Sign in with
Sign up | Sign in
Your question

Hyper Threading make my pc hang??

Last response: in CPUs
Share
June 18, 2004 4:10:03 PM

can anyone please tell me why my pc will hang when i on my hyper threading?
my pc is P4 2.8E,
msi 865PE neo2-p,
kingston KVR400X64C3A 256MB * 2
OS is Windows XP sp1
Anonymous
a b à CPUs
June 18, 2004 5:06:29 PM

While doing waht kind of work? surfing the web or heavy duty video editing etc?
Using a prescott it might be heat related what are your temps? maybe the cpu is trhotling itself I dont know give us more details.
IS dynamic OC enable ?

Asus P4P800DX, P4C 2.6ghz@3.25ghz, 2X512 OCZ PC4000 3-3-3-8, Leadtek FX5900 w/ FX5950U bios@500/1000, 2X30gig Raid0
June 18, 2004 7:36:10 PM

Yes, check your tempratures man, you can use software from the motherboard cd to know your tempratures.

Make sure they don't go over 60c.

CPU will protect itself by hanging the system when overheated.

Incase it overheat, make sure your fan is attached to your cpu in the proper position.

You might need to update your BIOS to fully support HT, some BIOS requires updates to work properly with HT.

<A HREF="http://www.clan-chaos.com" target="_new">clan CHAOS</A>
Related resources
June 19, 2004 2:09:34 AM

This is common with the "virtual" core enabled. Some programs don't like it when they are threaded into the virtul, verses real core.

My P4 locks up at random, too. It's fine with it set to single CPU. No, heat isn't the issue. Read up on this problem and you will find it simply has to do with where an application loads. Sometimes it gets the "real" core, sometimes it gets the "virtual" core.

This idea is pretty lame at best since it is not ubiquitous in its operation.
June 19, 2004 6:49:58 AM

>>>This is common with the "virtual" core enabled

Sorry, but saying it's "common" is just BS. Software that chokes on HT-enabled systems is actually pretty rare. I've been using a P4 with HT for a year, and have yet to encounter a program that has issues.

>>>This idea is pretty lame at best since it is not ubiquitous in its operation.

Nor could it be, given some of the ill-behaved software that's out there. And what do you think is going to happen with that same software when it encounters the upcoming dual-core systems? At least you can turn off HT if you need to, dual cores I'm not so sure about.
June 19, 2004 4:39:43 PM

Sorry folks, it is a common isuue with "virtual" core CPU's to be software sensitive. You said it yourself when you illuded to the fact that poor software doesn't know how to handle hyperthreading. Before you post, read up on how hyperthreading works, and what happens when an application is shoved off into the virtual thread.

Dual CPUs are NOT virtual processors! They are two LOGICAL processors. Software is designed for logical CPU's.

POST Screen display of physical versus logical processors
The word "Physical" is implied when the count of processors is stated on the BIOS POST screen. The processor count should refer to logical processor count only if explicitly stated as such. The system BIOS should recognize two logical processors, but only one physical processors. Intel® Desktop Boards supporting Hyper-Threading Technology† will display only one physical processor during POST.
- Intel site.

Don't defend illogical conclusions. Hyperthreading does lock. You WILL experience issues with more software than without it turned on. If rebooting every so often isn't a concern, turn it on. Me, I use my Email program (one it HATES) daily and the small performance improvement it provides is 99% useless. This is marketing, not applications driven.

I've posted the info before, but we decide to believe and support false pretenses on hyperthreading. I DON'T want anyone to have problems with hyperthreading, but people do, and will continue to.
June 19, 2004 5:26:11 PM

>I've posted the info before

You mean you keep repeating your unsubstantiated FUD ? Spare us. If you've got a credible link to prove enabling hyperthreading results in widespread stability issues, by all means share it. Otherwise, STFU, no one believes you. intel sells gazillions of HT enabled P4s and Xeons, if there where truly serious issues with it, we would have found out.

If I wanted to give you the benefit of the doubt, I would have suggested your PSU is too weak, and enabling HT might just raises the powerconsumption over the treshold of what your PSU could supply, but I don't, cause you're an ordinary troll. A stupid one at that.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
June 19, 2004 11:03:46 PM

rower, what the hell are you talking about.

Hyper threading rocks, never had one problem with it.

<A HREF="http://www.clan-chaos.com" target="_new">clan CHAOS</A>
June 20, 2004 2:12:09 PM

PeteRoy,

It rocks with software that is designed to thread and branch the CPU's cycles. If not, it can even slow a system down because software still has to determine if a line of code is threadable, per say. This information has been plainly illustrated by Tom and Anandtech in several articles on HT technology. To fully understand your "support" of a technology, you need to know where it helps verses adds to your system overhead and complexity. The marketing engine at Intel expects you to not think about all this when you buy their products.

Software and hardware "rocks" when it is a performance benefit. H.T. is slim to none, to slowing some systems down most of the time. Nothing of any real home PC nature uses it (see related Tom's web site documentation). Multi core physical CPU's are already taking its place and with just as few apps utilizing even this capability.

Your experience with HT technology with respect to "crashes" still doesn't address the root of the question, "does it really improve my PC's usability with current software?"

