PII vs PIII

[Steve Wolfe]

Actually I've owned two.

Then by all means, please tell us your experience with them!

And where are the stats to back up what you said?

Where are YOUR stats?

Since you've owned two dual-CPU systems, then you should CERTAINLY be able
to back up YOUR position.

Please, I'm dieing to know. Since you're the man who demands proof,
surely YOU, of all people, did benchmarks. Please, oh enlightened one, tell
us the numbers you got on your dual CPU systems. We're waiting.

steve

 

[Gregory L. Hansen]

All that and still nothing to back it up.

None from you, either.

The problem with Photoshop is the software, though its smp capable you will
never see a big improvement in processing. Dual machines have no advantage
to a single machine that has an equal amount of processing power and with up
to 20 percent of overhead will show a definite disadvantage.

Okay, here's a few URLs for you to look at. If you want more, use Google.

http://www.chaosmint.com/benchmarks/powermac-g5-ps7bench/

Not quite a straight comparison because it compares a G5 1.6 GHz with a G5
dual 2.0 GHz, but the dual machine is about twice as fast, sometimes
around 2.5 times as fast.

http://www.irb.uni-hannover.de/~brehm/publications/habil.pdf

See especially page 208.

http://www.kikumaru.com/pc/celeron/dcbench.html

Single versus dual Celeron.

http://www.cpuscorecard.com/cpufaqs/jul00e.htm

Dual PIII up to 80% faster than single PIII, if we can believe Intel.

http://www6.tomshardware.com/cpu/20030422/opteron-23.html

Benchmarking some graphics software besides Photoshop. The dual Xeons and
Opterons beat their similarly-clocked single processor counterparts in
every test by at least 50%.

 

[Gregory L. Hansen]

Here's a tip for you in life: Becoming a vehement, rabid dog over
something with which you have no experience is usually just going to make
you look like a fool.

But it will get you bombarded with more and higher quality information on
Usenet than you could ever get by asking politely.

 

[~misfit~]

was understand,

Actually I've owned two.

What CPUs were in those machines and what did you use them for Lane?

I've always read your posts and have thought you to be informed and helpful.
This thread is making me reconsider my opinion of you.

It's becoming an increasingly likely assumption that you mainly pick up your
information from websites and newsgoups you read and pass it on. That, with
a little experience of your own, is not a bad thing in and of itself, and
can help a lot of people who ask questions in newsgroups.

However, websites and benchmarks aimed at, for want of a better word,
'fan-boys', don't always cover real-world computing and are certainly no
substitute for hands-on experience. Just because something can't be
backed-up by a website or a benchmark doesn't make it untrue. "There are
more things in heaven and earth than are dreamt of in your philosophy
Horatio." to quote The Bard.

As an example, albeit a bit tangential, I frequent
alt.comp.hadware.overclockers. Several mods discussed over there include
breaking the pins off CPUs, a move always seen as being irreversible. I
asked for, and was given, a mod method which required just such an action.
However, I managed to break off the wrong pin. I was told by all the
regulars that I was screwed. This was the gospel according to a.c.h.o and
was widely accepted as bring true. Against all advice I attempted, and
succeeded in, soldering the pin back onto the CPU, rendering it functional
again. It's now running at a 50% overclock, running rock-solid and stable
and has been for months. Now the advice usually given over there is to make
sure you break off the right pin as it can be very difficult, if not
impossible to fix if you don't.

Listen to the other contributers in this thread Lane and you just might
learn something valuable. Something beyond the scope of the hardware review
sites and benchmarks.

 

[=?ISO-8859-1?Q?M=E1ximo_Casta=F1eda?=]

Steve Wolfe escribió:

Please, oh great one who knows everything. Expound to us the cost of
context-switching on x86 hardware, the effects of high levels of interrupts,
and the difference between PIC and APIC interrupt handling.

Now, I'd like to now about that. Any pointers, please?

 

[kony]

And even more nonsense, but not even one website, datasheet, review,
anything to back up any of their claims.

