Its really not surprising either, we have been using raid 0 with harddrives for ages. I bet in some games you'll probably see some huge gains, if it works properly. Thats the key, how well will it work?

Raid 0, huh? the one that doesn't give any performance at all? Oops.

Hopefully this thing is nothing like RAID 0 and improves performance. All power to AMD, but I'm not sure if this is going to work.

 

[imported_Questar]

Now don't shoot the messenger, I know it's from theinq.

In that case, you deserve to be shot twice.

 

[classy]

Originally posted by: guoziming

Originally posted by: classy
If you stop and think about it for a second it would explain why AMD has done all the things it has lately by canceling almost 50-60% of its product line. And consolidating others. As succesful as they have been, there is no way they would just throw in the towel. I would also imagine that they had this planned for quite some time, considering they have embedded hardware support for this. Its really not surprising either, we have been using raid 0 with harddrives for ages. I bet in some games you'll probably see some huge gains, if it works properly. Thats the key, how well will it work?

Raid 0, huh? the one that doesn't give any performance at all? Oops.

Hopefully this thing is nothing like RAID 0 and improves performance. All power to AMD, but I'm not sure if this is going to work.

The example of raid 0 just goes to show its certainly possible to create one virtual entity by using the resources of two seperate pieces of hardware. Performance is not the focus of my post at all and would not be as these are two totally different situations. But some seem to think its impossible or not likely and that is not the case at all.

 

[Duvie]

ok!!! So when are we going to see proof of this working? Conroe debut spoiler? I am skeptical considering the source, plus in their own words it wont be as effective in apps already smp aware....so for me core 2 duo is still more powerful....

Gamers may just abolut shite themselves though if this works considering the support for dual cores in games....

 

[classy]

Originally posted by: Duvie

ok!!! So when are we going to see proof of this working? Conroe debut spoiler? I am skeptical considering the source, plus in their own words it wont be as effective in apps already smp aware....so for me core 2 duo is still more powerful....

Gamers may just abolut shite themselves though if this works considering the support for dual cores in games....

Thats where I think something like this will have an impact, in games. I don't even think there are a lot games that are smp aware, are there?

 

[bunnyfubbles]

Originally posted by: guoziming

Originally posted by: classy
If you stop and think about it for a second it would explain why AMD has done all the things it has lately by canceling almost 50-60% of its product line. And consolidating others. As succesful as they have been, there is no way they would just throw in the towel. I would also imagine that they had this planned for quite some time, considering they have embedded hardware support for this. Its really not surprising either, we have been using raid 0 with harddrives for ages. I bet in some games you'll probably see some huge gains, if it works properly. Thats the key, how well will it work?

Raid 0, huh? the one that doesn't give any performance at all? Oops.

Hopefully this thing is nothing like RAID 0 and improves performance. All power to AMD, but I'm not sure if this is going to work.

Someone here doesn't have RAID 0...

 

[Sable]

Originally posted by: classy

Originally posted by: Duvie

ok!!! So when are we going to see proof of this working? Conroe debut spoiler? I am skeptical considering the source, plus in their own words it wont be as effective in apps already smp aware....so for me core 2 duo is still more powerful....

Gamers may just abolut shite themselves though if this works considering the support for dual cores in games....

Thats where I think something like this will have an impact, in games. I don't even think there are a lot games that are smp aware, are there?

Absolutely right. The whole point of it is to speed up single threaded apps/games.


Thing is, I'm very skeptical of this as well. Same as I was about Conroe. I want to see it working before I start going overboard. (isn't this what the AMD fans said about Conroe. nothing wrong with waiting for validated results)

I also want to know if this is specific to AM2 CPU's or if us poor 939 bods are gonna get a look in as well.

Can't wait to get some concrete facts on this.

edit:

btw, on the reliability of the source, it's appeared on a few other sites as well now. xbit, techreport

 

[MrFanel]

This will not happen. Inquirer is BS'ing out of their ass. The official word as of now is that a retard decided to speculate and does not know what he's/she's talking about.

 

[golem]

Thats where I think something like this will have an impact, in games. I don't even think there are a lot games that are smp aware, are there?

Wasn't one of the knocks against conroe, that most games are now GPU bound and extra CPU power doesn't have any affect beyond a certain resolution (Think Oblivion was one of the exceptions)? If that's the case, couldn't this technology actually slow down games because of the extra overhead this process might need?

 

[DrMrLordX]

Conroe sure does seem to be faster in FEAR and Oblivion. Obviously not all games are entirely GPU-bound.

 

[Fraggable]

Originally posted by: bunnyfubbles

Originally posted by: guoziming

Originally posted by: classy
If you stop and think about it for a second it would explain why AMD has done all the things it has lately by canceling almost 50-60% of its product line. And consolidating others. As succesful as they have been, there is no way they would just throw in the towel. I would also imagine that they had this planned for quite some time, considering they have embedded hardware support for this. Its really not surprising either, we have been using raid 0 with harddrives for ages. I bet in some games you'll probably see some huge gains, if it works properly. Thats the key, how well will it work?

Raid 0, huh? the one that doesn't give any performance at all? Oops.

Hopefully this thing is nothing like RAID 0 and improves performance. All power to AMD, but I'm not sure if this is going to work.

Someone here doesn't have RAID 0...

Raid 0 rules. Don't act like it doesn't boost performance, because it does when researched and used right. It doesn't do much for Raptors but it does a LOT for drives like the 2 in my sig. give me any hard drive benchmarking program and I'll give you my transfer speeds to compare to seagate's specs for an individual drive. It'll blow every one of them way out of the water. My 2 80GBs in RAID 0 kill a 36GB Raptor.

 

[n19htmare]

Doesn't intel have a similar thing... CMT or Core Multiplexing technology, already seen on some new motherboards and bios.

http://www.xtremesystems.org/forums/showthread.php?t=104178

or from intel

http://www.intel.com/technology/magazine/research/speculative-threading-1205.htm

Well there is some talk on intel side, atleast directly from intel.

CMT on a Core 2 Duo

or

CMT on Kentsfield = 2 multiplexed cores yummmmmmy

 

[Griswold]

Let's wait a month and see if this was just a bubble. Worst that can happen, is that some people here who are abso-fvcking-lutely sure that this is just bullshit, will look really stupid if AMD, somehow, pulls this off.

Me? I'm sceptical. It wouldnt affect me anyway, as I dont have a AM2 system and wont be getting one anytime soon.

 

[Madwand1]

Originally posted by: MrFanel
This will not happen. Inquirer is BS'ing out of their ass. The official word as of now is that a retard decided to speculate and does not know what he's/she's talking about.

They say that Microsoft will be providing a patch. So this would imply that there should be two "retards" -- one in MS and one in AMD. The original article explicitly mentioned plying the over-worked AMD source with alcohol. If the INQ has gotten involved and up'd the ante to women and song, then such results could be predictable.

However, if the INQ is reading this -- you have the wrong sources. I know the real story, am thirsty, have no specific preference in hair colour in the dark, and would like to hear Messiaen's quartet live again.

 

[bunnyfubbles]

Originally posted by: Griswold
Let's wait a month and see if this was just a bubble. Worst that can happen, is that some people here who are abso-fvcking-lutely sure that this is just bullshit, will look really stupid if AMD, somehow, pulls this off.

Me? I'm sceptical. It wouldnt affect me anyway, as I dont have a AM2 system and wont be getting one anytime soon.

It would effect me; I want Conroe to play (and to be my yearly upgrade), yet if this is true for AM2, there would most likley be two toys I would want to play with...more money to spend, booo - more toys to play with, yaaay!

 

[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'.