PII vs PIII

[Steve Wolfe]

And more nonsense
Again do you have anything that shows that they are even remotely close
in benchmarks even while multitasking. Dual machines make good servers and
ok single app workstations but they are terrible desktop machines.

Please, tell me how many dual CPU systems you use as desktops. I've got
quite a few here. I'll bet you're talking out of your butt.

The vast
majority of software is designed for single processors and are virtually
worthless on a dual machine and it's getting worse every day as software
developers code for the P4 single CPU.

Hahahahahaha! You don't have a clue how many things are actually vying
for CPU time, do you? And you don't have a clue about interrupts, do you?

To sit there and wait for a program
that can't take advantage of the extra CPU doesn't make any sense with
today's prices.

You're telling me that when you run X, and you have the X server trying
to get CPU time, your app trying to get CPU time, your kernel using CPU
time for disk I/O, your kernel using CPU time for network I/O, and quite a
few others, that one CPU is going to get left idle? Like I said, you've
never used a dual-CPU desktop, have you? Go home and play with your toys.

steve

 

[Steve Wolfe]

So your trying to tell me that dual 600 system is better than a single
1.4

gig system when every benchmark in the world shows a huge gap between the
two. When every game plays twice as fast on the celeron, when photoshop,
excel, word, music software etc perform at least 40 to 60 percent better
maybe even more. Do you have anything at all that would back up what you
just said.

You skipped over or ignored every point I made. If you like to live
your life looking at people's benchmarks, and deciding that because
someone got a 1% higher FPS score on their favorite game, or that
Photoshop finished 10% faster, then by all means, keep doing it.

I don't live like that. I judge the machine on how much it gets done
for me while I'm using it. My single-CPU machines don't get as much done
as my dual-CPU machines. I have single-CPU desktops from a P2/233 to an
AthlonXP 2700+. I have dual-CPU desktops from a dual P-133 to a dual
2.8-GHz Xeon. Desktop systems, mind you, not servers. And I can tell you
from experience that the dual-CPU desktops let you get more done.

Now, like I said, if I were going to play 3D games all day, a single-CPU
machine would do at least as well, usually better. If I had other
computationally-bound tasks (like applying Photoshop filters to
30-megabyte images), I'd go for the single-CPU system, and sit on my
thumbs while I waited.

I, for one, would *much* rather have a system that's responsive under
load, which single-CPU systems are *not*. When I launch the
computationally-bound processes, yes, they take longer if I'm using slower
CPU's. But while they're working, I can continue on other tasks just as
if there were nothing on the system.

I have a feeling that you don't keep up much on kernel development. If
you had, you would realize that one of the largest pushes in the 2.6
series has been to increase responsivness under load, because with
single-CPU systems, you just don't have it. They've been jumping through
hoops for months so that rendering pages in mozilla while you listen to
..mp3's doesn't produce any audio crackle. Guess what, I'll let you in on
a secret: My dual CPU desktops (even the P133) haven't ever had that
problem. And I'll let you in on another secret: The reasons don't have a
thing to do with CPU speed, or elapsed time in benchmarks.

As in other messages, I ask you: How many dual-CPU desktop systems do
you use day in and day out?

steve

 

[Stacey]

It looks like the machine I have (an HP Kayak XAS, you may have seen me
mention it already) can be upgraded from a PII 400MHz to dual PIII 600MHz.
But the PIII's are substantially more expensive than the PII's, around $70
each compared to $10 each.

For $150 you can get an XP tbred chip, board and DDR ram that will absolutly
KILL a dual P3 system.

 

[SIOL]

Please, tell me how many dual CPU systems you use as desktops. I've got
quite a few here. I'll bet you're talking out of your butt.


Hahahahahaha! You don't have a clue how many things are actually vying
for CPU time, do you? And you don't have a clue about interrupts, do you?


You're telling me that when you run X, and you have the X server trying
to get CPU time, your app trying to get CPU time, your kernel using CPU
time for disk I/O, your kernel using CPU time for network I/O, and quite a
few others, that one CPU is going to get left idle? Like I said, you've
never used a dual-CPU desktop, have you? Go home and play with your toys.

steve

My two centiEuros go with Steve's perspective. I have several machines in my
workshop and I can tell you that dual PIII- 1 GHz feel much more responsive
than one Tualatin at 1.7 GHz. Yeah, there could be overhead, but dual CPU is
certainly worth it.

Thing is, almost all benchmarks measure "linear speed", something like drag
racing, they don't take into account real traffic performance with curves,
crossroads etc. With mothern machines, task switching, intetrrupts etc. take
considerable efforts and time.
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.

ALL my machines will be at least dual CPU in future (workstations, servers
etc). I'm still waiting for a decent dual CPU Opteron board for good
price...