Does anyone else start to see a pattern here. The reason they can't provide
any proof is because it doesn't exist. The dual cpu on the desktop is a
joke, it provides less than a ten percent improvement over a same sized
single cpu system and that's only with special programs written for a dual
machine. This has been know for years so why this nonsense keeps popping up
again and again is beyond me.

Lane

Hi Lane... you sure got a huge thread going, eh? ;-)

I agree, SMP on a desktop is never anywhere near 2X the performance of
one, but I'd put the performance in two categories, single-treaded
apps and multi-threaded (with supportive OS, etc)... for multithreaded
the performance boost can approach 60% or so, but only in these
specific tasks... makes for a good specific-purpose workstation but in
everyday use it's a lot closer to the 10% you mentioned.


Dave

 

[w_tom]

No support is provided by Steve Wolfe, et al for these
claims of improvement. Instead of posting facts and numbers,
they discuss some sort of 'usability'. IOW its called junk
science reasoning in its most classic form. They *feel* the
dual processor system works better. Feel is what junk
scientists must use to prove a point - as if they are in
contact with mystical spirits.

At best, all he can demonstrate is a mild improvement for
some unique desktop applications and only during extreme CPU
intensive processing. In the meantime, he does not even
demonstrate what is necessary to make that dual system
effective - just more missing facts.

Lurkers should note two types of responses. Some agree with
Steve Wolfe et al because of emotional viewpoints - who they
feel is posting nicer. Real people first look for facts and
citation; brutally and aggressively need the irrefutable
fact. Its called reality. Plenty of citations provided that
don't prove anything other than Steve Wolfe, et al have made
claims they cannot support. Just many feelings that are
irrelevant to significant improvement in system performance.

Lane Lewis keeps asking for one simple little thing - the
irrefutable fact. He is not getting it. When will Steve
Wolfe post in specific detail one short fact to prove his
point - with a paragraph to summarize his point? Just more
examples of what junk scientists do: try to confuse the issue,
like a deer caught in headlights, because someone demanded
facts. One response instead was a long useless chain of
posts. No specific fact. Bottom line remains that Steve, et
al provided no relevant fact - just many irrelevant numbers to
confuse the issue - leaving those who demand facts do his work
for him - to wade through that long useless citation.

If 'usability' is a major improvement - then numerical specs
can demonstrate that advantage. I paste wax the car. That
day, the car engine 'feels' smoother. I know paste waxing did
not affect the engine. But the car's motor always *feels*
smoother after a paste wax. Its called 'hands on'
experience. And so we have more proof dual processor
advantage? Its called emotion - the source of so much junk
science. One *feels* it is better - therefore he just knows?
Rubbish.

Another nonsense post: "SMP takes some load of the foirst
CPU and makes it possible for one CPU to deal with real time
data and enables other to go for performance." SMP does not
work that way. But that is proof of why dual processors are
superior? His sentence demonstrates more junk science
reasoning.

"I have experience and you don't. Therefore I am the expert
and you don't know anything." What kind of reasoning is
this? More examples of what junk scientists do - simply
because they have the divine knowledge? Lane asked for
specific facts - and got none. He does not have to prove
anything. Steve Wolfe, et al are making blanket and
emotionally inspired statements they cannot support.

René Descartes did not say "I feel; therefore I am". Steve
Wolfe's reasons are in direct contradiction to those who deal
in reality; not in junk science. And that is what we have in
the computer industry - too many 'experts'who need not first
learn facts - junk scientists.

Just because something like dual processors "can't be
backed-up by a website or a benchmark" *does* make it
probably untrue. No benchmarks are provided by Steve Wolfe
who is selling this dual processor concept just like
Listerene. It must work because "I feel something". Classic
junk science. If that is not obvious to the lurker, then the
lurker should ask whether he is easily made a victim of
propaganda and advertising; things proven simply by emotion.

 

[Lane Lewis]

snip

Hi Lane... you sure got a huge thread going, eh? ;-)

Its not so big by past standards. There were probably more heated threads
over the Buffer statement in DOS. That goes back a few years though.

