Interesting read about upcoming K9 processors

On Sun, 1 Aug 2004 16:17:14 -0400, "Carlo Razzeto"
wrote:
"Keith" wrote in message
news
On Sun, 01 Aug 2004 00:43:17 -0400, Carlo Razzeto wrote:

My point is that M$ *DOESN't* deal with the average user. OEM's are stuck
dealing with the average user.

--
Keith


Ok... Perhaps I didn't state what I meant very clearly... What I meant by
deal with the average user is create a system which will do all of the
things that the average user thinks it needs to do...

Windows *DEFINITELY* does not do this!

If that means run a
badly written application perfectly then that is what Windows has to do.
Even if they aren't dealing with customer directly it's still their
problem....

It's not really their problem and they don't really have a solution.
There are lots of old, badly written applications that just won't run
at all in WinXP... and MS doesn't really care.

And Microsoft's solution to this? "Call your OEM". Seriously. MS
support is essentially non-existent, and when you've got a monopoly
and don't have to worry about support, it doesn't much matter what you
do.

-------------
Tony Hill
hilla underscore 20 at yahoo dot ca

Even if they aren't dealing with customer directly it's still their
problem....
No it isn't! That's the point. They wash their hands of *all* support.

Have you watched Pat Gelsinger's talk at Standford? At one point he mentions
going to the developer of MS Flight Simulator and begging for a change in some
peculiarity of the VM implementation, because he wanted to change it in the
next chip, but some version of MS FS relied on it to work. He was immensely
relieved at being told that code relying on that feature was no longer in
current version of MS FS, so he could go ahead with making the change.

And this was for a non-architected detail of implementation.

Microsoft has a _HUGE_ backward-compatibility problem. Sometimes, they
can give reasons such as "security!" for breaking something.

Jan

glen herrmannsfeldt wrote:
+---------------
| Nick Maclaren wrote:
| Tim Shoppa wrote:
| DEC C V6.0-001 (which was rather current as of late 1998)
| under Alpha VMS 7.2:
| I remember now (and have just got a colleague to check). Yes, VMS
| uses that model, but Tru64 uses the normal I32LP64 one. As far as I
| know, the sum total of C compilers on systems that anyone normal has
| ever heard of that use IL32LLP64 is two, and both of those are relics
| (i.e. I believe that Microsoft's future direction is I32LP64).
|
| When was "long long" invented? It isn't part of C90, but
| seems to have been implemented in many compilers. I don't
| believe that it existed yet when Alpha came out, though.
+---------------

Well, the Amdahl mainframe C compiler supported "long long" at least
as early as 1985, since I used it to write an Ethernet device driver
for Amdahl's UTS-5 (running on an IBM 3081), and "unsigned long long"
was commonly used for CAWs & CCWs (Channel Address & Control Words).


-Rob

-----
Rob Warnock
627 26th Avenue URL:http://rpw3.org/
San Mateo, CA 94403 (650)572-2607

Seongbae Park wrote in message ...
Yousuf Khan wrote:
Tony Hill wrote:
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.

I'd never heard of that until now. Doing a Yahoo search only revealed a few
articles from 2003 (too old now to be really useful), and some Sun articles
being suitably vague ("real soon now"). You got something to link to?

http://www.theregister.co.uk/2004/07...lives_opteron/

And I can say that there's been a lot more progress than
what's reported in this article.

Seongbae

Does Solaris-AMD64 run Solaris-386 apps?

In article ,
Rob Warnock wrote:
glen herrmannsfeldt wrote:
|
| When was "long long" invented? It isn't part of C90, but
| seems to have been implemented in many compilers. I don't
| believe that it existed yet when Alpha came out, though.

Well, the Amdahl mainframe C compiler supported "long long" at least
as early as 1985, since I used it to write an Ethernet device driver
for Amdahl's UTS-5 (running on an IBM 3081), and "unsigned long long"
was commonly used for CAWs & CCWs (Channel Address & Control Words).

Well, it was also in Algol 68 :-)

"long long" was widespread but very restricted in use before C99,
and was typically used ONLY for such esoteric purposes. K&R and C90
both required that "long" be the longest integer type which, inter
alia, meant that it had to be large enough to address the largest
allocatable object. There was a slight kludge on 16-bit systems,
where the extra size allowed for by unsignedness was used.

Neither required nor assumed that it could address all of memory
(i.e. map all pointers) nor address all files. Unix did assume
the latter, but the assumption had already broken down by 1990;
"long long" was sometimes (NOT often) used for file addressing.

This mess is related to the ISA one which puns integers and pointers.
There is no need for this, and it is not the case on the AS/400, nor
would it be on capability machines. It was not a major hardware
problem at the time it got established, but is a noticeable headache
now, when it is normal for the bits used in a register file to be
much less than the size of the file.

In both cases, the root cause is a failure to separate different
aspects of the architecture, and ending up constraining the design
to match a small number of misdesigned programs rather than the
(surprisingly) larger number of better ones.


Regards,
Nick Maclaren.

What you mean by the "Unless of "course"..."?

BTW: are your sure

-Jouni

As always, AMD is good and Intel is bad. Same old spiel. YAWN!

Sure, when Intel tries to twist the market (i.e. consumer) to their
benefit and there is an alternative with a consumer-friendly alternative,
you bet Intel is *BAD*! To think otherwise is simply stupid. Unless of
"Course"...

BTW, top-posters suck!

Michael S wrote:
Does Solaris-AMD64 run Solaris-386 apps?

It should.

Yousuf Khan

glen herrmannsfeldt wrote in message news:hHSOc.193341$IQ4.139784@attbi_s02...
When was "long long" invented? It isn't part of C90, but
seems to have been implemented in many compilers. I don't
believe that it existed yet when Alpha came out, though.

In a 1985 "draft proposed C Standard" it was listed as a "common
extension". That said, I didn't encounter it until a couple of
years later in a spanking-new compiler for a rather bizarre 64-bit-data-wide
DSP built out of ECL and with a Harvard architecture (the C was
compiled to microcode, the instruction word was a hundred-and-something
bits wide...)

Tim.

Available in the released 32 bit version.

"Yousuf Khan" wrote in message
t.cable.rogers.com...
Alexander Grigoriev wrote:
Did you check WinServ2003's versioned file system? Keeps multiple
version of a file, allowing you to roll back, or fetch an old version.

Is that a feature of the 64-bit WinServ2003 only, or is it available even
on
the 32-bit version?

Yousuf Khan

On Mon, 02 Aug 2004 04:08:50 -0700, Jouni Osmala wrote:

What you mean by the "Unless of "course"..."?

Would it have made more sense if he spelled it "Corse"?

It probably won't unless you were reading c.s.i.p.h.c a year or two ago.

Google for it if curious.

Cheers
Anton