AMD's reverse HT

[Redstorm]

AMD Socket AM2 has a secret weapon

If this turns out to be a winner, my recommendation would be to go for a Dual Core Athlon 64 AM2 for your gaming PC.

Having both cores working on a single thread should give AMD 6 IPC vs Conroes 4 IPC which would completely errode Conroes adavantage..

Bring on the 23rd of July..

 

[Bateluer]

Repost.

Still just a rumor too.

 

[theteamaqua]

just like core multiplex technology

 

[zsdersw]

Having both cores working on a single thread should give AMD 6 IPC vs Conroes 4 IPC which would completely errode Conroes adavantage..

Unlikely.. even if it's more than a rumor, although if the AMD fanboy in you desires to seek *anything* to downplay Conroe I can see how you'd jump to those conclusions. The most it is likely to do is have some impact in some situations and no effect at all in most others. That's not a "complete erosion" of any advantage Conroe brings to the table.

It's also not really a "secret". Intel's got it too (the above mentioned core multiplex technology).

Nothing Intel or AMD do is a "secret" between them.

 

[dmens]

Originally posted by: Redstorm
Having both cores working on a single thread should give AMD 6 IPC vs Conroes 4 IPC

Even if this technology exists (and I assure you it does not), the "IPC" (a dubious metric to begin with) will not double, not even close.

 

[theteamaqua]

no i do think AMD reverse HT exist, they r just trying to work out some sort of function that can beat conroe liek the 4 x 4. its stupid ! why would ppl spend 2 times hte money jsut to get probably marginally better performance (maybe not better, no one has done any benchrmakr compare 3.9Ghz conroe with AM2)

 

[Kromis]

What is this fabled "reverse HyperThreading" that you speak of?

 

[BrownTown]

AMD fanboy's pipedreams....

but really, there is not evidence either RHT or CMT are based on specualtive threading, and if they are there is nothing but rumors to suggest they are anything but a tech demo.

 

[DrMrLordX]

Read the article linked by the OP or read any of the threads that came before this one. This is a repost two times over.

 

[Trey22]

Being under NDA, I cannot either confirm or deny this technology does or does not exist .

 

[Conky]

Yeah, but the new Core 2 Duo's feature reverse-3DNow! instructions so it's pretty much back to the benchmarks. :laugh:

 

[treskin]

According to my sources the Core 2 will infact have Anti-Reverse-HyperThreading With Duplex Matrixing and Bi-Operational Signal Flux Co-Processor capabilities.

 

[Fox5]

Current multithreaded apps often see as little as a 10% to 15% performance boost.
Assuming AMD's reverse HT matches that, that puts them about on par with a Conroe for performance per mhz, at much slower mhz, running much hotter, drawing more power, and the advantage not carrying over to true multithreaded apps.

 

[kknd1967]

AMD X2 users, can you benchmark the difference with this?
"AMD reserved HT" utitlity so called by INQ

 

[coldpower27]

Rumors for the current time, more like AMD Fanboy's pipedream for the time being we will see.

 

[Fox5]

Originally posted by: kknd1967
AMD X2 users, can you benchmark the difference with this?
"AMD reserved HT" utitlity so called by INQ

That sounds more like a bug-fix than anything else.

 

[Imyourzero]

Conroe 2 is rumored to feature RTPHT (Reverse Transient Polygamous HyperThreading), so even if AM2 does successfully master Reverse HT, BIG DEAL!

 

[Pederv]

This is going to be needed sooner or later.
Quad core comes out, frequency is equal to mid range dual core.
Octa core comes out, frequency is equal to mid range quad core.
Etc.,.....

But anyway for applications written to take advantage of dual cores, combining quad and octa cores to act as duals would probably provide a bigger boost than running them in native format.