IDE vs. SCSI
This one is just about over I hope.
Dos vs. Windows
Windows vs. OS2 (The OS Wars)
This was probably the worse time in the computer newsgroups (ECHOs?).
IBM unleashed "Team OS2" on the unsuspecting Win crowd with a vengeance.
They were supposed to help people with OS2 but quickly became some of the
worse thugs in Usenet history. You couldn't post a question about window 3.1
without some member of the team telling you what an idiot you were for using
Windows. This generally started flame wars that went on for weeks if not
months.

PIII vs. Athlon
This started multiple crossposting between AMD and Intel newsgroups and
it was hard to get any other topic discussed
Via chipset vs. all the others
A bad one here because no matter how many problems were pointed out,
some of the VIA crowd insisted they were the best motherboards around.

Lots more that aren't mentioned but it might be better not to :O)

Lane

snip

 

[Lane Lewis]

snip


What CPUs were in those machines and what did you use them for Lane?

I've always read your posts and have thought you to be informed and helpful.
This thread is making me reconsider my opinion of you.

It's becoming an increasingly likely assumption that you mainly pick up your
information from websites and newsgoups you read and pass it on. That, with
a little experience of your own, is not a bad thing in and of itself, and
can help a lot of people who ask questions in newsgroups.

However, websites and benchmarks aimed at, for want of a better word,
'fan-boys', don't always cover real-world computing and are certainly no
substitute for hands-on experience. Just because something can't be
backed-up by a website or a benchmark doesn't make it untrue. "There are
more things in heaven and earth than are dreamt of in your philosophy
Horatio." to quote The Bard.

As an example, albeit a bit tangential, I frequent
alt.comp.hadware.overclockers. Several mods discussed over there include
breaking the pins off CPUs, a move always seen as being irreversible. I
asked for, and was given, a mod method which required just such an action.
However, I managed to break off the wrong pin. I was told by all the
regulars that I was screwed. This was the gospel according to a.c.h.o and
was widely accepted as bring true. Against all advice I attempted, and
succeeded in, soldering the pin back onto the CPU, rendering it functional
again. It's now running at a 50% overclock, running rock-solid and stable
and has been for months. Now the advice usually given over there is to make
sure you break off the right pin as it can be very difficult, if not
impossible to fix if you don't.

Listen to the other contributers in this thread Lane and you just might
learn something valuable. Something beyond the scope of the hardware review
sites and benchmarks.

I have no misconceptions about benchmarks being accurate but they do
give you something to work with. Discussing something without any data at
all and it just becomes a I know better than you argument. Real world test
are the best way to go but that involves a lot of work and unless there's a
clear reason to do so I don't want to get into it.
Hold your opinion into you see how this pans out, we all might learn a
little more about computing and how to carry on a discussion about sensitive
subjects. I spend more of time now in other groups and some of the debaters
there have been doing this since the early 90s and will severely denigrate
you for not following the rules such as backing up any assertions you might
profess to be true. You soon learn not to post what you believe if you don't
have piles of websites that at least seem to agree with some of your
assumption.
This group has changed quite a bit and part of the problem is that I
probably come off as a know it all stranger that has no business telling
anyone about anything, but I weathered thru worse than this and hopefully it
will end with the group being a little better.

As far as the systems I used to have were dual P2s and dual celerons.
They were for a while the fastest machines on the market but once the 533a
and the 600 O/C 900 celerons came out and with their ease at overclocking
they just killed the dual boards with brute force.

You would think that a Dual P2 450 or a dual celeron O/C 550 could keep
up with a single celeron O/C 900 but they couldn't. I always attributed it
to overhead of the OS but I think most of the smp programs were not capable
of taking full advantage of the dual CPUs. So anyway I parted out the dual
machines and have recommended single processors for the desktop ever since.

Lane

 

[Lane Lewis]

snip

Okay, here's a few URLs for you to look at. If you want more, use Google.

http://www.chaosmint.com/benchmarks/powermac-g5-ps7bench/

Not quite a straight comparison because it compares a G5 1.6 GHz with a G5
dual 2.0 GHz, but the dual machine is about twice as fast, sometimes
around 2.5 times as fast.

