G
A
AD.
I think he meant the problems MS have making it work - wider issues than
most people will come across. If there weren't any problems it would've
been released many months ago.
Cheers
Anton
Y
Yousuf Khan
Tony Hill said:
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?
It's possible that the NX protected pages are breaking more apps than
initially thought.
Yousuf Khan
A
Andrew Reilly
Sander said:
That model is surely going to make the famous trick of using xor
to build doubly-linked lists more interesting. (Unless they used
a typedef to mean "an integer that can hold a pointer". They
certainly wouldn't have been using "long long" for that purpose in
the 16- or 32-bit days.)
Weren't we just discussing "issues with C and pointers"
S
Seongbae Park
Yousuf Khan said:
http://www.theregister.co.uk/2004/07/15/sun_solarislives_opteron/
And I can say that there's been a lot more progress than
what's reported in this article.
Seongbae
K
Keith
I don't think so. More likely Windows is a nightmare to code/modify. Some
people like conspiracy theories, however.
Moi? Come on. Either M$ is incompetent or they're holding back. You
choose!
No (micro)$hit! Linux doesn't have the corporate skulldugery impeeding
its progress. There is a market, it will fill it. ...kinda like AMD
these days.
Perhaps not impending.
Well, there is no longer a "Power" architecture (you really should
know by now that that is a silly marketeer's term). If you're
alluding to "PowerPC", well it seems to be invading from top to bottom;
IBM's Power(TM) and blades, Apple G5 and XServe, Nintendo,
PlayStation3, X-Box2, and a ton of embedded stuff.
Yeah, it seems to be doing a tad better than Itanic! ;-)
Who me? o;-)
BTW, Dean a decent newsreader is in order. Lookout sucks.
D
Dean Kent
Keith said:
How many Windows programmers does it take to change a lightbulb? ;-).
Far too early to tell for AMD, unfortunately. 15% marketshare is a bit less
than their best. Server segment share is not too shabby, but still a long
way to go to be compared to Linux.
Hell, they're not only porting Solaris to x86-64, but are considering
PowerWhatever/IPF as well. Sun sees the light, and it is coming from
somewhere else. ;-).
It has impressive numbers, for sure. What happened to SPEC int, though?
So I've heard. I haven't made the investment (timewise) to put Linux on,
and don't have the funds to upgrade my K7 box to a K8 yet. I figure to do
that all at the same time - will you stop your whining then?
Besides,I keep hoping IBM will let me play with a Thinkpad with FLEX-ES and z/OS on
it (or they approve z/OS to run on Hercules along with an affordable
single-user license) so I can pretend I am on a *real* computer while at
home... <g>.
Regards,
Dean
B
Bill Davidsen
Pleasant said:
I think this is a serial problem, time to solution is at least the time
to solution for the most dificult problem, or more depending on how many
things can be done in parallel.
People are still doing graduate thesis on NUMA, it's not clear that more
people will mean shorter time to solution. The first cut may settle for
being stable and running well where Ntask <= Ncpu. Getting it wrong
means bigtime bad throughput, which is probably an issue, since there
are more mature Linux and Solaris (I'm told) models as competition.
I think the rest of the steps are pretty well understood and can be
solved in reasonable time. Or course things like keeping the CPU number
in a byte will need work ;-)
C
chrisv
Dean Kent said:
Does anyone care about integer performance anymore? 8)
K
Keith R. Williams
That's easy; none. But that certainly doesn't answer the question
(quite the opposite, in fact). ;-)
Linux has 15%? The server segment is not too shabby, but...
When one has so many lamps to choose from, why would one keep lamp-
designers on payroll?
Don't look at me! I didn't take it! ...which SPEC Int were you
referring to?
So I've heard. I haven't made the investment (timewise) to put Linux on,
and don't have the funds to upgrade my K7 box to a K8 yet. I figure to do
that all at the same time - will you stop your whining then?
Linux isn't needed to get a decent reader. Agent works fine under Win,
though I prefer Gravity. There are many choices, most *free*, and all
better than outhouse. It doesn't take a lot of time to switch. It took
me an hour or so to switch from Win/Gravity to Linux/PAN.
I can't remember the last time I logged onto a 'Z' system. Must be
close to five years for VM (I've long forgotten my password) and
fifteen for MVS.
H
Hellmark
Dean Kent's last words before the Sword of Azrial plunged through his body
were:
You can get different email and news readers for windows too, ya know
were:
So I've heard. I haven't made the investment (timewise) to put Linux on,
and don't have the funds to upgrade my K7 box to a K8 yet. I figure to do
that all at the same time - will you stop your whining then?
