Re: Box restarting with Windows XP & Hyperthreading
From: NoNoBadDog! (mypants_bjsledgeATpixi.com)
Date: 08/22/04
- Next message: Nathan McNulty: "Re: Windows detects wrong printer"
- Previous message: Nathan McNulty: "Re: Network Printing Slow"
- In reply to: Nathan McNulty: "Re: Box restarting with Windows XP & Hyperthreading"
- Next in thread: NoNoBadDog!: "Re: Box restarting with Windows XP & Hyperthreading"
- Reply: NoNoBadDog!: "Re: Box restarting with Windows XP & Hyperthreading"
- Messages sorted by: [ date ] [ thread ]
Date: Sat, 21 Aug 2004 19:44:17 -1000
Nathan;
I agree that there has been some progress with HT, but with AMD and Intel
migrating to Hypertransport for the x86, I see little potential for further
development of HT. The advantage of Hypertransport is that it does not
require that software be recompiled to use it. Intel has confirmed that
it's x86 extensions will conform to the standards developed by AMD, and
Hyperthreading is one of those standards.
I still maintain that unless you are a "power user" and you use
predominantly those apps that currently work with HT, then it is probably
better to have HT disabled. For the majority of people, who buy their
computer to do email, VIM, surf the web, d/l music and make CDs, then Ht is
definitely not needed.
If you remember the introduction of the NetBurst P5 architecture, we were
promised that by this point in time we would be using 10 GHz processors.
But the engineers quickly learned the limitations of the 586 die, which is
why we have been stuck at the 3 Ghz level for so long. HT was supposed to
be another tool to squeeze out a little more performance, and the idea and
intentions are good, but the software developers were not exactly chomping
at the bit to recompile their code. Perhaps if HT had been introduced early
in the development of the Pentium 4 (586) strategy, then it could have been
much more successful. As it stands, HT is pretty much a moot issue, as is
the Pentium 4. It is time to move on to other architectures (PCI express,
Hyperthreading, x86 64 bit processing, etc).
Please note that I do own HT procs myself, and I am not bashing it. I
just think it's value has been blown completely out of proportion. I wish
it could have lived up to it's promise.
Bobby
"Nathan McNulty" <nospam@msn.com> wrote in message
news:u$h1%23dAiEHA.356@tk2msftngp13.phx.gbl...
> You make a very good point. I would like to show you a comparison of HT
> when it first came out to HT more recently. More recent programs have
> been better compiled and more and more are supporting HT. See the
> differences between these two:
> Old - http://www.2cpu.com/articles/43_1.html
> New - http://www.2cpu.com/articles/42_1.html
>
> As for the fix, try installing Service Pack 2. When this is finished and
> has rebooted, turn off the computer. Turn on the computer, then enter the
> BIOS, enable HyperThreading, then let Windows load. It should have no
> problems after this point.
>
> ----
> Nathan McNulty
>
>
> Nospam wrote:
>> NoNoBadDog,
>>
>> You are correct to a certain extent as the benefits of hyperthreading
>> will depend upon what the user is doing. However, software will be
>> complied more so in the future to accomodate hyper-threading. I
>> definitely notice the difference on my PC since I have several apps
>> running at the same time, some active, some in the background. Tom's
>> Hardware Guide is a respected website with factual information. This
>> link http://www6.tomshardware.com/cpu/20021227/ has a good article on
>> Hyperthreading.
>>
>> Nospam
>>
>> "NoNoBadDog!" <mypants_bjsledgeATpixi.com> wrote in message
>> news:eKjgE8whEHA.3928@TK2MSFTNGP11.phx.gbl...
>>
>>>As a side note;
>>>
>>> Unless you are using software that is HT "aware" (Photoshop being one
>>> of the FEW programs that are HT "Aware"), you will take a performance
>>> hit by having HT enabled. This has to do with the way that the OS and
>>> the proc handle long-chain branching predictions and also with the way
>>> that signaling is handled on the extra pipelines. Typical performance
>>> hit is somewhere in the range of 7% (slower), but there have been
>>> recorded instances of up to 17% (on Intel 2.8 HT P4s). In most cases,
>>> it is better to leave HT off. HT is nothing more than a marketing tool
>>> for Intel, and it really does not benefit the average user. The techs
>>> actually did you a favor by turning it off. Before you lamers and
>>> flamers start responding, be aware that I also own a HT based system,
>>> and I leave the HT off. I also own AMD based systems. and both are very
>>> worthy systems. I am not "Intel bashing". No need to respond if I have
>>> hurt anyone's delicate "Intel Rocks" feelings.
>>>
>>>Bobby
>>>
>>>"Cari (MS MVP)" <Newsgroups1@coribright.com> wrote in message
>>>news:%23pQRn0whEHA.1656@TK2MSFTNGP09.phx.gbl...
>>>
>>>>Did you contact the manufacturer of the PC?
>>>>--
>>>>Cari (MS-MVP Windows Client - Printing, Imaging & Hardware)
>>>>www.coribright.com
>>>>
>>>>"Kirk" <stormstaff@sbcglobal.net> wrote in message
>>>>news:a5fc01c48709$cdf64220$a601280a@phx.gbl...
>>>>
>>>>>My comp started rebooting all of a sudden. No viruses were
>>>>>found, so I took it into the local shop. They told me
>>>>>there is a known issue with Windows XP & Intel's
>>>>>Hyperthreading.
>>>>>
>>>>>They turned off the HT and everything is fine. They said I
>>>>>needed to contact Microsoft for a patch. I can't find it
>>>>>on the site. I called support and they don't wanna help me
>>>>>because it's an OEM version.
>>>>>
>>>>>I just want to be able to run my HT & XP together. Can
>>>>>anyone help me please?
>>>>
>>>>
>>>
>>
- Next message: Nathan McNulty: "Re: Windows detects wrong printer"
- Previous message: Nathan McNulty: "Re: Network Printing Slow"
- In reply to: Nathan McNulty: "Re: Box restarting with Windows XP & Hyperthreading"
- Next in thread: NoNoBadDog!: "Re: Box restarting with Windows XP & Hyperthreading"
- Reply: NoNoBadDog!: "Re: Box restarting with Windows XP & Hyperthreading"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
