1 |
On Mon, Nov 28, 2011 at 1:51 PM, Michael Mol <mikemol@×××××.com> wrote: |
2 |
> On Mon, Nov 28, 2011 at 12:26 AM, Pandu Poluan <pandu@××××××.info> wrote: |
3 |
|
4 |
[snip] |
5 |
|
6 |
> FWIW, I strongly suspect that N should be your number of *logical* |
7 |
> cores, not your number of physical cores. I believe most of the |
8 |
> overhead |
9 |
|
10 |
...and finishing my sentence. |
11 |
|
12 |
I believe most of the overhead comes from context switching between |
13 |
two threads, and hyperthreading adds CPU hardware specifically for the |
14 |
purpose of avoiding that costly context switch when controlling which |
15 |
thread's instructions get sent up the pipeline. |
16 |
|
17 |
|
18 |
-- |
19 |
:wq |