 

[kony]

On Fri, 10 Oct 2003 18:31:51 +0000 (UTC),

I didn't realize there was such a difference between a PIII 600MHz 512K
cache and a PIII 600MHz 256K cache. Are all PIIIs with 512K Katmais, and
all PIIIs with 256K coppermines?

Regarding the previous paragraph this is in reply to, you could check
the motherboard manual or spec sheet if available, or take the numbers
off the voltage regulator controller chip, head to the respective
manufacturer's website, check the datasheet for the voltage levels it
supports. For an extra level of certainty you could take a continuity
meter and check the corresponding VID pins on that regulator with the
VID pins in the socket (with CPU removed), as per the socket pin
functions seen on Intel's CPU data sheet for the CPU. Hmm, I just
remembered, it's a slot one board... same procedure but checking
continuity at the slot contact instead of a socket pin.

Then there's the BIOS support... see if there's a setting in the BIOS
for "Halt on CPU error" or similar. If so, disable it. You may need
to update the bios, and in rare cases (some Intel boards) you don't ev

 

[kony]

Continuing where I left off in the previous reply, the rest is
embedded below:

Regarding the previous paragraph this is in reply to, you could check
the motherboard manual or spec sheet if available, or take the numbers
off the voltage regulator controller chip, head to the respective
manufacturer's website, check the datasheet for the voltage levels it
supports. For an extra level of certainty you could take a continuity
meter and check the corresponding VID pins on that regulator with the
VID pins in the socket (with CPU removed), as per the socket pin
functions seen on Intel's CPU data sheet for the CPU. Hmm, I just
remembered, it's a slot one board... same procedure but checking
continuity at the slot contact instead of a socket pin.

Then there's the BIOS support... see if there's a setting in the BIOS
for "Halt on CPU error" or similar. If so, disable it. You may need
to update the bios, and in rare cases (some Intel boards) you don't

even want to use the last BIOS updates, but rather a few earlier from
the last, as Intel "caused" the board to halt while previous BIOS
updates worked.

To use a different VRM without solid evidence that it's compatible
you'd need to trace the circuits and determine if it has a compatible
pinout, appropriate voltage support, and can supply needed amperage.

It really is a lot of work and expense for so little gain, a single
modern CPU would run circles around whichever CPU(s) you can get
working, and would support more memory, less expensively, in addition
to a lot of other modern features lacking in a BX board.


Dave

 

[Lane Lewis]

snip

You're telling me that when you run X, and you have the X server trying
to get CPU time, your app trying to get CPU time, your kernel using CPU
time for disk I/O, your kernel using CPU time for network I/O, and quite a
few others, that one CPU is going to get left idle? Like I said, you've
never used a dual-CPU desktop, have you? Go home and play with your toys.

steve

I'm telling you for the second time here to back up your statements. Why did
you ignore the first request. Are you one of those people who thinks they
can make any claim they want and never have to support those claims.
Theories are nice, facts are better.

Either back up what you say or retract it, it's that simple.

Lane

 

[Lane Lewis]

So your trying to tell me that dual 600 system is better than a single 1.4

You're missing the point. I'm not talking about benchmarks. I'm talking
about usability.

Have YOU used a dual-CPU desktop? I didn't think so. Now go back home
and play with your toys.

steve

Nice dodge

now you can show me some "usability" tests or retract your statements.

Lane

 

[Lane Lewis]

You skipped over or ignored every point I made. If you like to live
your life looking at people's benchmarks, and deciding that because
someone got a 1% higher FPS score on their favorite game, or that
Photoshop finished 10% faster, then by all means, keep doing it.

I don't live like that. I judge the machine on how much it gets done
for me while I'm using it. My single-CPU machines don't get as much done
as my dual-CPU machines. I have single-CPU desktops from a P2/233 to an
AthlonXP 2700+. I have dual-CPU desktops from a dual P-133 to a dual
2.8-GHz Xeon. Desktop systems, mind you, not servers. And I can tell you
from experience that the dual-CPU desktops let you get more done.

Now, like I said, if I were going to play 3D games all day, a single-CPU
machine would do at least as well, usually better. If I had other
computationally-bound tasks (like applying Photoshop filters to
30-megabyte images), I'd go for the single-CPU system, and sit on my
thumbs while I waited.

I, for one, would *much* rather have a system that's responsive under
load, which single-CPU systems are *not*. When I launch the
computationally-bound processes, yes, they take longer if I'm using slower
CPU's. But while they're working, I can continue on other tasks just as
if there were nothing on the system.

I have a feeling that you don't keep up much on kernel development. If
you had, you would realize that one of the largest pushes in the 2.6
series has been to increase responsivness under load, because with
single-CPU systems, you just don't have it. They've been jumping through
hoops for months so that rendering pages in mozilla while you listen to
.mp3's doesn't produce any audio crackle. Guess what, I'll let you in on
a secret: My dual CPU desktops (even the P133) haven't ever had that
problem. And I'll let you in on another secret: The reasons don't have a
thing to do with CPU speed, or elapsed time in benchmarks.