The athlon tests also show the problems with a dual system. Comparing
the athlon dual 2200 with the athlon 3000 and you see somewhat mixed results
but mostly the dual 2200 is faster. Heres the problem, if we compared an
athlon 3000 system with a dual 1500 system there just would be no comparison
of the two. The 3000 would win hands down, so whats the point of a dual
system if it loses in every benchmark ? If one of the advantages is running
smp programs and it can't even win in that catagory imagine how poorly it
will do with the average desktop program. You'll be running the average
program at half speed and that doesn't even take into consideration the
extra OS overhead. So why is this a good idea for the average destop user
when its not even a good idea for the workstation photoshop user.

Lane

snip

 

[Johan Kullstam]

snip

And even more nonsense, but not even one website, datasheet, review,
anything to back up any of their claims.

Does anyone else start to see a pattern here. The reason they can't provide
any proof is because it doesn't exist. The dual cpu on the desktop is a
joke, it provides less than a ten percent improvement over a same sized
single cpu system and that's only with special programs written for a dual
machine. This has been know for years so why this nonsense keeps popping up
again and again is beyond me.

Up until last year, I was using a quad ppro 200 with 256MB of ram. It
was a tank. Not very fast but it *never* *slows* *down* (at least
interface wise). You can open programs, run jobs in the background,
et cetera and your input is not sluggish, keypress gives immediate
result, your mouse moves, menus open. It just feels more resiliant.
With a single CPU machine (I had a 650MHz Cu-mine PIII at work) things
would get choppy or herky-jerky when under heavy loads.

The only reason I retired it is that it sounde like a jet airplane
trying to take off with all the fans that beast.

I realize that you are skeptical. Sure it doesn't get long haul jobs
any quicker than single CPU MHz would say and memory contention means
that it might be worse. On the other hand, for interactive use, it
just seems to be a tough, solid system when you have more than one CPU.

Everybody here who has used a multi-CPU box knows what I am talking
about and is trying to tell you how it is. You might trying believing
us. We aren't trying to mislead you.

 

[John-Paul Stewart]

snip

The athlon tests also show the problems with a dual system. Comparing
the athlon dual 2200 with the athlon 3000 and you see somewhat mixed results
but mostly the dual 2200 is faster. Heres the problem, if we compared an
athlon 3000 system with a dual 1500 system there just would be no comparison
of the two. The 3000 would win hands down, so whats the point of a dual
system if it loses in every benchmark ?

Who said we're comparing dual CPUs to a single CPU of twice the clock
rate? Go back to the start of this thread. The OP has an HP Kayak
system that'll take CPUs upto 600MHz, and either 1 or 2 of them. His
choice is limited to single 600 vs. dual 600. Dual wins. Whether or
not a single 1200MHz CPU would be better is irrelevant---he'd have to
replace the system to get that and that's not what he was asking about!

BTW, I'm running dual 3.06GHz Xeons. Just what single CPU system is
going to be faster? A single 6.12GHz Xeon???

One of my favourite vendors currently list AthlonXPs upto "3200+" rating
and AthlonMPs upto "2800+". Now I ask you, which is faster: dual 2800
or single 3200? Clearly the dualie wins. Frankly the perfomance of a
dual 1600 has little to no bearing on the discussion at hand.

 

[Steve Wolfe]

No support is provided by Steve Wolfe, et al for these

claims of improvement. Instead of posting facts and numbers,
they discuss some sort of 'usability'. IOW its called junk
science reasoning in its most classic form. They *feel* the
dual processor system works better. Feel is what junk
scientists must use to prove a point - as if they are in
contact with mystical spirits.

What I posted was fact, not opinion or viewpoint. If one machine is bogged
down to where I can't use the interface and the other isn't, that's not
"feeling".

Lane Lewis keeps asking for one simple little thing - the
irrefutable fact. He is not getting it.

And, after all of the times we asked Lane for irrefutable fact, did he
provide it?

When will Steve
Wolfe post in specific detail one short fact to prove his
point - with a paragraph to summarize his point? Just more
examples of what junk scientists do: try to confuse the issue,
like a deer caught in headlights, because someone demanded
facts.

I posted a lot of facts. They were ignored. I talked about context
switching, interrupt floods, and other things. They were all conveniently
dropped on the floor.

