Rage3D Discussion Area

Rage3D Discussion Area (http://www.rage3d.com/board/index.php)
-   Front Page News (http://www.rage3d.com/board/forumdisplay.php?f=75)
-   -   Intel to Demo it's Dual-Core CPU at IDF (http://www.rage3d.com/board/showthread.php?t=33806425)

1stFlight Mar 1, 2005 03:39 PM

Intel to Demo it's Dual-Core CPU at IDF
 
Bringing you only the freshest news daily, Intel has announced they will be highlighting their dual-core CPU at this years Intel Developers Forum.
At its developer forum next week in San Francisco, Intel will make dual-core processing a key theme of the show. Included will be demonstrations of dual-core chips built using the 65-nanometer manufacturing process.

Source:Ziff-Davis' eWeek


tristancarton Mar 1, 2005 03:45 PM

scared at what the move to 65nm will entail let alone dual core.

i just remember. oh yeah 90nm p4s are going to run so cool and be able to clock up to 5ghz easily, 130nm p4s are going to become extinct and suck.

p4 90nm comes out. proclaimed as worlds smallest heat generator. what does the future entail for intel's 65nm and dual core process??? (sure hope that they do a good job on the process change. smp processing kicks serious ass compared to single)

1stFlight Mar 1, 2005 04:31 PM

Quote:

Originally Posted by tristancarton
scared at what the move to 65nm will entail let alone dual core.

i just remember. oh yeah 90nm p4s are going to run so cool and be able to clock up to 5ghz easily, 130nm p4s are going to become extinct and suck.

p4 90nm comes out. proclaimed as worlds smallest heat generator. what does the future entail for intel's 65nm and dual core process??? (sure hope that they do a good job on the process change. smp processing kicks serious ass compared to single)

Just think of them as insole warmers... very expensive insole warmers.

Seriously, I'm not looking as forward to the dual-core move as most. In looking at SMP and gaming, the only games that really support SMP have been the iD family of games, and to date no one really runs them in that configuration. In short, for desktop apps, dual core is almost worthless

Now server side that's a whole other animal, figure with the myriad of processes going SMP makes a serious difference. Even in just a mail server (Exchange2K for example), the SMP difference can be seen.

So all in all, personally if I purchase a dual-core it'll be more for my normal upgrade process, I doubt because of a real need.

mahi Mar 1, 2005 07:15 PM

funny...athlon fx's give off the similar heat....they are all around 70-90 watts...all hot..

and no one here remembers paliminos? i do...7000rpm fans bleeding ear drums..

HOPE intel is a success with this one...so you can get cheaper parts from amd (genius)...

then hope amd with trump intel so you can get cheap tech from intel...etc...
thats the way of the enthusiast...the way of the fanboi is to cry for their "company" who pays them nothing when they fail or succede.....

i like both intel and amd...

amd64 for gaming...
intel for ocing and day to day use...they both serve me well...

dreddlox Mar 1, 2005 11:58 PM

SMP for games will take off quickly, as more and more gamedevs look at Cell programming, etc. it will become closer and closer to double perf. with dual-core.
At the moment, the main concern about making games multi-threaded(and thus able to use more than one CPU) is that with only 1 CPU, having the second thread immediately causes an ~1% performance loss and a 5-20% increase in code complexity. Also if you launch a second thread, the threads will alternate at (very)approx every 5-20ms, unless they're tightly synchronized, this will cause delays before the code can move between jobs.

