| Author |
Message |
Niels Jørgen Kruse
Guest
|
 Posted:
Sun Nov 27, 2005 5:15 pm Post subject:
Re: The Emperor's new clothes |
|
|
Charles Richmond <richchas@comcast.net> wrote:
| Quote: | Joe Seigh wrote:
So these processor manufacturers all have these
nice new multi-core cpu's but apart from market
hyperbole (these cpu's will save the environment, etc...)
I don't see them actually doing anything to exploit their
potential. By "them", I mean them not us. We of course
know to do. But what's going on to get all the applications
to start exploiting this? The magic parallelization fairy?
Perhaps the magic paralization fairy... ;-)
|
FWIW "paralization" doesn't show up in the Dashboard dictionary widget
(Oxford American Dictionaries), but "parallelization" does.
--
Mvh./Regards, Niels Jørgen Kruse, Vanløse, Denmark |
|
| Back to top |
|
 |
Stephen Fuld
Guest
|
| Posted:
Sun Nov 27, 2005 5:15 pm Post subject:
Re: The Emperor's new clothes |
|
|
"Andrew Reilly" <andrew-newspost@areilly.bpc-users.org> wrote in message
news:pan.2005.11.27.08.01.26.37664@areilly.bpc-users.org...
| Quote: | On Fri, 25 Nov 2005 17:10:06 +0000, Stephen Fuld wrote:
I don't think so. Desktop searching is probably either totally disk
bound
or "embarassingly parallel" such that no big advance in programming
technology or expertise would be required.
Why would you expect that the bulk of applications with non-trivial
completion times (i.e., that make you wait, and consequently desire better
throughput) will all turn out to be such: i/o bound or embarrasingly
parallel (or at least fairly trivially parallel, if not actually
embarrasing)?
|
Well, first of all, I was responding to an earlier post that mentioned
desktop searching, not making a general statement. That said, except for
graphics manipulations (which, as I said earlier seems amenable to tailored
instructions (e.g. SSE, etc.) or to better use of the streaming architecture
of graphics processors), and games, which have far better than average
programmers to take advantage of multiple cores, SMT, etc., I cannot think
of a high use example of an application that meets your stated criteria.
I'm not saying that there aren't any, but if you can think of some, please
let us all know.
--
- Stephen Fuld
e-mail address disguised to prevent spam |
|
| Back to top |
|
|
Joe Seigh
Guest
|
| Posted:
Sun Nov 27, 2005 5:15 pm Post subject:
Re: The Emperor's new clothes |
|
|
Niels Jørgen Kruse wrote:
| Quote: | Charles Richmond <richchas@comcast.net> wrote:
Joe Seigh wrote:
So these processor manufacturers all have these
nice new multi-core cpu's but apart from market
hyperbole (these cpu's will save the environment, etc...)
I don't see them actually doing anything to exploit their
potential. By "them", I mean them not us. We of course
know to do. But what's going on to get all the applications
to start exploiting this? The magic parallelization fairy?
Perhaps the magic paralization fairy... ;-)
FWIW "paralization" doesn't show up in the Dashboard dictionary widget
(Oxford American Dictionaries), but "parallelization" does.
I think he meant paralyzation. Probably the result of corporate execs |
having too many botox treatments.
--
Joe Seigh
When you get lemons, you make lemonade.
When you get hardware, you make software. |
|
| Back to top |
|
|
Joe Seigh
Guest
|
| Posted:
Sun Nov 27, 2005 5:15 pm Post subject:
Re: The Emperor's new clothes |
|
|
Stephen Fuld wrote:
| Quote: | "Andrew Reilly" <andrew-newspost@areilly.bpc-users.org> wrote in message
news:pan.2005.11.27.08.01.26.37664@areilly.bpc-users.org...
On Fri, 25 Nov 2005 17:10:06 +0000, Stephen Fuld wrote:
I don't think so. Desktop searching is probably either totally disk
bound
or "embarassingly parallel" such that no big advance in programming
technology or expertise would be required.
Why would you expect that the bulk of applications with non-trivial
completion times (i.e., that make you wait, and consequently desire better
throughput) will all turn out to be such: i/o bound or embarrasingly
parallel (or at least fairly trivially parallel, if not actually
embarrasing)?
Well, first of all, I was responding to an earlier post that mentioned
desktop searching, not making a general statement. That said, except for
graphics manipulations (which, as I said earlier seems amenable to tailored
instructions (e.g. SSE, etc.) or to better use of the streaming architecture
of graphics processors), and games, which have far better than average
programmers to take advantage of multiple cores, SMT, etc., I cannot think
of a high use example of an application that meets your stated criteria.