If you're going to throw around the term "Junk Scientist", I think it
would apply more to someone who has no experience, and bases his opinions on
what he's read, rather than someone who has years of first-hand experience.

Another nonsense post: "SMP takes some load of the foirst
CPU and makes it possible for one CPU to deal with real time
data and enables other to go for performance." SMP does not
work that way. But that is proof of why dual processors are
superior? His sentence demonstrates more junk science
reasoning.

I certainly didn't say that, and I suspect that nobody else did, either.
I searched groups.google.com for "smp takes some load of the foirst cpu
group:comp.os.linux.hardware" in comp.os.linux. hardware, and yours was the
only post to come up. I fixed your spelling error, and again, your message
was the only one to come up.

So, you're attributing a direct quote that was, in fact, never made. That
makes it awfully hard to take you seriously.

"I have experience and you don't. Therefore I am the expert
and you don't know anything." What kind of reasoning is
this? More examples of what junk scientists do - simply
because they have the divine knowledge?

Again, that quote was never made. Are you sure that you aren't Lane in
disguise?

Lane asked for
specific facts - and got none.

Bull. He got them, and ignored them.

If you don't believe my claims, then you are free to follow your own
advice, and provide some non-junk science disproving them. Otherwise,
you're just a pot calling the kettle black.

steve

 

[Steve Wolfe]

The athlon tests also show the problems with a dual system.
Comparing

the athlon dual 2200 with the athlon 3000 and you see somewhat mixed results
but mostly the dual 2200 is faster. Heres the problem, if we compared an
athlon 3000 system with a dual 1500 system there just would be no comparison
of the two. The 3000 would win hands down, so whats the point of a dual
system if it loses in every benchmark ?

In terms of a single task, yes, like I've been saying, the 3000 would win.
I've never argued that at all. I've said since that sort of thing since the
first of the discussion, like when I pointed out that if I wanted to play a
3D game, I'd never choose the SMP system. Why do you keep beleagering the
point?

steve

 

[~misfit~]

I have no misconceptions about benchmarks being accurate but they do
give you something to work with. Discussing something without any data at
all and it just becomes a I know better than you argument. Real world test
are the best way to go but that involves a lot of work and unless there's a
clear reason to do so I don't want to get into it.
Hold your opinion into you see how this pans out, we all might learn a
little more about computing and how to carry on a discussion about sensitive
subjects. I spend more of time now in other groups and some of the debaters
there have been doing this since the early 90s and will severely denigrate
you for not following the rules such as backing up any assertions you might
profess to be true. You soon learn not to post what you believe if you don't
have piles of websites that at least seem to agree with some of your
assumption.
This group has changed quite a bit and part of the problem is that I
probably come off as a know it all stranger that has no business telling
anyone about anything, but I weathered thru worse than this and hopefully it
will end with the group being a little better.

As far as the systems I used to have were dual P2s and dual celerons.
They were for a while the fastest machines on the market but once the 533a
and the 600 O/C 900 celerons came out and with their ease at overclocking
they just killed the dual boards with brute force.

You would think that a Dual P2 450 or a dual celeron O/C 550 could keep
up with a single celeron O/C 900 but they couldn't. I always attributed it
to overhead of the OS but I think most of the smp programs were not capable
of taking full advantage of the dual CPUs. So anyway I parted out the dual
machines and have recommended single processors for the desktop ever

since.

Thanks for the reply Lane.

Funny you should mention celeron 600s running at 900. It was a celeron 600
that I broke the wrong pin off, in the process of getting it to run in a
non-coppermine slocket. Soldered it back on, broke off the correct pin,
jumpered the back of the socket with a strand of IDE cable and it's running
perfectly at 900Mhz now. Still quite usuable too, although not my main
machine.

 

[w_tom]

Instead of posting repeatedly what you did post (a waste of
bandwidth), instead just post the one fact. I don't care how
you 'feel' that one machine performs. That is only an opinion
based upon an emotion. That emotion does not prove dual
processor machines are superior for workstations. Where is
the irrefutable facts - basic specification - the numbers? I
don't even see your list of workstation programs that operate
nearly twice as fast due to multithreading. I do see Lane
Lewis's summary of where dual processors should provide
utility. But Lane is not making your unsubstantiated claims -
claims made without relevant numbers.

