Disabling Hyperthreading in XP

I have heard that disabling Hyperthreading in the BIOS of XP will give you more power in optical burning applications. I have read a couple of articles and I would love to get some feed back on this issue.

First I’ve heard of it.

Well I have done it and my speed has increased from a avg 11X to an avg of 12.5X for a full 4.3 gig. I was just wondering if was just my Intel 2.8 mhz chip or other chips with hyperthreading have this result. I don’t really know why this is the case because I don’t know how hyperthreading works, but if someone does your input would be appreciated.

That doesn’t make sense to me. Why would disabling HT actually improve performance?

Also, nice avvy, you play often?

Yes. You are the first person to recognize my avatar.

Dude, I dunno why you’d notice ANY performance increase in burning by playing with hyperthreading! If you’re having trouble making 16x burns, your hard drive may not be providing data fast enough. You should check to see if you have a fragmented hard drive, maybe some sort of virus scan accessing the hard drive slowing stuff down, or any of that stuff. My friend’s system has a 3.0c northwood, and it doesn’t puke at all with hyperthreading enabled. I dunno, I think you’re barking up the wrong tree, my burner goes above 15x all the time when I burn stuff. It’s also only got a 2mb buffer and operates in UDMA2. There’s a ton of stuff you could check to figure it out, but CPU speed has little to do with burn speed unless there’s something interfering with burning. :doh:

Putting data down onto a DVD is essentially a serial process, and one which is unlikely to benefit from being broken up again into a series of parallel threads. As Deer Slayer says, there’s very little processing to do because once you’ve decided to burn, the data should be ‘ready-to-go’. I would suggest that comparing averages of 11x and 12.5x are hardly significant, unless you know how these are calculated.

The advice on investigating HDD fragmentation seems highly relevant.

Bear in mind that the test system was DUAL CPU, hyperthreading, so it is possible that scheduling mistakenly put two tasks on one CPU when hyperthreading was avaialable.

Hyperthreading means that two logical CPUS make use of certain doubled units in the CPU core, and hopefully alternate units that are not doubled, while occasionally stalling on busy units, or sufferring from cache starvation.

Ther are possibly a very few examples where the stall from two threads requring the same execution unit might be longer than the scheduling delay without allow HT, but you almost have to design code to work against HT - paradoxically, two threads of P4 optimized code may not hyperthread as cleanly as unoptimized code, but then most code will not be optimized.

The one thing I can think of, is that maybe hyperthreading could increase contention for hard disk access, as Windows could begin a pre-emptive swap activity at the same time as the writer thread is trying to access the drive.

That was a great link, as I said I am looking for a solution, and I really did not know how Hyperthreading worked so I appreciate your input and now believe my problem is elsewhere. I will continue to try to optimize my system because as the previous poster said burning at 15x is possible with an avg of 12x. My LG 4163 give me a great burn at 12x but craps out, it seems as if the system buffer runs down and refills and when that happen the speed slows to 8x and stays there for the remainder of the burn. I just want to find out why. Thanks.

ok, id like to add some input to this.i have an 3.0 ht processer. & when enabled it seens to drag!! progs are slow opening& internet slows to a crall. the 1st hd is allways defraged & the second one (movies&music ect…) i,ve never defraged, dont want to destroy any vid files!!!did it once & lost losts of info!!! well when disabled in bios it seens ok. mabie like a 2.5 but nothing special.any sugestions??often wondered what was wrong with it?? but could never come up with an logical answer.hopefully someone can help!!!

HT is sometimes ‘broken’ under high load such as SQL for instance.
//Danne

Not using Nero by any chance, are you?

The fact is that I am getting a better transfer rate with HT disabled in the BIOS. I do not know why. I do have Nero my my machine but do not use it that much. I use VSO CopytoDVD to burn.

Try this link there is a patch which might help with your problem

http://forums.pcper.com/showthread.php?t=404179

Thank you Mike, I’ll be back with the results.

Ah yes, noticed the behaviour on my P4 640 (Prescott) with EIST enabled (any Windows power profile other than “Desktop” or “always on”).

Unless a a multthreaded load is seen, it equates a full single threaded load to 50% on each logical CPU unit, so does not raise the performance state - though for EIST, the minimum state is still 2.8GHz

The tracking of load across multiple CPU units, and a better reaction to it in terms of raising the power state is what this tackles, though I switched to using RMCLOCK to control my power state.

http://cpu.rightmark.org/products/rmclock.shtml

It’s a must have for any CPU with multiple power states, as it’s monitoring and control is far better than Windows.

Same situation. But I have been helped by the thread above by MikeP and evryone else who responded, thankyou. I am currently working on getting that patch from Microsoft.

alan1476 if you are having any problems getting the patch try link.

http://www.hardforum.com/showthread.php?t=961914