[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Multithreading question
From: |
Matteo Morini |
Subject: |
Re: Multithreading question |
Date: |
Wed, 29 Jan 2003 19:23:22 +0100 |
User-agent: |
Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.0.2) Gecko/20021120 Netscape/7.01 |
Well, the thread is getting a little bit too hardware-oriented for me to
follow; still I'd like to give you my 2c.
If what you need from your model is a noticeable performance boost I'm
afraid you have no other choice than making it able to run on multiple
cpus (or virtual cpus, if I understand correctly what HT is about), but
*splitting it into multiple processes*. I'm quite sure partitioning OS
tasks to one of the virtual cpus in an HT physical cpu (or one of the
physical cpus in a 'proper' MP pc) is not that big advantage, since I
doubt your OS has much else to do besides idling and routing some stray
interrupt, when there aren't other 'heavy' processes running on the
machine, unless you have a big deal of I/O caused by your code
(continuous disk access? Network traffic?). This is especially true,
IMHO, when we're talking about >3Ghz processors.
Cheers, Matteo
Darold Higa wrote:
[snip]
I understand that Intel's HT will give some boost in performance even if your
application is a monolithic block since OS tasks are partitioned off to one of
the virtual processors.
[snip]
Darold Higa
==================================
Swarm-Support is for discussion of the technical details of the day
to day usage of Swarm. For list administration needs (esp.
[un]subscribing), please send a message to <address@hidden>
with "help" in the body of the message.
Re: Multithreading question, Marcus G. Daniels, 2003/01/29
- Ok, what about benchmarks, Darold Higa, 2003/01/29
- Re: Ok, what about benchmarks, Marcus G. Daniels, 2003/01/29
- RE: Ok, what about benchmarks, Darold Higa, 2003/01/29
- RE: Ok, what about benchmarks, gepr, 2003/01/29
- Re: Ok, what about benchmarks, Marcus G. Daniels, 2003/01/29
Optimization Ideas I collected: [Re: Ok, what about benchmarks, Paul E Johnson, 2003/01/29