True. And if you go to some of our local car dealers you can still buy a 2000 model. That doesn't mean they're still in production. All it means is that the old inventory hasn't been cleared yet.
Actually he was listing the versions of Windows that are current distributions not a list of service packs and variations sold throughout the last decade. (Don't know where you came up with Windows 97, though)
Windows 2000 comes in 4 variations (plus an embedded version for hardware manufacturers - but we're talking about commercial OS's here)
Windows 2000 Professional (Desktop OS)
Windows 2000 Server (Small dedicated servers)
Windows 2000 Advanced Server (Large dedicated servers)
Windows 2000 Datacenter Server (Huge mainframe replacement dedicated servers)
Datacenter Server is only sold with hardware that supports it like the Unisys 32 processor systems.
The Windows 98 family currently has one member
Windows Millenium Edition (ME)
So, two Operating Systems with the same API and same device driver model with five total commercial distributions for different audiences. Is that too hard to remember?
In fairness, most benchmarks do multiple passes of the test to correct for code load times and to get long enought times so that the differences don't get lost in the timer precision. If anything, conventional benchmarks should help the Crusoe since loops of test are just about ideal for their "morphing".
You test with typical apps and typical data. Whether that helps or hurts any specific design isn't really important since it is testing a reasonable usage scenario rather than a specific processor feature.
This would actually hurt the Crusoe, though, since most standard benchmarks do LOTS of looping which should play to the Crusoe's strengths. After all, most benchmarks seem to have a "do the next test 1000 times" part so that load time isn't as much of a factor. That Crusoe's "learning" should help it a lot in this kind of artificial looping test.
And if they exposed any of the interesting parts, there'd be a reason for all this discussion. Since they don't, the Crusoe is just another x86 clone. I fail to see why that's worth discussion. Now, flaming them for keeping the microcode API proprietary and secret would be worth discussion but we won't see that since Linus works there.
When you have as much free cash as IBM does the best strategy is to do a little seeding with it. If one out of twenty pays off you still win more than enought to cover the 19 losses.
True, but lets be honest. Most Linux people (including/.ers) run Linux on standard Intel, "Designed for Windows" hardware. Any advantage of portability to other architectures is not seen in practice since nobody bothers with any other options.
Y'know, as a software developer rather than a hardware developer, I'm much more concerned with the instruction set rather than how easy it is to implement caching hardware. RISC is nice for the hardware people since it makes it easier to compensate for lousy RAM access speed but it makes software harder to do.
Now, don't interpret this as saying that the PentiumX is a good CISC instrution set, it isn't, but why do we not have ANY choices besides having to reinvent 80% of the instruction set in software every time?
Actually, Microsoft seems to like having multiple architectures running the Windows NT architecture. They've had it on:
Intel iAPX32
Fairchild/Intergraph Clipper
Motorola/IBM/Apple PowerPC
MIPS Rx000
Digital/Compaq Alpha
and have announced support for Intel iAPX64 when it ships. So,it looks like the problem is that nobody buys other architectures since none of these sold well enough to keep in development.
...Bell Labs for the first 50 years of its existence had little or no government backing at the point of its invention and that includes UNIX, C and radio astronomy.
Bell Labs at the time of all these developments was totally government subsidised. Not directly but remember that AT&T was given a guaranteed profit margin as part of their being a government established monopoly. The more they spent, the more they made. Putting money into Bell Labs meant that they were given a guaranteed profit on all those expenses.
Some interesting results from all of this but probably the most inefficient use of funding. (With the exception of DOD funded research that remains classified so nobody even knows it was done)
If we skip the ones that are only sold to hardware vendors we get a choice of:
Windows ME (Home desktop)
Windows 2000 Professional (Business desktop)
Windows 2000 Server (Small Servers)
Windows 2000 Advanced Server (Big Servers)
Four choices in total. Two choices for servers, Two choices for desktops. All use the Win32 API. All use the same driver architecture. Be honest here, MS does make the choices pretty easy.
One of the biggest reasons that killed PReP and CHRP was that IBM said that having a parallel printer port was mandatory and Apple said that not having a parallel printer port was mandatory.
Isn't it amazing what entrenched bureaucracies can do to each other (with the users as innocent bystanders)
Exactly. RISC isn't inherently faster. What makes the RISC chipsets faster is that the cache management is easier and the big reason that's needed is that modern PCs don't have anywhere near enough fast RAM. Originally, PCs had 100% of their address space available to the processor in zero wait state memory (either DRAM or SRAM). Now, we treat the L1 and L2 cache the way we used to use RAM and system RAM the way we used to use the disk - as slow backing store.
What we need is a new architecture that supports direct 1 clock access to RAM. With that, we don't need complex caching algorithms and can actually do interesting things with the instruction sets like the self-optimizing microcode that Zilog was working on for the Z-80000 or actually having instruction sets optimized for programming rather than ease of cache design.
Steve Ciarcia did this in his Circuit Cellar column in Byte magazine around 1978. At the time it was a cheap display if you couldn't afford a terminal for your state of the art 8-bit computer.
Slashdot Mirror
True. And if you go to some of our local car dealers you can still buy a 2000 model. That doesn't mean they're still in production. All it means is that the old inventory hasn't been cleared yet.