I'm not saying that there aren't any, but if you can think of some, please
let us all know.
|
That ubiquitous piece of bloat, the browser. Having the gui events handles
by multiple event handling threads would definitely speed things up. Of
course the gui architecture would have to be completely rewritten since
the current one is deadlock prone enough as it is. But that's not a problem
of having threads. It's a problem of ignoring threads in the initial
architectural design.
--
Joe Seigh
When you get lemons, you make lemonade.
When you get hardware, you make software. |
|
| Back to top |
|
|
Stephen Fuld
Guest
|
| Posted:
Mon Nov 28, 2005 1:15 am Post subject:
Re: The Emperor's new clothes |
|
|
"Andrew Reilly" <andrew-newspost@areilly.bpc-users.org> wrote in message
news:pan.2005.11.27.21.46.21.718437@areilly.bpc-users.org...
| Quote: | On Sun, 27 Nov 2005 16:57:12 +0000, Stephen Fuld wrote:
Well, first of all, I was responding to an earlier post that mentioned
desktop searching, not making a general statement.
Sure, but it was such a nice statement, I thought it worth using as a
launching board, to get back to the topic of the Subject.
That said, except for
graphics manipulations (which, as I said earlier seems amenable to
tailored
instructions (e.g. SSE, etc.) or to better use of the streaming
architecture
of graphics processors), and games, which have far better than average
programmers to take advantage of multiple cores, SMT, etc., I cannot
think
of a high use example of an application that meets your stated criteria.
I'm not saying that there aren't any, but if you can think of some,
please
let us all know.
Well, there aren't really many things on desktops that cause any
perceptable delay at all. Hence the Forest curve.
Of those that many people encounter, media manipulations of various sorts
and searching stand out, and both of those are observably both trivially
parallelizable and/or IO bound.
Joe Seigh has mentioned browsers, and by extension most GUIs. I suspect
that most perceived browser slowness is really IO limitation (network
latency), but it's true that most existing "WIMP" systems have fairly
strong single-threading constraints in their historical implementation and
design, if not of necessity. High-end "visualization" and CAD systems
show that parallelization is useful for other heavy graphics problems, so
we should expect similar results for video games.
In the more specialized world of software development, compilation is
mostly parallelizable (and "make" handles that without much end-user
effort) up to the point where it becomes I/O bound.
What do you do that induces "wait" that you think isn't I/O bound or
trivially parallelizable?
|
Well, for the things I do (which I certainly don't claim is exhaustive or
even representative), I said I could think of nothing. But fixing the
scheduling bugs that hurt responsivness in Windows would be very nice. :-(
--
- Stephen Fuld
e-mail address disguised to prevent spam |
|
| Back to top |
|
|
Andrew Reilly
Guest
|
| Posted:
Mon Nov 28, 2005 1:15 am Post subject:
Re: The Emperor's new clothes |
|
|
On Sun, 27 Nov 2005 16:57:12 +0000, Stephen Fuld wrote:
| Quote: | Well, first of all, I was responding to an earlier post that mentioned
desktop searching, not making a general statement.
|
Sure, but it was such a nice statement, I thought it worth using as a
launching board, to get back to the topic of the Subject.
| Quote: | That said, except for
graphics manipulations (which, as I said earlier seems amenable to tailored
instructions (e.g. SSE, etc.) or to better use of the streaming architecture
of graphics processors), and games, which have far better than average
programmers to take advantage of multiple cores, SMT, etc., I cannot think
of a high use example of an application that meets your stated criteria.
I'm not saying that there aren't any, but if you can think of some, please
let us all know.
|
Well, there aren't really many things on desktops that cause any
perceptable delay at all. Hence the Forest curve.
Of those that many people encounter, media manipulations of various sorts
and searching stand out, and both of those are observably both trivially
parallelizable and/or IO bound.
Joe Seigh has mentioned browsers, and by extension most GUIs. I suspect
that most perceived browser slowness is really IO limitation (network
latency), but it's true that most existing "WIMP" systems have fairly
strong single-threading constraints in their historical implementation and
design, if not of necessity. High-end "visualization" and CAD systems
show that parallelization is useful for other heavy graphics problems, so
we should expect similar results for video games.
In the more specialized world of software development, compilation is
mostly parallelizable (and "make" handles that without much end-user
effort) up to the point where it becomes I/O bound.
What do you do that induces "wait" that you think isn't I/O bound or
trivially parallelizable?