If you posted lots of facts, then post the 'best one' right
here and now. A single irrefutable fact. One simple
paragraph at the very top of your reply will do just fine
stating why dual processors are so superior and numbers that
demonstrate that superiority. Not reams of speculative
comments, provided without numbers, and intertwined with my
post to confuse the issue and to make those posts difficult to
read. Please don't post more URLs that only vaguely relate to
your claim. Where is that best irrefutable fact and
supporting numbers? Simply state the summary of your claim and
provide supporting digits.

 

[w_tom]

Do you happen to know what the function or signal name for
that pin was? Rather interesting experiment.

 

[Gregory L. Hansen]

In terms of a single task, yes, like I've been saying, the 3000 would win.
I've never argued that at all. I've said since that sort of thing since the
first of the discussion, like when I pointed out that if I wanted to play a
3D game, I'd never choose the SMP system. Why do you keep beleagering the
point?

It's been explained to me that games tend not to be written for multiple
processors because they tend to be written for home versions of Windows,
which doesn't support multiprocessing at all. So naturally you get little
advantage from the second processor, aside from handling interruptions
from the OS. But that has nothing to do with "overhead", rather with the
software not even trying to use the second processor. I suspect that's
the sort of computing that Lane Lewis is drawing on for his opinion, and
then just generalizing it to anyone doing anything on anything.

But in one web page I found on game benchmarks the author said that
multiprocessing in Apple's G5 gave such a high frame rate that he didn't
believe Apple's number until he tried it himself and got a slightly higher
number than they did. I couldn't find stats that clearly (to me) showed
frame rate with single and dual processors with all else equal, so I
didn't post a link. But it seems multiple processors *could* speed up a
game if the software supports it. And if the next version of Windows
supports multiprocessing in both the pro and home versions (I don't
keep up with Windows so I'll have to take my friend's word for it), we'll
see how important that second processor becomes to gamers after that.

 

[Jeroen Geilman]

One of my favourite vendors currently list AthlonXPs upto "3200+"
rating
and AthlonMPs upto "2800+". Now I ask you, which is faster: dual
2800
or single 3200? Clearly the dualie wins.

I'm smelling a bit of an Intel bias here...

Both Athlon XP and Athlon MP are fully dual-processor capable.

The difference (AFAIAA) is that the MP can be used in N-way
multiprocessing systems, while for the XP 2 CPUs is the limit.

So you could run the XP3200+ in a dual system - Tyan makes the best
MoBo's for this.

 

[Gregory L. Hansen]

Who said we're comparing dual CPUs to a single CPU of twice the clock
rate? Go back to the start of this thread. The OP has an HP Kayak
system that'll take CPUs upto 600MHz, and either 1 or 2 of them. His
choice is limited to single 600 vs. dual 600. Dual wins. Whether or
not a single 1200MHz CPU would be better is irrelevant---he'd have to
replace the system to get that and that's not what he was asking about!

I think the thread is no longer about the original poster at all. But
you're right, PIII 600 is as high as I can go. I already have a PII 400,
so my choice is really single or dual PII 400, or single or dual PIII 600
or 550 or something. For two PIII's and a VRM I'm aiming for around $90
including shipping, half of that cost due to the VRM (thanks a lot, HP).
Maybe I'll spend $10 on a second PII while waiting for deals on PIIIs to
appear on eBay. I'm not really in any hurry.

There's the other opinion that a motherboard and newer processor can be
had for $150 or so. And I'd have to get a case and power supply because I
have an HP ("Highly Proprietary"?), could be another $30 not including
shipping, maybe cheaper at a computer show. And then I could in principle
port everything else -- disk drives and memory and things. But it's not
really just about maximizing performance. I *like* the Kayak, I want to
keep it intact. I'm just wondering how to get the best I can out of it,
but I don't really expect to be doing anything that will overly tax it in
the near future. I won't be running Windows on it, for instance.