Perhaps you confused the statements that it would be the last 16-bit Windows. That was the case.
Windows 2000 comes in 4 variations (plus an embedded version for hardware manufacturers - but we're talking about commercial OS's here)
- Windows 2000 Professional (Desktop OS)
- Windows 2000 Server (Small dedicated servers)
- Windows 2000 Advanced Server (Large dedicated servers)
- Windows 2000 Datacenter Server (Huge mainframe replacement dedicated servers)
Datacenter Server is only sold with hardware that supports it like the Unisys 32 processor systems.The Windows 98 family currently has one member
- Windows Millenium Edition (ME)
So, two Operating Systems with the same API and same device driver model with five total commercial distributions for different audiences. Is that too hard to remember?Operating Systems for desktops:
- Windows ME (for homes)
- Windows 2000 Professional (for corporate desktops)
Operating Systems for dedicated servers:In fairness, most benchmarks do multiple passes of the test to correct for code load times and to get long enought times so that the differences don't get lost in the timer precision. If anything, conventional benchmarks should help the Crusoe since loops of test are just about ideal for their "morphing".
This would actually hurt the Crusoe, though, since most standard benchmarks do LOTS of looping which should play to the Crusoe's strengths. After all, most benchmarks seem to have a "do the next test 1000 times" part so that load time isn't as much of a factor. That Crusoe's "learning" should help it a lot in this kind of artificial looping test.
And if they exposed any of the interesting parts, there'd be a reason for all this discussion. Since they don't, the Crusoe is just another x86 clone. I fail to see why that's worth discussion. Now, flaming them for keeping the microcode API proprietary and secret would be worth discussion but we won't see that since Linus works there.
When you have as much free cash as IBM does the best strategy is to do a little seeding with it. If one out of twenty pays off you still win more than enought to cover the 19 losses.
Companies A & B are in the same industry. Both have $100,000,000 budgeted for a new product.
Company A spends $99,900,000 developing a product (paying programmers and engineers) and spends the remaining $100,000 on marketing.
Company B spends $ 100,000 making a copy of Company A's product and spends their remaining $99,900,000 on marketing their copy.
Guess which one succeeds in the market?
Would you rather see those salaries go to marketing people or programmers/engineers?
True, but lets be honest. Most Linux people (including /.ers) run Linux on standard Intel, "Designed for Windows" hardware. Any advantage of portability to other architectures is not seen in practice since nobody bothers with any other options.
Now, don't interpret this as saying that the PentiumX is a good CISC instrution set, it isn't, but why do we not have ANY choices besides having to reinvent 80% of the instruction set in software every time?
- Intel iAPX32
- Fairchild/Intergraph Clipper
- Motorola/IBM/Apple PowerPC
- MIPS Rx000
- Digital/Compaq Alpha
and have announced support for Intel iAPX64 when it ships. So,it looks like the problem is that nobody buys other architectures since none of these sold well enough to keep in development.Or like reporting on a /. review that an MS product sucks
For more on the IOmega Click of Death problem, check out the TIP (Trouble In Paradise) section of Steve Gibson's web site
If Microsoft loses the appeal then Linux becomes the dominent and only relevant operating system in the world and Windows 2000 is dead.
If Microsoft wins the appeal then Windows 2000 becomes the dominent and only relevant operating system in the world and Linux is dead.
There's so much wrong with this picture I don't even know where to start!
Bell Labs at the time of all these developments was totally government subsidised. Not directly but remember that AT&T was given a guaranteed profit margin as part of their being a government established monopoly. The more they spent, the more they made. Putting money into Bell Labs meant that they were given a guaranteed profit on all those expenses.
Some interesting results from all of this but probably the most inefficient use of funding. (With the exception of DOD funded research that remains classified so nobody even knows it was done)
Windows 3.0 had real, standard and enhanced
Windows 3.1 had standard and enhanced
Windows 3.11 was just a bug fix
Windows for Workgroups 3.1 added networking
Windows for Workgroups 3.11 added new features
If we skip the ones that are only sold to hardware vendors we get a choice of:
- Windows ME (Home desktop)
- Windows 2000 Professional (Business desktop)
- Windows 2000 Server (Small Servers)
- Windows 2000 Advanced Server (Big Servers)
Four choices in total. Two choices for servers, Two choices for desktops. All use the Win32 API. All use the same driver architecture. Be honest here, MS does make the choices pretty easy.Isn't it amazing what entrenched bureaucracies can do to each other (with the users as innocent bystanders)
What we need is a new architecture that supports direct 1 clock access to RAM. With that, we don't need complex caching algorithms and can actually do interesting things with the instruction sets like the self-optimizing microcode that Zilog was working on for the Z-80000 or actually having instruction sets optimized for programming rather than ease of cache design.
Steve Ciarcia did this in his Circuit Cellar column in Byte magazine around 1978. At the time it was a cheap display if you couldn't afford a terminal for your state of the art 8-bit computer.
Right, and neither does anybody else. Care to make up another one. Halloween is coming up.
As for the list of processors and lies listed, here's a more accurate one.
Last I'd seen it was still supposed to ship with the processor. It can't ship earlier and still be tested against the real production chipset.
You mean like Linux is native to the PDP-8? :-)