I keep hoping IBM will let me play with a Thinkpad with FLEX-ES and z/OS on
it (or they approve z/OS to run on Hercules along with an affordable
single-user license) so I can pretend I am on a *real* computer while at
home... <g>.
You can get different email and news readers for windows too, ya know
T
Tony Hill
The 32-bit version of Solaris 9 is indeed an option on the Opterons,
but that's not really what we're interested in here. It's the 64-bit
version of Solaris 10 that people are waiting for.
T
Tony Hill
The biggest issue I've been hearing about (admittedly mostly
second-hand) seems to be with actually getting the service pack
installed. A lot of people seem to have had all sorts of problems,
ranging from changing settings and breaking things right up to making
the machines blue-screen and not boot.
Once installed I have heard of a few application compatibilities,
particularly with software that does kind of funky things with
hardware (some CD burning applications, virtual drives, etc.). Some
of these problems are probably almost by design, ie they tie in to the
new security features of SP2, but in the process break compatibility
with some odd-ball software. This isn't entirely a bad thing, MS
really NEEDED to break compatibility with some software in order to
improve their security, but there do still seem to be a few rough
edges to be worked out.
Of course, that being said, "release candidates" are never really
actual candidates for release, so I shouldn't really be directing too
much blame at MS for this one. They're really just late-beta releases
that are going through their final testing and bug-fixing before an
actual release candidate can be created. In this regards RC2 seems
fine. Hopefully MS will get the rough edges smoothed out and get SP2
out by the end of the year (only about a year late) so that all the
other new operating systems can get here.
Y
Yousuf Khan
Bill Davidsen said:
I just remembered that pretty soon there will be dual-core Opterons too, so
that in itself will add another level of NUMA to take into account. Internal
chip-connect vs. Hypertransport connect vs. customized board-to-board
connect.
Yousuf Khan
S
Stephen Sprunk
Russell Wallace said:I don't know whether he's being sarcastic...
...but I hope you are!
NT was 64-bit clean, since the Alpha and MIPS ports seemed to make that a
requirement, as does the upcoming PPC port for Xbox2.
The biggest problems are getting all those 32-bit drivers converted to
64-bit, and getting 32-bit apps to work under a 64-bit OS, neither of which
was a problem for the prior ports (since they were 64-bit only).
Some "apps" like .NET were gratuitously dependent on the register size, and
those will need serious reworking, but the OS core was ported to AMD64 over
a year ago.
S
S
Stephen Sprunk
Dean Kent said:
Given MS just announced all AMD64 platforms have slipped to 1Q05,
embarrassment is pretty much a given -- not to mention Linux has been
shipping for AMD64 for over a year.
G
Greg Lindahl
Stephen Sprunk said:
I can't comment on the MIPS port, but the Alpha port was 32 bits, and
was pretty famous for it, because it was Plain Weird. I see a ton of
other wrong info in this thread, despite people's attempt to correct
it.
Followups reduced to groups I read.
-- greg
K
Keith
I don't know how you could come to any of these conclusions, given the
public information anyway.
I agree with the others here. Force the issue by eliminating those who
won't convert. It seems Linux has done a reasonable job of supporting
AMD64 *without* all the resources M$ can bring to bare.
Ah, so we have another conspiracy theory. So you're on the "incompetence"
side?
D
Dean Kent
Keith said:
Some things are designed to be easily portable (such as DB2, perhaps?),
while others are not. The incompetence does not have to be with today's
coders, but could be due to yesterday's designers...
Regards,
Dean
A
Anton Ertl
Keith said:
The Alpha and MIPS ports of WNT were 32-bit. The 64-bit WNT for Alpha
never was released publically (and AFAIK Linux/MIPS died before work
on 64-bit WNT started). However, WNT for IA-64 is 64-bit, so 64-bit
WNT exists ("WNT" here includes W2K, WXP etc.).
How does a good marketing machine and the ability to buy competitors
help in driver development? Or what kind of resources are you
thinking of that MS has and that Linux lacks?
Linux has profited here from something that is sometimes seen as a
disadvantage: That hardware vendors usually don't do Linux drivers, so
Linux developers had to write them themselves. Making them work on
64-bit kernels is less work than writing them from scratch or getting
the hardware vendor to do the 64-bit port (for a Windows version that
is not released yet).
To make 32-bit apps to work under a 64-bit OS, you need to convert the
32-bit system calls to 64-bit ones, and to provide all the 32-bit
libraries (and a way for the apps to link to the right libraries). If
you have a more complex infrastructure (as Debian does), you have to
adapt that as well (and that's why Debian has a pure64 port, with
the multiarch i386/amd64 port waiting for upstream extensions to the
infrastructure last time I looked). However, WNT should have already
tackled this for the IA-64 port, no?
Followups to comp.arch.
- anton