--
Andrew |
|
| Back to top |
|
|
Guest
|
| Posted:
Wed Nov 30, 2005 4:36 pm Post subject:
Re: The Emperor's new clothes |
|
|
Bernd Paysan <bernd.paysan@gmx.de> wrote:
| Quote: | The antivirus software works basically the same way the immune system of
our
|
I concur with the original poster, the whole concept of of an antivirus
program is flawed and is a symptom of an insufficently secure OS.
| Quote: | This won't work if the malware authors had techniques to completely remove
any trace of pattern on their code, but so far, they haven't learned that.
And no, I won't give them hints here ;-).
|
They did learn, there were some really quite clever viruses around however
it became much much easier to attack scripting vunerabilities in browsers
and email clients and to write seemingly useful trojans. Things like
polymorphically encrypted checksum preserving viruses are incredibly hard to
write.
-p
--
"What goes up must come down, ask any system administrator"
-------------------------------------------------------------------- |
|
| Back to top |
|
|
Chris Barts
Guest
|
| Posted:
Wed Nov 30, 2005 5:15 pm Post subject:
Re: The Emperor's new clothes |
|
|
David Magda <dmagda+trace050401@ee.ryerson.ca> wrote on Tuesday 22 November
2005 17:43 in comp.arch <m2ek58dwi8.fsf@gandalf.local>:
| Quote: | Ken Hagan <K.Hagan@thermoteknix.co.uk> writes:
David Magda wrote:
Under Windows I'm sure Norton will chew up one of the cores. (I'm
only half-kidding.)
Lots of corporate customers have no choice but to run such filth, so
they will see an immediate benefit. (I'm not even half-kidding.)
Yes, including where I work. Didn't seem to stop the latest Sober
variants (nor did it prevent Sony's rootkit from being installed). :-/
Oh well.
|
All antivirus software, at least as the term is understood now, is a fraud
and /cannot/ work as advertised, even in principle. Trying to ferret out
bad code using static methods is equivalent to solving the Halting Problem,
something I'm sure Norton and the rest of that mob understands but
willfully ignores as long as the ignorant crowds are willing to pay money
for the software and endless stream of updates.
The only way to prevent bad code from doing damage is through strict
run-time controls on resource usage. Practically speaking, this means
running the software in a good OS. Realistically speaking, this means
dumping or severely curtailing use of Microsoft products.
(Microsoft /could/ make a good OS. They just haven't yet.)
--
My address happens to be com (dot) gmail (at) usenet (plus) chbarts,
wardsback and translated.
It's in my header if you need a spoiler.
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =---- |
|
| Back to top |
|
|
Bernd Paysan
Guest
|
| Posted:
Wed Nov 30, 2005 5:15 pm Post subject:
Re: The Emperor's new clothes |
|
|
Chris Barts wrote:
| Quote: | All antivirus software, at least as the term is understood now, is a fraud
and /cannot/ work as advertised, even in principle. Trying to ferret out
bad code using static methods is equivalent to solving the Halting
Problem, something I'm sure Norton and the rest of that mob understands
but willfully ignores as long as the ignorant crowds are willing to pay
money for the software and endless stream of updates.
|
The antivirus software works basically the same way the immune system of our
body works: It uses rather simple patterns against known evil-doers. The
pattern generation process is even less automated than in our body; but in
principle, malware that wants to spread out could be detected automatically
(by observing traffic - normal communication involves a high variety of
messages, each sent to few recipients, while worms and spam send basically
the same message to loads of others).
This won't work if the malware authors had techniques to completely remove
any trace of pattern on their code, but so far, they haven't learned that.
And no, I won't give them hints here ;-).
--
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/ |
|
| Back to top |
|
|
Dan Koren
Guest
|
| Posted:
Thu Dec 01, 2005 7:12 am Post subject:
Re: The Emperor's new clothes |
|
|
"Chris Barts" <puonegf+hfrarg@tznvy.pbz> wrote in message
news:1133364930_6285@corp.com...
| Quote: |
All antivirus software, at least as the term is understood now, is a fraud
and /cannot/ work as advertised, even in principle. Trying to ferret out
bad code using static methods is equivalent to solving the Halting
Problem,
something I'm sure Norton and the rest of that mob understands but....
|
You may be giving "Norton" and the
"mob" more credit than they deserve....
;-)
dk |
|
| Back to top |
|
|
Chris Barts
Guest
|
| Posted:
Thu Dec 01, 2005 9:15 am Post subject:
Re: The Emperor's new clothes |
|
|
Dan Koren <dankoren@yahoo.com> wrote on Wednesday 30 November 2005 18:12 in
comp.arch <438e4df0$1@news.meer.net>:
| Quote: | "Chris Barts" <puonegf+hfrarg@tznvy.pbz> wrote in message
news:1133364930_6285@corp.com...
Trying to ferret out bad code using static methods is equivalent to
solving the Halting Problem, something I'm sure Norton and the rest of
that mob understands but....
You may be giving "Norton" and the
"mob" more credit than they deserve....
;-)
|
Yea, Hanlon's Razor probably applies here as well as anywhere else. ;)
--
My address happens to be com (dot) gmail (at) usenet (plus) chbarts,
wardsback and translated.
It's in my header if you need a spoiler.
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =---- |
|
| Back to top |
|
|
Chris Barts
Guest
|
| Posted:
Thu Dec 01, 2005 9:15 am Post subject:
Re: The Emperor's new clothes |
|
|
Bernd Paysan <bernd.paysan@gmx.de> wrote on Wednesday 30 November 2005 09:05
in comp.arch <l6v063-564.ln1@miriam.mikron.de>:
| Quote: | The antivirus software works basically the same way the immune system of
our body works: It uses rather simple patterns against known evil-doers.
The pattern generation process is even less automated than in our body;
but in principle, malware that wants to spread out could be detected
automatically (by observing traffic - normal communication involves a high
variety of messages, each sent to few recipients, while worms and spam
send basically the same message to loads of others).
|
I know how the immune system works, and I think it is an ugly kludge as
well. I don't complain about it because patches don't seem to be applicable
to already existing systems. ;) The body doesn't have the luxury of
sandboxing proteins* the way software engineers can sandbox code, so the
proved methods of system security don't apply to meat machines.
*Unless they get a research grant and can buy a few scratch monkeys.
| Quote: |
This won't work if the malware authors had techniques to completely remove
any trace of pattern on their code, but so far, they haven't learned that.
And no, I won't give them hints here ;-).
|
They don't need hints: The current crop of crap is more than intelligent
enough to infect large numbers of well-connected machines. It doesn't need
to evolve any more than the flu does: there are plenty of morons with the
sanitation habits of two-year-olds to infect.
--
My address happens to be com (dot) gmail (at) usenet (plus) chbarts,
wardsback and translated.
It's in my header if you need a spoiler.
----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =---- |
|
| Back to top |
|
|
Ken Hagan
Guest
|
| Posted:
Thu Dec 01, 2005 4:16 pm Post subject:
Re: The Emperor's new clothes |
|
|
Chris Barts wrote:
| Quote: |
All antivirus software, at least as the term is understood now, is a fraud
and /cannot/ work as advertised, even in principle. Trying to ferret out
bad code using static methods is equivalent to solving the Halting Problem,
something I'm sure Norton and the rest of that mob understands but
willfully ignores as long as the ignorant crowds are willing to pay money
for the software and endless stream of updates.
|
The real beauty of the AV business model is that as Microsoft patch up
their software (and fewer viruses get through to dumb users) it's the
AV vendors who get the credit (coz it must be due to that new "filter"
software running on our firewall, right?).
That's money for someone else's old rope! |
|
| Back to top |
|
|
Stefan Monnier
Guest
|
| Posted:
Thu Dec 01, 2005 5:15 pm Post subject:
Re: The Emperor's new clothes |
|
|
| Quote: | The real beauty of the AV business model is that as Microsoft patch up
their software (and fewer viruses get through to dumb users) it's the
AV vendors who get the credit (coz it must be due to that new "filter"
software running on our firewall, right?).
|
I think the AV's success is mostly due to the fact that it tells you "here,
I've detected and deactivated this virus", which gives you a warm&fuzzy
feeling of security.
Of course, it doesn't tell you that detection&deactivation was not needed
because it's an old virus that your system is already immune to (not because
of AV but because of actual bug fixes).
So it seems useful even when it's completely useless.
Stefan |
|
| Back to top |
|
|
Brian Hurt
Guest
|
| Posted:
Fri Dec 02, 2005 1:15 am Post subject:
Re: The Emperor's new clothes |
|
|
nospam@ab-katrinedal.dk (Niels Jørgen Kruse) writes:
| Quote: | Charles Richmond <richchas@comcast.net> wrote:
Perhaps the magic paralization fairy... ;-)
FWIW "paralization" doesn't show up in the Dashboard dictionary widget
(Oxford American Dictionaries), but "parallelization" does.
|
It's an amalgamation of parallel and paralyze. The magic paralization
fairy allows you to deadlock eight cpus simultaneously. Boy, I wish I
was joking.
Brian |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in