PDA

View Full Version : Please confirm that this guy is a moron...


Daneel Olivaw
12-12-04, 07:32 AM
http://www.tacticalgamer.com/showthread.php?t=3770

HyperThreading

To DRASTICALLY increase performance in both games and benchmarking, DISABLE hyperthreading, as it allows NO process to EVER run over 50% CPU usage. Thats how it load balances, by splitting the CPU into two "virtual" concurrent processors. While this works well for playing a FPS and burning a DVD at the same time, if you have a game that needs more cpu power, it will be left lacking.

By turning off HT on your P4C, you make it more like an AMD CPU, as in your processing power isn't being artificially "split." Thus allowing the CPU to totally dedicate itself to the task at hand.

You disable HyperThreading in the BIOS under the Advanced Menu.

Example:

Without HT on a 3.0GHz CPU, the CPU will give all 3GHz of it's power to an application. While multitasking may be hurt, your performance in a game by itself will be MUCH higher. Now if you turn HT on, you have two "virtual" 1.5GHz CPUs, and your game will only get one of them. Now you can see why gamers aren't thrilled about HT very much, as very few programs are able to "see" HT and take advantage of it. (As in using both virtual processors to speed performance)

nIghtorius
12-12-04, 09:22 AM
hereby:

*confirmed*

Riptide
12-12-04, 09:24 AM
No way. If it were true a 2.53/533 (non HT) would bench significantly faster than a 2.8/800 and we know that isn't the case.

jAkUp
12-12-04, 11:20 AM
Or a 2.4/533 vs 2.4/800.

Bunch of balony

$n][pErMan
12-12-04, 01:04 PM
Unless he has benchmarks or timedemos to prove this.... he is a moron. Need proof people!!! (cheers)

Nutty
12-12-04, 01:51 PM
Probably stems from the fact that windows is rubbish at showing cpu usage on a HT machine.. A single process at full cpu usage is only shown to be using 50% cpu. Which is rubbish.

Sazar
12-12-04, 02:04 PM
this guy reminds me of a previous nvnews regular who was absolutely convinced of the unbelievable efficiency of an HT cpu v/s amd's processors because it never hit 100% usage :rofl

btw... confirmed... this guy is a moron..

Daneel Olivaw
12-12-04, 05:37 PM
P4 2.4C @ 3.2GHz, see sig.

confirmed

Viral
12-12-04, 11:47 PM
this guy reminds me of a previous nvnews regular who was absolutely convinced of the unbelievable efficiency of an HT cpu v/s amd's processors because it never hit 100% usage :rofl

btw... confirmed... this guy is a moron..

Hehehe, that's who I thought of when I read this. I forgot his argument though.. I remember asking him countless times how a processor only using 50% of its potential in games would be a good thing. He was still convinced that it was for some reason or another.. even though it was using all of it's potential. It was like, "whoa, P4 may not beat AMD CPU's anymore, but it doesn't even have to use half of it's power and it's still close!", yet he didn't realise how stupid such a situation would be unless you discovered a way to have the CPU use all its potential in games (like this guy is claiming).

Kickus_assius
12-13-04, 12:20 AM
He is a moron on that front but benchmarks have shown that CPUs with HT are not faster than those without for gaming. But I am referring to an apples to apples comparison in GHz here. I have seen benchmarks that put 3GHZ 800 non-HT against a 3GHz with hyperthreading and the one without HT had slightly better numbers. It was only like 1-5 max difference in FPS and I think Farcry was what they tested it on. It certainly is not a "1.5 GHz CPU" but it does perform at about 95-98% of what the same speed CPU with HT turned on performs at for games.

retsam
12-13-04, 06:16 AM
i guess this guy know more about cpu's then ibm... ibm uses this type of tech ( hyperthreading) on there power 5's and multi core cpu's.. what an idiot....

Daneel Olivaw
12-13-04, 07:26 AM
I gotta say retsam, posting there and reviving that one+ year old thread might not have been a good idea :lol:

retsam
12-13-04, 11:04 AM
hahah i didnt see that before oh man that was bad hahah oh well

bluntman
12-14-04, 09:29 AM
Just show that person this photo...
http://members.rogers.com/alegardo/problem.jpg

Pandora's Box
12-14-04, 09:45 AM
lmao

nVestor
12-14-04, 10:05 AM
Wow, if that were true... you'd think someone would have pointed it out already, HT is like over a year old. (xmasmile) Thing I have noticed about HT vs. A64 is with regard to onboard sound. There is a sizable CPU usuage hit that does result in a significant fps loss as compared to onboard on a HT system.

Easy way to test is with Teamspeak. (xmasgrin)

That fella from tactical gamer should stick to games, seems hardware is not his strong point. (xmasmile) conclusion: hogwash

jbirney
12-15-04, 05:49 PM
Slighlty OT:

At the company I work for I have done some benchmark using EE programs that our engineers use (Matlab, XFDTD, LabVIEW, ADS, MentorGraphics). We saw a few cases were HT actually made the workstations (Dell 360/370s) run a bit slower. Some times it was faster. So for us it was more of a mixed bag. Since the times it did run slower was fewer and still pretty dam fast, we left it on. However some of the custom LINIX software used by the Semiconductor group did not get along well with HT. But those were speicalized apps desiged for IC verification so thats not too suprising. I just wanted to say that HT is pretty neat but not a sure fire win every were as there are some cases were it can slow you down a bit....

retsam
12-16-04, 07:45 PM
At the company I work for I have done some benchmark using EE programs that our engineers use (Matlab, XFDTD, LabVIEW, ADS, MentorGraphics). We saw a few cases were HT actually made the workstations (Dell 360/370s) run a bit slower. Some times it was faster. So for us it was more of a mixed bag. Since the times it did run slower was fewer and still pretty dam fast, we left it on. However some of the custom LINIX software used by the Semiconductor group did not get along well with HT. But those were speicalized apps desiged for IC verification so thats not too suprising. I just wanted to say that HT is pretty neat but not a sure fire win every were as there are some cases were it can slow you down a bit....

were all the aps compiled for HT ?

jbirney
12-17-04, 10:07 AM
were all the aps compiled for HT ?

Labview is a multi-threaded app. I did not have time to re-write the VI to make the most of HT. But Labview did have a section on HT. However our engineers are NOT going to take the time to re-write their VIs to make use of HT. The use Labview to test and develope products so 99% of their time is already used up trying to meet deadline, so they use already working programs.

The other programs we have to wait for the vendor to tweak them to make better use of HT.

CDsDontBurn
12-17-04, 05:39 PM
moron confirmed

Guess somebody is more of a moron than jakup :lol: