Slashdot Mirror
Hyper-Threading Speeds Linux
developerWorks writes "The Intel Xeon processor introduces a new technology called Hyper-Threading (HT) that makes a single processor behave like two logical processors. The technology allows the processor to execute multiple threads simultaneously, which can yield significant performance improvement. But, exactly how much improvement can you expect to see? This article gives the results the investigation into the effects of Hyper-Threading (HT) on the Linux SMP kernel. It compares the performance of a Linux SMP kernel that was aware of Hyper-Threading to one that was not." Ah, the joys of high performance.
"The Intel Xeon processor introduces a new technology called Hyper-Threading (HT) that makes a single processor behave like two logical processors."
I'm glad we can get this sort of up-to-the-decade insight here on slashdot.
Of course the article goes on "Linux is faster on an SMP system".
Blows my mind. HyperThreading in Intel Xeons, what's next? HyperThreading in Pentium 4s?
I don't need no instructions to know how to rock!!!!
tHATs wHAt i wAAASS gOInG tO sAy!!! iMAgINe hTAt!
(`>
/( )
LL
Something that makes your computer run faster also makes free operatings systems faster too?!
I wonder what it does for commercial OSes.
Sorry for the sarcasm, but isn't that obvious? If you have a processor that can do more work than another processor at equivalent MHz, it, by most estimations, will speed something up.
Not true for everything, but pretty close.
That what was all this school was for... to teach us how to solve our own problems. -- janeowit
I just heard some sad news on talk radio - Computer Science pioneer Don Knuth was found dead in his California home this morning. There weren't any more details. I'm sure everyone in the Slashdot community will miss him - even if you didn't enjoy his work, there's no denying his contributions to computer science. Truly an American icon.
Once again Linux gets the jump on Micro$oft. It's hard to believe that some people still don't understand that Open Source results in better software.
Karma: Good (despite my invention of the Karma: sig)
... Is a graphics card able to view my beloved child pornography at the resolutions I require. All this fancy 3D stuff is unnecessary when you know that a real kid's been made to suck their dad's cock for your pleasure.
I hope I cum before they get old
We've been doing some research in the Wagner Labs and we've seen many cases where an app is optimzed for hitting the level II cache and thus reducing the pipeling done by optimizing on modern day compilers and when you use these apps on a hyperthreaded proc you actually see a performance DECREASE by the order of Olog(n) due to the fact that the insctuction set is running parallel in the CPU and never leave the LII cache, thusly never getting a chance of utilizing the advantages of hyperthreading.
Once again this proves the point made by Fred Brooks in "The Mythical Man Month" that even if you increase the technical levels of optimiztion you can will only see actual real-world speed improvements in Olog(n)/4ac, on the average.
That said I do think there is quite a bit of potential for hyperthreading when the compilers are able to catch up, so to speak.
Warmest regards,
--Jack
Wagner LLC Consulting Co. - Getting it right the first time
What is a company like Microsoft to do about a per Processor licensing fee on machines with "virtual" SMP?
- Kill Yourself, spare us all! -
According to /. Intel is the Devil, since they monopolize the market and hurt Linus. But find some favorable morsel in their product that benefits Linux and now they're God? Make up your minds, people.
but surely someone who was qualified to comment would know how to spell kernel?
It's not that I want to pull you up for bad spelling but in order to speak with authority one must get the simple things right.
Apologies if you have some sort of linguistic problem but people who seriously study kernel performance see the word kernel constantly and therefore one would expect them to spell kernel kernel.
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter