T
Tiberius
Provided to you "as is" without my "smarty-pants" comments I usually add.
http://arstechnica.com/news.ars/pos...f-windows-to-be-fundamentally-redesigned.html
Microsoft executive Ty Carlson spoke about the future of Windows recently
during a panel discussion at the Future in Review 2007 conference held in
San Diego, California. Carlson said that future versions of Windows would
have to be "fundamentally different" in order to take full advantage of
future CPUs that will contain many processing cores.
"You're going to see in excess of eight, 16, 64 and beyond processors on
your client computer," said Carlson, whose job title is director of
technical strategy at Microsoft. Windows Vista, he said, was "designed to
run on one, two, maybe four processors."
Carlson is tipping his hat to the fact that little growth is expected from
straight MHz scaling of single CPU cores over the coming years. Multi-core
is the only way to go (for now), but Microsoft isn't exactly behind the
times. The Windows kernel has supported multiple processors since the first
release of NT (which for marketing reasons was called version 3.1) back in
1993. The NT kernel can allocate various processes and threads to different
CPUs, and the maximum number of CPUs that it supports is generally an issue
of licensing, not technical capability. (There is a hard limit, however, on
NT systems: 32-bit Windows can have only 32 total processor cores, and
64-bit Windows has a 64-core limit, no matter how many physical processors
are in the system).
It only makes sense that as the multi-core scene matures, so will
Microsoft's embrace of it. Whether or not this embrace will result in
something "fundamentally different" is not particularly clear, and given
that Carlson is more of a marketing person than a technical one (he
previously held the position of manager of the rapid deployment program),
there's always the possibility that "fundamentally different" means nothing
more than "different." Still, let's give him the benefit of the doubt.
Carlson might be talking about features that exist in specialty versions of
its operating systems, such as Windows DataCenter Server, trickling down to
the consumer level. Alternatively, Carlson might be implying that the layers
of software, including the user interface, that run on top of the Windows
kernel may need to receive an overhaul.
Currently, Windows already spawns many different processes and threads as it
goes about its business, but there are still areas where it could be
improved. Back in 1991 when ex-Apple executive Jean-Louis Gassée was
starting development of a new operating system called BeOS, its designers
tried to make the entire operating system "pervasively multithreaded" in
anticipation that multiple CPUs would be much more common in the future.
This ensured that any one window that became unresponsive would not
interfere with any other windows, although the forced multithreaded
programming model increased the risk of programming errors such as race
conditions and deadlocks. It does seem highly unlikely, however, that
Microsoft would make major changes to the GUI model, given that they just
rewrote the 20 year-old GDI/GDI+ model for Windows Vista.
It will still take some time before many applications take full advantage of
multithreading and thus full advantage of multiple CPU cores. Even game
developers, who by necessity need to stay on the cutting edge of
performance, have only recently started taking advantage of multiple
threads.
Exactly how Windows in particular will adapt to the reality of dozens of
cores on a single chip is still not yet known, although I bet that Ars
readers can come up with all sorts of interesting and unusual ideas. When
contacted for clarification of Carlson's statements, a Microsoft
spokesperson would only say "We are not giving official guidance to the
public yet about the next version of Windows, other than that we're working
on it. When we are ready, we will provide updates."
http://arstechnica.com/news.ars/pos...f-windows-to-be-fundamentally-redesigned.html
Microsoft executive Ty Carlson spoke about the future of Windows recently
during a panel discussion at the Future in Review 2007 conference held in
San Diego, California. Carlson said that future versions of Windows would
have to be "fundamentally different" in order to take full advantage of
future CPUs that will contain many processing cores.
"You're going to see in excess of eight, 16, 64 and beyond processors on
your client computer," said Carlson, whose job title is director of
technical strategy at Microsoft. Windows Vista, he said, was "designed to
run on one, two, maybe four processors."
Carlson is tipping his hat to the fact that little growth is expected from
straight MHz scaling of single CPU cores over the coming years. Multi-core
is the only way to go (for now), but Microsoft isn't exactly behind the
times. The Windows kernel has supported multiple processors since the first
release of NT (which for marketing reasons was called version 3.1) back in
1993. The NT kernel can allocate various processes and threads to different
CPUs, and the maximum number of CPUs that it supports is generally an issue
of licensing, not technical capability. (There is a hard limit, however, on
NT systems: 32-bit Windows can have only 32 total processor cores, and
64-bit Windows has a 64-core limit, no matter how many physical processors
are in the system).
It only makes sense that as the multi-core scene matures, so will
Microsoft's embrace of it. Whether or not this embrace will result in
something "fundamentally different" is not particularly clear, and given
that Carlson is more of a marketing person than a technical one (he
previously held the position of manager of the rapid deployment program),
there's always the possibility that "fundamentally different" means nothing
more than "different." Still, let's give him the benefit of the doubt.
Carlson might be talking about features that exist in specialty versions of
its operating systems, such as Windows DataCenter Server, trickling down to
the consumer level. Alternatively, Carlson might be implying that the layers
of software, including the user interface, that run on top of the Windows
kernel may need to receive an overhaul.
Currently, Windows already spawns many different processes and threads as it
goes about its business, but there are still areas where it could be
improved. Back in 1991 when ex-Apple executive Jean-Louis Gassée was
starting development of a new operating system called BeOS, its designers
tried to make the entire operating system "pervasively multithreaded" in
anticipation that multiple CPUs would be much more common in the future.
This ensured that any one window that became unresponsive would not
interfere with any other windows, although the forced multithreaded
programming model increased the risk of programming errors such as race
conditions and deadlocks. It does seem highly unlikely, however, that
Microsoft would make major changes to the GUI model, given that they just
rewrote the 20 year-old GDI/GDI+ model for Windows Vista.
It will still take some time before many applications take full advantage of
multithreading and thus full advantage of multiple CPU cores. Even game
developers, who by necessity need to stay on the cutting edge of
performance, have only recently started taking advantage of multiple
threads.
Exactly how Windows in particular will adapt to the reality of dozens of
cores on a single chip is still not yet known, although I bet that Ars
readers can come up with all sorts of interesting and unusual ideas. When
contacted for clarification of Carlson's statements, a Microsoft
spokesperson would only say "We are not giving official guidance to the
public yet about the next version of Windows, other than that we're working
on it. When we are ready, we will provide updates."