Have you noticed any indication of a CPU shoot out being decided by HT technology in an Intel verses AMD comparison, and why not? The HT spread is so small that it is largely irrelevant. The basic CPU physical single line code execution pipeline is far more important than multi-proc like comparisons.

If a user is REALLY multi-proc aware, he/she will be a server client with code designed SPECIFICALLY for dual CPU's and their hardware drivers to insure stability.

That's what I'm talking about.
June 20, 2004 3:34:45 PM

have you seen the toms article about the HT enabled p4 against the OC'd by 400 mghz p4? HT really makes professionally desighned programs fly, like photoshop and premier(maybe adobe just makes HT programs or something)...i know this because i was goofing off with my friends video editing compueter(he didn't like that much haha) but i noticed an enormous increase in performance when HT was enabled.

however, rower is right in that it doesn't give any 3d performance gains. dont need any "substantial evidence" from a review site, just run the benchmarks yourself, you'll see.

so HT only benefits you if you are in a professionaly syle enviroment where you use professional programs. premier is 700 dollars! i dont know how my friend could afford it!

"I see" said the blind man to his deaf dog
June 21, 2004 2:49:01 PM

normally it works well.. but when i try to surf net then it will automatic hang there.. i m using 56k modem.. USB port..
Anonymous
a b à CPUs
June 21, 2004 3:17:16 PM

BTW in WIndows 2k/XP PRo you can set whats called Processor affinity.
Im pretty sure that if you set affinity to only one cpu(the physical one), you will resolve your so called conflict. I even think THG made an article and a little app for that...

And why so much hate about HT you had any bad experience?Ht is a nice move to keep the P4's long pipeline filled and I never had any issue cause of it.
Running some Dual Xeon HT server and its really smooth...

Asus P4P800DX, P4C 2.6ghz@3.25ghz, 2X512 OCZ PC4000 3-3-3-8, Leadtek FX5900 w/ FX5950U bios@500/1000, 2X30gig Raid0
June 21, 2004 3:20:07 PM

zkian, you might want to see if the modem manufacturer has a driver update. When HT first came out, I recall reading about some modem driver that didn't work with HT. With a bit of luck they may have an update to fix the issue by now.
June 21, 2004 4:50:03 PM

I use a HT P4 at work and I think it's great. I've only ever heard of two things that had a problem with it: some obscure modem driver, and Interbase 6.0 OpenSource. I ran into the second one myself, but through doing some research of my own I simply found switching to FireBird solved the problem - and it's a better database anyway! Of course I could have started bitching about it, but that wouldn't have helped any future customers that had a P4.... :tongue:

I could just as easily say that all these modern PCs that are faster than a 486 are crap because some old DOS games I own run so fast as to be unplayable :eek:  . It's badly developed software to blame, not the hardware.

---
Epox 8RDA+ V1.1 w/ Custom NB HS
Summer's here! so ease off the overclock...
XP1700+ @166x12 (~2Ghz), 1.475 Vcore
2x256Mb Corsair PC3200LL 2-2-2-4
Sapphire 9800Pro 400/730
June 26, 2004 7:07:01 PM

Hate HT? No, I don't hate it at all. It doesn't like drivers and software that want to see a single CPU, that's all. Yes, it does cause instability with those few drivers and software programs. And yes, there are bound to be work around.

I use XP Home and besides, I can select single CPU in BIOS as well. The problem isn't a showstopper but it is still interesting to define the exact issues and solve them. This is a hobby after all.

To change the rules with a new CPU and then say that all the older programs or drivers that worked with single CPU's and don't with HT turned on are broken is silly. They met the old rules, and are as such not broken. They also work fine with AMD CPU's and about any other (Cyrix) even older CPU's. This doesn't say HT technology is "BAD" and you can't run around and say every non-compliant driver / program is also bad written BEFORE HT even existed.

All sort of drivers needed to be redone for XP as an example, and much older hardware was ditched due to lack of support. Was this XP's "fault"?. Well, it depends on how realistic you are. It is from an absolute sense, and isn't from a continued technology sense. Either way, a solution could exist with new drivers or revisions to software.

I am simply trying to find the driver or revisions to make it work. There is enough hate being managed without my help.
Admitting that a problem exists is the first step towards a solution. Some driver or software is the problem and once identified is there a solution?

My hunch, It's the modem driver and not the actual program itself.
June 26, 2004 7:22:08 PM

How is it that you can "blame" software designed before HT was even around?

And, where on earth are you getting this complaining complex? Solving a driver or software issue is complaining? OK, if you say so (good grief).

I don't have a single issue with HT until it crashes, either. But I can't ignore Email. So I turn it off. No big deal until I fix it.

Now, you did point out one of the, I can't count how many times I've said this.... FEW applications, that cause it to crash. My modem is suspect based on triangulation of the problem and it is what the news group are all about. Does anybody else have the issue and what was the solution and not whose fault it is. Pretty soon we'll be standing outside the poor guy's house who wrote the exact line for the driver and saying, "HA, it is YOU WHO did this!!"

So what. It worked with single CPU's when he wrote it. So it isn't any more broke than HT. Lets move on shall we?

I have a US Robotics modem, any body else have a work around or see an issue with this modem? A better drive may not be available. Any users with analog modems and experiencing any problems (better driver)?
!