Interesting read about upcoming K9 processors

  • Thread starter Thread starter Yousuf Khan
  • Start date Start date
Yousuf said:
I thought the point was getting a working 64-bit Microsoft system? That
would mean not just the OS, but also the apps and drivers. If it's just the
OS, then Microsoft is already done, the OS is already ready for Opteron. But
Microsoft has said that the only thing holding them back from releasing the
OS is the drivers, and a few apps which might do things and get away with in
the 32-bit OS which they won't be allowed to get away with in 64-bit.

I've been waiting since Windows 3.1 for a working 32-bit version, don't
ya know?
 
My money is on embarassement...

I guess it's somewhat embarrassing if Sun ships Opteron servers, and
can't offer Solaris to go with them. Microsoft can still afford to
wait, the vast majority of their market is still Intel and 32 bits.

-kzm
 
Yousuf said:
It would be supremely embarrassing to Microsoft if Sun gets Solaris for
Opteron out before Windows.

Yousuf Khan

It is apparently ready to ship on their 1,2,and 4-way
Opty servers and W/S that were announced on Monday.
I wonder if Sun had the hardware ready to go for a while
but were waiting on Solaris ?
 
Ketil said:
I guess it's somewhat embarrassing if Sun ships Opteron servers, and
can't offer Solaris to go with them. Microsoft can still afford to
wait, the vast majority of their market is still Intel and 32 bits.

I just found a link I'd been looking for:
http://www.sun.com/desktop/workstation/w2100z/
which says that this system ships with Solaris, with
a 64 bit Solaris available "soon".
 
You don't understand the problem... the o/s needs to make all sorts of
decisions about moving a process to another processor to load balance
vs. cost of moving, etc. It is a nasty problem!

Sure it is a nasty problem and I do understand that. But, I'm talking
about the total amount of work for Microsoft here.

How many people do you think would need to be working on the kernel
development for this particular problem as opposed to migrating the
entire Win32 API and attendant APIs such as DirectX to 64-bit? That was
my point. The real world is more than some academic abstraction that
"NUMA is hard" and so on. The real world is about delivering a complete
shrink-wrapped 64-bit Windows XP I can buy.

For those that think the issue was solved with the Itanium/Alpha ports
you obviously aren't Windows programmers nor conversant with the various
APIs.

And if anybody wants to further this discussion, please stick OT. We're
talking _Windows_ here, much as some of you might not like.
 
Rob said:
It is apparently ready to ship on their 1,2,and 4-way
Opty servers and W/S that were announced on Monday.
I wonder if Sun had the hardware ready to go for a while
but were waiting on Solaris ?

I think it's just the 32-bit Solaris that is shipping right now. I was
referring to the upcoming 64-bit Solaris.

Yousuf Khan
 
Dean Kent said:
I don't think so. More likely Windows is a nightmare to code/modify. Some
people like conspiracy theories, however. :-).


Linux isn't Windows, and therefore is a completely different argument.

Are you being sarcastic? I'd be amazed if Win32 was not 64-bit clean
from day one. The industry was a lot more mature at that point, and
hopefully learned from the migration of 16- to 32-bit...
 
chrisv said:
Are you being sarcastic? I'd be amazed if Win32 was not 64-bit clean
from day one. The industry was a lot more mature at that point, and
hopefully learned from the migration of 16- to 32-bit...

Surely if Win32 were 64-bit clean, MS wouldn't have had to ship separate
Win64 headers, which they did, to the general horror of everyone who
expected a 64-bit "long".

Furthermore, at the time of its inception, it was far more important for
Win32 code to be Win16 clean, and I doubt if MS could produce headers
that are clean for all three sizes simultaneously.
 
Are you being sarcastic?

I don't know whether he's being sarcastic...
I'd be amazed if Win32 was not 64-bit clean
from day one. The industry was a lot more mature at that point, and
hopefully learned from the migration of 16- to 32-bit...

....but I hope you are! :P
 
In comp.arch Bill Davidsen said:
Yes, since Linux is already NUMA capable 64 CPU is a configuration
option. Still, finding the bandwidth to feed those CPUs is easier if
they have some dedicated RAM (read as: Opteron).

The problem is that if Windows claimed 64 bit compat the same way linux did,
everybody would simply laugh and tell them to make up a better joke...
 
In comp.arch Bill Davidsen said:
I've been waiting since Windows 3.1 for a working 32-bit version, don't
ya know?

So you are a loser who bashes MS without knowing any actual details about
teh OS?
 
It would be supremely embarrassing to Microsoft if Sun gets Solaris for
Opteron out before Windows.

Sun currently has 64-bit Solaris for Opteron scheduled for Dec. of
this year. Word so far is that they are pretty much right on schedule
and that the OS is up and running in their labs.

FWIW I don't think the MS delay is simply an issue of drivers.
They've also delayed Win2003 SP1 for apparently the same reason as
their delay of WinXP 64-bit. All of this actually seems to tie back
in to WinXP (32-bit) SP2, which is continuously being pushed back.
All the future OSes are going to be built off the SP2 code-base and
Microsoft seems to be having no end of problems getting this update
out.

