Another Conroe Thread

[Pederv]

What if Conroe has Reverse Hyperthreading?
With no BIOS control, what if it is enabled by default?

I figured what the heck, got to put up with these Conroe threads for a few more weeks. May as well add to the FUD while it lasts.

 

[DrMrLordX]

Originally posted by: Pederv
What if Conroe has Reverse Hyperthreading?
With no BIOS control, what if it is enabled by default?

Intel is already working on speculative threading. Their initiative is named Mitosis. You haven't heard of it yet? It's not ready for prime-time (not by a long shot).

May as well add to the FUD while it lasts.

Please don't.

 

[dmens]

Originally posted by: Pederv
What if Conroe has Reverse Hyperthreading?
With no BIOS control, what if it is enabled by default?

It doesn't, and even if it did, it wouldn't be user-controllable. Too many complications.

 

[zsdersw]

I like this thread. It gives those who complain about all the Conroe threads something else with which to get their undies in a bunch.

 

[Pederv]

Well if you take the time to consider the questions that I'm asking.
It would help to explain why the performance increase varies so much from benchmark to benchmark.

 

[DrMrLordX]

Originally posted by: Pederv
Well if you take the time to consider the questions that I'm asking.
It would help to explain why the performance increase varies so much from benchmark to benchmark.

No it doesn't. What this entire thread explains to me is that many of us on this board (myself included) really don't understand modern CPU microarchitecture. Few people do. It's useless to throw wild theories like this out there when we can't even begin to explain why CPUs perform as well as they do based on microarchitecture.

Asking those in the know is one thing, but making silly declarations is quite another.

What I do know, and what I will repeat, is that Intel has "reverse hyperthreading" (they call it speculative threading) in the works (mitosis). It's not implemented in silicon yet, so therefore it can not be affecting Conroe's performance. It's not on the Kentsfield CPUs circulating in the wild either.

 

[DrMrLordX]

Originally posted by: zsdersw
I like this thread. It gives those who complain about all the Conroe threads something else with which to get their undies in a bunch.

Eh? Since when did dmens complain about Conroe threads? *P

 

[dmens]

Originally posted by: Pederv
It would help to explain why the performance increase varies so much from benchmark to benchmark.

What variation? The only outlier reported thus far is with certain floating point applications, and that has been analyzed to death already. Otherwise, merom is pretty consistent beating the competition.

 

[zsdersw]

Eh? Since when did dmens complain about Conroe threads?

He didn't.. but others have.