Gates: 64-Bit Desktop Computing Will Be Mainstream In Two Years

CRN ^ | 11:38 AM EST Wed., May 05, 2004 | Paula Rooney

[Ernest_at_the_Beach]

Microsoft Chairman and Chief Software Architect Bill Gates predicted that 64-bit computing will become mainstream in the PC industry over the next two years because of 32-bit extended technology that preserves customers' existing investments.

"64-bit will be common sense in the next couple of years," Gates said, noting that the high-performance processor engine will take off on workstations and desktops as 64-bit/32-bit capable chipsets supercede stand-alone 32-bit processors in the next couple of years. "It puts us at the very high end of computing, and it will be a smoother transition than those previously."

To that end, Microsoft announced at WinHec 2004 in Seattle that its Windows XP 64-Bit Edition for 64-Bit Extended Systems is scheduled to be released in the fourth quarter.

The Redmond, Wash., software giant also unveiled the release of an enhanced Driver Development Kit for the 64-bit Extended environment that will support Visual Studio.Net 2003 and Microsoft .Net Framework 1.1, which will enable more application development for the 64-bit Extended environment.

Gates said Microsoft will take advantage of processors with "multiple cores" that enable parallel processing on the PC desktop and server. The company will support typical eight-core server processors and two-core desktop processors, Gates said. He also urged ISVs to begin developing applications that exploit the new architectures.

Microsoft's Windows Server 2003 for 64-Bit Extended Systems, now in beta testing, also is scheduled for release during the second half.

Extended is the second 64-bit architecture supported by Microsoft. The Windows XP and Windows Server 2003 Extended offerings are based on 64-bit extensions to the x86 instruction set and support both AMD 64 and Xeon processors with 64-bit extension technology.

[Swordmaker]

Apple has never been big on backward compatibility. I have some programs left over from the DOS 1.0 days that still run on XP.

The only reason you say this is ignorance. There are Apple Mac applications that ran on the original Mac that will still run on my G5 in Classic Mode.

Apple got this reputation because they did not provide for a way to run Apple II applications on the Mac until much later. The Apple II and the Mac were not in any way related except by maker.

[Myrddin]

The new AMI BIOS released for my motherboard was incorrectly auto-detecting the DVD-RW drive. When I put a disk into the DVD drive, Windows 2000 would access it and crash. The solution was to set the BIOS entry to "Not Installed". There is a BIOS choice for CDROM/DVD, but the choice was inaccessible. Poor testing on the part of AMI. Windows still probed the interface and found the DVD drive. I've updated the DVD drive firmware as well (DVR-105). All better now.

Can you recommend a motherboard that will host the new AMD hyperthreaded, dual-core CPU? If not, I can just wait for the Tom's Hardware crew to review the new chips and motherboards.

[bigcheese]

Hmm... is it anything like the autocad boxes that feed coordinates to a big box machine that then mills an object in 3d? I had a client who had a couple of those, they were really cool. Amazing, actually. It (the 'big box') would pick up different cutting/drilling/grinding devices and then go to work on a chunk of metal. I wanted one, but he said he had bought them used for a quarter of a mil each. A leeeee-til out of my budget for a toy.

[Hodar]

First off, BIOS is an unholy mis-mash of legacy (and poorly documented) settings; mixed with the latest advancements that will accelerate your board. It's not unusual for BIOS to change very basic settings for a given device numerous times during the boot (and post) cycle; with the correct setting hopefully being the final setting (before Windows takes over; and BIOS is removed from the loop). A very large number of settings in BIOS may be over-ridden by Windows. But, I digress.

My personal favorite at the moment is the NVidia chipset. Specifically, I would look for a NForce3 350 chipset. These are made by numerous suppliers (MSI, Asus, etc.). This gives you the option of checking with the mobo manufacturer for specialized updates, or going directly to Nvidia for their latest releases.

Tom's hardware recently finished a good round of these mobo's; with the nForce 3 scoring very well. These will accept the Anthlon 64. But, here's the catch (ie. where I plead ignorance); I do NOT know which package the dual core processors will be produced in. There is a ~700 pin package, and a ~900 pin package. I suspect the dual core will be in the ~900 pin package, but I DO NOT KNOW FOR CERTAIN. This may be something for you to dig into.

[theFIRMbss]

>Microsoft Chairman and Chief Software Architect Bill Gates predicted that 64-bit computing will become mainstream in the PC industry over the next two years because of 32-bit extended technology that preserves customers' existing investments.
>>DOS Forever.

FORTH, Inc. 25+ Years of Productivity

[justlurking]

"640k ought to be enough for anybody." -Bill Gates, 1981

It's a funny joke, but Gates denies he ever said that.

I'm not generally included to believe him, but no one has ever produced any evidence (other than hearsay) that he said it.

[justlurking]

I'm wondering if the hyperthreaded Intel CPU makes life more interesting when testing a multi-threaded program. A real two processor machine introduces the possibility of concurrent access to the same memory location, thus the need to include flawless concurrency locking schemes.

It does. The critical section problem still exists: for all practical purposes, you must treat it as if there are two CPUs. In addition, there is a "yield" instruction that should be used by spin-locks, which will give control to the other CPU in a hyper-threaded processor.

If the hyperthreaded CPU acts identically to a real 2 CPU machine, I'll spring for one of those ASAP.

No, it doesn't. The two CPUs in a hyperthreaded processor share some functional units. I'm not sure which ones, but the floating point unit is a likely candidate. So, if both CPUs try to do a floating point operation at the same time, one must wait for the other to finish.

Intel claims that running two instances of the same application on a hyper-threaded CPU yields a 15-30% increase in throughput. SPECint_rate2000 measures about 10% increase in throughput.

[William Terrell]

Probably a good idea for my company to start doing away with our old Commodore 64s.

You're going to upgrade to slide rule, pen and ink?

[NoClones]

yeeehawwww!!!

[js1138]

Heck, I have a WordStar (CP/M) and Turbo-Pascal v1.0 (CP/M) which runs under CP/M 2.2 that runs on a z-80A emulator which runs in the DOS emulation of Windows XP Professional (SP1). Talk about your back-wards compatibility...

And NASA claims they can't read data from old tapes where they've lost the format definition. Poopheads.

[N3WBI3]

If youre going to run Oracle, run it on *nix and use existing 64bit technology..

[Myrddin]

My personal favorite at the moment is the NVidia chipset. Specifically, I would look for a NForce3 350 chipset.

That's a good start. I've been using MSI and Asus boards for the last 4 years. Tyan was my preference before that time. I will hold the purchase until I know whether the 700 or 900 pin package is the preferred arrangement.

I was dismayed to see the AMI BIOS 1.70 with a defect as simple as pre-selection of CDROM/DVD for the IDE device type. It's a simple menu selection that is visible to the end user. How hard can it be to properly test the ability to set the value? It reminds me of the software quality quip, "If carpenters built houses like programmers write software, the first woodpecker of the morning would destroy civilization".

[Myrddin]

The critical section problem still exists: for all practical purposes, you must treat it as if there are two CPUs. In addition, there is a "yield" instruction that should be used by spin-locks, which will give control to the other CPU in a hyper-threaded processor.

I figured the critical section issues would be more exacting. I'm accustomed to doing that already. UNIX device drivers need to be protected from interrupts with spl8()/splx(old) guards. In QNX, I put a "sleep(0)" statement in selected loop threads to yield the CPU back to another thread. The monitor facilities in Java and C# make locking a critical section easier. I suspect the impacts on each of these types of critical sections will differ a bit on a hyperthreaded processor.

[CyberCowboy777]

Very much like that.... small enough to fit in a dental office.