mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Hardware

Reply
 
Thread Tools
Old 2003-09-27, 23:32   #1
GP2
 
GP2's Avatar
 
Sep 2003

A1B16 Posts
Default Pentium V may have optional 64-bit "module"

Pentium V will launch with 64-bit Windows Elements
http://www.theinquirer.net/?article=11785

Quote:

According to this source, and the details have not been confirmed, a module sitting on top could provide 64-bit extensions.

And the source claimed, Microsoft is ready to launch a version of Windows called Elements with 64-bit extensions.

The idea seems to be that people can buy a 32-bit module, and then add in the 64-bit processor.
GP2 is offline   Reply With Quote
Old 2003-09-29, 06:26   #2
SalemTheCat100
 
Oct 2002

2·13 Posts
Default Re: Pentium V may have optional 64-bit "module"

Quote:
Originally posted by GP2
Pentium V will launch with 64-bit Windows Elements
http://www.theinquirer.net/?article=11785
Here we go AGAIN!

intel may this, intel may that.

AMD HAS 64 bits NOW, intel doesn't.

SALEM
SalemTheCat100 is offline   Reply With Quote
Old 2003-09-29, 08:12   #3
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

803010 Posts
Default

http://arstechnica.infopop.net/OpenT...6&m=7060986785
Xyzzy is offline   Reply With Quote
Old 2003-09-29, 15:36   #4
QuintLeo
 
QuintLeo's Avatar
 
Oct 2002
Lost in the hills of Iowa

7008 Posts
Default Re: Re: Pentium V may have optional 64-bit "module"

Quote:
Originally posted by SalemTheCat100
AMD HAS 64 bits NOW, intel doesn't.
Not entirely true - remember the Itanic.

9-)
QuintLeo is offline   Reply With Quote
Old 2003-09-29, 16:02   #5
GP2
 
GP2's Avatar
 
Sep 2003

13·199 Posts
Default Re: Re: Re: Pentium V may have optional 64-bit "module"

Quote:
Originally posted by SalemTheCat100
AMD HAS 64 bits NOW, intel doesn't.
Salem,
what some of us are speculating is that Intel does have 64-bits now, ready to go, and it's x86-64 bits and never mind the Itanic. Because Intel is neither stupid nor insanely stubborn, and they've been "secretly" working on it all along. They're just carefully preparing the timing of the announcement in a way that's least politically damaging to them, and most likely to steal AMD's thunder. They wait for AMD to fully reveal its hand, and then swoop in.

Remember AMD still has a fair bit of work to do to improve yields and move to 90-nm manufacturing so that they can actually make these things affordably in large numbers. So Intel still has a wide window of opportunity to grab the lion's share of the market share for x86-64 desktop chips.

Last fiddled with by GP2 on 2003-09-29 at 16:03
GP2 is offline   Reply With Quote
Old 2003-09-29, 17:11   #6
Dresdenboy
 
Dresdenboy's Avatar
 
Apr 2003
Berlin, Germany

5518 Posts
Default

Unfortunately in the next quarters Intel won't have a AMD64 compatible CPU ready that has enough performance to steal significant amounts of the x86-64 marketshare.

What several rumour sources write is, that Tejas will be the first Intel CPU which will provide AMD64 compatibility.

And test samples of Prescott show a performance increase of roughly 5% against the P4C - as expected (by taking into account the known architectural changes).

But even in 64bit mode a speculative 64bit Prescott - compared to Athlon 64/FX - can't be faster if it's not faster in 32bit mode. To make it easier to understand: if Prescott does 2 adds per cycle in 32bit mode, it can't do 3 64bit adds in 64bit mode. Maybe even only 1 because the operation is a bit more complex (32 bit add is split into 2 16bit adds while a 64 bit add would have to combine the result of 4 16bit adds).

Currently it is hard to get a higher IPC while lengthening the pipeline. With a pipeline length of 25 and an average IPC of 2.0 the CPU would have to throw away 50 instructions after a mispredicted branch. With an 12 stage pipeline and an avg. IPC of 2.8 that would be about 34 instructions.

The whole topic is much more complicated (I study such things since several years for creating a CPU making game ).

Without branches we could just take IPC*clock frequency for a rough performance estimation. But there are a lot of branches. That's a reason for the nice effect of profile guided optimizations.

Regards,
DB
Dresdenboy is offline   Reply With Quote
Old 2003-10-02, 02:31   #7
QuintLeo
 
QuintLeo's Avatar
 
Oct 2002
Lost in the hills of Iowa

26·7 Posts
Default

The most likely holdback for "Yamhill" is that Intel doesn't want to KILL the Itanic - it's taking forever to get any market momentum going as it is, and if some form of x86-64 gets popular soon enough, the Itanic WILL finally sink - along with the $BILLIONS (several of them) Intel has invested in design, building, and marketing the Itanic.

Even Intel would notice THAT sort of hit on it's bottom line....
QuintLeo is offline   Reply With Quote
Old 2003-10-02, 20:27   #8
Maybeso
 
Maybeso's Avatar
 
Aug 2002
Portland, OR USA

2·137 Posts
Default

It sounds like we need a grass-roots effort to encourage Intel to hasten the Itanic off to its reward. Perhaps one of the following:
  1. The Itanium Memorial (and R&D Recuperation) Fund
  2. We the undersigned do hereby agree to purchace an Itanium in exchange for a hastening of its demise:
    Name Qnty Email
    Spaz 1 spaz7@imlost.com
    Bob 3 bob99@help.com
    ...
  3. The Itanium for Tejas Exchange Program
    (buy an Itanium now, get a trade-in discount later.)
Post your ideas now. Hey, even if Intel won't go for it, you'll help pass the time while we wait!
Maybeso is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Aouessare-El Haddouchi-Essaaidi "test": "if Mp has no factor, it is prime!" wildrabbitt Miscellaneous Math 11 2015-03-06 08:17
"Optional user ID" automatically reset g0vegan PrimeNet 3 2008-11-20 03:58
Query: Pentium D 805 in "Thermally-Advantaged Chassis" ? cheesehead Hardware 2 2006-03-28 00:24
Would Minimizing "iterations between results file" may reveal "is not prime" earlier? nitai1999 Software 7 2004-08-26 18:12

All times are UTC. The time now is 12:56.

Tue Apr 13 12:56:07 UTC 2021 up 5 days, 7:36, 1 user, load averages: 3.46, 2.68, 2.61

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.