For those that aren't familiar with WinXP SP2, it is a pretty
significant change to WinXP (many have referred to it more as "WinXP
Second Edition" rather than just a service pack). Lots of positive
changes with regards to the basic security concept of the system, but
MS seems to be having HUGE problems making it work. Reports from the
recently released "Release Candidate 2" suggest that this is still
definitely beta software (certainly not an actual candidate to be
released). I suspect that until MS gets these sorted out they aren't
going to try to push the other new OSes and service packs out.
 
Sander said:
So you are a loser who bashes MS without knowing any actual details about
teh OS?

No, he sounds like someone who has tried them all
and doesn't feel they are good enough to be labelled
as "working".

For crapware like Win 3.x, Win9x, and WinMe he's got a
point, but I would disagree with him on NT4 and W2K.
 
No, he sounds like someone who has tried them all
and doesn't feel they are good enough to be labelled
as "working".

For crapware like Win 3.x, Win9x, and WinMe he's got a
point, but I would disagree with him on NT4 and W2K.

agree but some people can't get anything to work no matter what you give
them. ;p
 
Rob said:
No, he sounds like someone who has tried them all
and doesn't feel they are good enough to be labelled
as "working".

For crapware like Win 3.x, Win9x, and WinMe he's got a
point, but I would disagree with him on NT4 and W2K.

I have to agree with him on NT4, specifically file shares across long
and unreliable links. Bad idea to do, but it sucks that you have to
reboot the --ing file server to make it drop the stale locks when the
connection breaks. For some reason I never saw NT 3.51 have that same
problem... Did they ever fix that little doozey ? I lived with that
bug for 4 years all told at various PPOEs. I particular resented the
one where I lost a ton of sleep fixing batches that blew up because
the dipsticks in Hong Kong lacked the "expertise" to set up a FTP
server.

"Enterprise Ready", my arse.

Cheers,
Rupert
 
In comp.arch Ketil Malde said:
I guess it's somewhat embarrassing if Sun ships Opteron servers, and
can't offer Solaris to go with them. Microsoft can still afford to
wait, the vast majority of their market is still Intel and 32 bits.

I though Solaris 9 something was an option on the Opterons?
 
In comp.arch Pleasant Thrip said:
How many people do you think would need to be working on the kernel
development for this particular problem as opposed to migrating the
entire Win32 API and attendant APIs such as DirectX to 64-bit? That was
my point. The real world is more than some academic abstraction that
"NUMA is hard" and so on. The real world is about delivering a complete
shrink-wrapped 64-bit Windows XP I can buy.

Well... I would imaghine it depends a lot on how throughout work they want
to make of it. With teh IL32P64 model they are using, almost no extra work
might be needed for an initial api and lilbrary conversion.
 
In comp.arch Rob Stow said:
No, he sounds like someone who has tried them all
and doesn't feel they are good enough to be labelled
as "working".

For crapware like Win 3.x, Win9x, and WinMe he's got a
point, but I would disagree with him on NT4 and W2K.

Precicely - and now with WinMe being dead and essentialy superceded
with WinXP Home, his "I have been waiting for 32bit os since win 3.1"
simply makes no sense.
 
Tony Hill said:
Sun currently has 64-bit Solaris for Opteron scheduled for Dec. of
this year. Word so far is that they are pretty much right on schedule
and that the OS is up and running in their labs.

FWIW I don't think the MS delay is simply an issue of drivers.
They've also delayed Win2003 SP1 for apparently the same reason as
their delay of WinXP 64-bit. All of this actually seems to tie back
in to WinXP (32-bit) SP2, which is continuously being pushed back.
All the future OSes are going to be built off the SP2 code-base and
Microsoft seems to be having no end of problems getting this update
out.

For those that aren't familiar with WinXP SP2, it is a pretty
significant change to WinXP (many have referred to it more as "WinXP
Second Edition" rather than just a service pack). Lots of positive
changes with regards to the basic security concept of the system, but
MS seems to be having HUGE problems making it work. Reports from the
recently released "Release Candidate 2" suggest that this is still
definitely beta software (certainly not an actual candidate to be
released). I suspect that until MS gets these sorted out they aren't
going to try to push the other new OSes and service packs out.


Kinda curious where you found the reports on SP2 RC2.
Have three systems running it, with zero problems.

I'd like to read the reports so I can tell what problems
to look for and test.

--

... Hank

http://horedson.home.att.net
http://w0rli.home.att.net
 
Sander Vesik said:
Precicely - and now with WinMe being dead and essentialy superceded
with WinXP Home, his "I have been waiting for 32bit os since win 3.1"
simply makes no sense.

It was probably intended to be a joke, which you seem to have missed.

By the way, your quote is wrong, he said "working", which is a fudge
big enough to drive a big opinion through.

Followups reduced.

-- greg
 
Back
Top