Intel17 Mar 2, 2005 05:54 AM

Thing is, with games, most of the work is offloaded to the GPU rather than the CPU in terms of graphics. So we won't need that much of a CPU power increase in the coming years.

dreddlox Mar 2, 2005 02:22 PM

Well, yes and no, though many things like visibility checks, lighting and shaders are becoming available in the GPU, there are many things that just cant be replicated on the GPU efficiently.
Matrix trees, for example. They cannot be done efficiently on the GPU, because for almost every matrix calculation on the GPU, the CPU must give the GPU information on how it needs to be done(although matrix buffers are a viable option, just not avaliable in the forseeable future).
Physics is the prime example. Physics can benefit hugely from SIMD but not so much from SMP. The problem is that in its most accurate state(and most likely to be used in the future), events such as collisions must happen in order, otherwise you'd get freaky stuff like a train hitting someone after they had jumped out of the way(presuming this all took place in a single physics step).
Once perfect physics is achieved, im sure we'll start looking for accurate ambient lighting and raytraced effects, neither of which can run on GPU

Elysian Mar 2, 2005 02:38 PM

Intel unveils five dual-core processors

Quote:

Originally Posted by Tomshardware.com
San Francisco (CA) - Intel took the wraps off five of its 15 currently developed multicore processors at IDF. The company believes the technology will allow its processors to gain performance in the next four years faster than in any other timeframe in their history.

The manufacturer unloaded an enormous amount of information about its future platforms and processor that left the impression, that Intel wants to be absolutely sure that it has the right products in place to maintain its dominating position in the microprocessor business.


As expected, the firm displayed dual-core chip products across all of its market segments, ranging from the Pentium D 800-series ("Smithfield"), its 65nm successor "Presler", the mobile version "Yonah"which also will be the firm's 65 nm processor on the market next year, to a dual-core Itanium ("Montecito") as well as the 65 nm "Dempsey", which will become the next-generation Xeon processor in 2006.

10 more dual-core processors are under development right now, but information rleased was largely limited to new codenames on the roadmap. "Bensley" is a Xeon dual-processor (DP) platform that will include the Dempsey processor and will be first to integrate Intel's I/O acceleration technology I/OAT in the 2006. The Xeon MP family in the 2006 timeframe is code-named "Truland" and consists of the 90 nm "Paxville" processor and a demand-based switching technology (DBS) that enables teh system to decrease power consumption.

Following Paxville, "Tulsa" will be the first Xeon MP chip in 65 nm and become a member of the "Reidland" platform. Another generation into the future, "Whitefield" will be the first Xeon MP processor that shares "some" platform architecture elements with the future Itanium processor "Tukwila", formerly code-named "Tanglewood".

The update for the Itanium family will be the 90 nm "Montecito", which is scheduled to make its debut in the fourth quarter of this year. The processor with 1.72 billion transistors will include up to 24 MByte of cache, deliver about 50 to 100 percent more performance than its predecessor with significantly less power consumption: Instead of swallowing 122 watts, Montecito will consume 100 watts, according to Intel. Following Montecito, Intel outlined "Millington" for dual-processor and low-voltage systems and Montecito's successors "Montvale" and Tukwila. "Dimona, based on Tukwila, will succeed the Millington processor.

Pat Gelsinger, head of Intel's Digital Enterprise Group, expects that these changes will "result in the fastest rate of performance increase of our time." Parallelism will allow to chip developers to speed up processors ten-fold between 2005 and 2008, the executive said. "By the end of the decade, mainstream desktops will handle eight threads, mainstream servers 32 threads. You have to think different of developing applications for the marketplace," he told conference attendees. "We are dead serious about developing dual-cores."

For the desktop, the Pentium D 800-series will make its debut as a dual-core processor that integrates two merged dies in one package. Its successor Presler will move to two individual dies that allows Intel to select and match dies into one package. "If you yield a 3.2 and a 3.8 GHz processor die in a single-die package, then the clock speed of the processor will be 3.2 GHz," explained Intel spokesman Howard High. "With two individual dies, we can select dies that match in clock speeds which enables us to achieve the best possible frequency for the processor.

http://www.tomshardware.com/hardnews...02_024903.html


All times are GMT -5. The time now is 02:35 PM.

Powered by vBulletin® Version 3.6.5
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.
All trademarks used are properties of their respective owners. Copyright 1998-2011 Rage3D.com