[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: To hyper-thread or not to hyper-thread



Considering Intel's hyperthereading - it seems to be there are two teams and two different opinions about hyperthreading:
It looks like team which designed P4 and I7 is saying - hyperthreading is the best.
On the other hand team which designed Core Duo2 is saying - let's not waste time with it...

I'm more sympathetic with latter so I'm looking forward for next generation without hyperthreading ;)

On Mon, Dec 7, 2009 at 4:50 PM, Chris Adams <cmadams hiwaay net> wrote:
I believe that the Linux kernel scheduler takes all of this into
account.  You are better off enabling HT (and letting the kernel worry
about taking advantage of it) than disabling it.

Can anybody point me to the source where can I find the details how is kernel making the difference between hyper-thereading cores and real cores?

I also know it is possible to set affinity of the process for the specific core but there does not seem to be any difference between hyperthreading cores and real ones.

Thank you.
[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]