As in other messages, I ask you: How many dual-CPU desktop systems do
you use day in and day out?

steve

The computers I own are none of your business.

Back up your statements or retract them its that simple. If you can't back
up what you say then don't say it to begin with or at least have the honesty
to say its your opinion.

So where's the data that will back up anything you've said.

Lane

 

[Lane Lewis]

snip

My two centiEuros go with Steve's perspective. I have several machines in my
workshop and I can tell you that dual PIII- 1 GHz feel much more responsive
than one Tualatin at 1.7 GHz. Yeah, there could be overhead, but dual CPU is
certainly worth it.

Thing is, almost all benchmarks measure "linear speed", something like drag
racing, they don't take into account real traffic performance with curves,
crossroads etc. With mothern machines, task switching, intetrrupts etc. take
considerable efforts and time.
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.

ALL my machines will be at least dual CPU in future (workstations, servers
etc). I'm still waiting for a decent dual CPU Opteron board for good
price...

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

 

[SIOL]

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

When your knowledge level is so low, anything is beyond you.

What kind of proof do you want from me ?
I have single CPU and dual CPU systems. Latter work much better for me,
especially under Linux.
Difference is so obvious I don't even bother quantifying it. Besides, what
benchmark should I use ?
And why ? Real experience is what counts for me.

I need benchmarks just for the things that I can not evaluate from real
experience. And even then, I have to be carefull which benchmark to use as a
guide.

So, what would be proof for you ? Mpeg file showing that I really can do all
that on this system at the same time ?

I am aware of pale results of dual CPU systems on classic benchmarks.
Judging the system by them, one would never opt for dual CPU system. It just
doesn't make sense. Usually very expensive board, expensive SMP CPUs that
are often a step or more behind leading single CPUs AND much more expensive
etc...

But just try it once (O.K. for a few days or so ) on desk and you would
never want to leave it for uniCPU machine.

My next personal choice would be an early Tyan board fitted with two Bartons
(if I could get BArtons to work with it) or XPs. BArton would be much better
due to bigger L2, but Xp would be nice also. Note that early Tyans can work
with XPs and MPs alike...

But since Opterons are around corner, I'll better wait for one decent board
and stick two 24x in it...

 

[Gregory L. Hansen]

And more nonsense
Again do you have anything that shows that they are even remotely close
in benchmarks even while multitasking.

Why do you keep asking about benchmarks? Who said anything about
benchmarks? Dual processor machines are about twice as fast on Photoshop
filters than a single processor of the same type, those sorts of stats
get thrown around comp.sys.mac.advocacy now and then. But we're not
talking about number crunching speed.

Dual machines make good servers and
ok single app workstations but they are terrible desktop machines.

Everyone I've talked to (including people in real life) seem to think that
dual processor machines make wonderful desktop machines -- smoother and
more responsive. And faster. Certainly faster even for a single
application if it's multithreaded, but also just because it doesn't have
the operating systme interrupting it all the time.

Two processors versus a single processor of the same type, of course. Two
processors of half the clock speed as the single processor may or may not
be as fast depending on things like how well factored the code is and how
important it is to have twice the cache. It depends on what you're doing
with it.

I happen to have a machine sitting beside me that I can plug a second
processor into. Can't think of a reason not to.

 

[Mike Walsh]

Even the single processor on my home PC is inhibited by the relatively slow memory speed on memory intensive applications, e.g. image editing, so I imagine it would be much worse with dual processors.

 

[Lane Lewis]

snip

Why do you keep asking about benchmarks? Who said anything about
benchmarks? Dual processor machines are about twice as fast on Photoshop
filters than a single processor of the same type, those sorts of stats
get thrown around comp.sys.mac.advocacy now and then. But we're not
talking about number crunching speed.


Everyone I've talked to (including people in real life) seem to think that
dual processor machines make wonderful desktop machines -- smoother and
more responsive. And faster. Certainly faster even for a single
application if it's multithreaded, but also just because it doesn't have
the operating systme interrupting it all the time.

Two processors versus a single processor of the same type, of course. Two
processors of half the clock speed as the single processor may or may not
be as fast depending on things like how well factored the code is and how
important it is to have twice the cache. It depends on what you're doing
with it.

I happen to have a machine sitting beside me that I can plug a second
processor into. Can't think of a reason not to.

Great even more claims and nothing to back them up.

[qt]
"Dual processor machines are about twice as fast on Photoshop filters
than a single processor of the same type."
[end]

This one here is blatantly false, nothing runs twice as fast on a dual
machine due to overhead. At best you might see a 10 percent improvement and
a 1.2 gig single will smoke a dual 600 in loading and processing Photoshop
filters.

Are you starting to get the picture yet. Why are there all these claims
about dual CPU systems and nothing to confirm them with. The fact is their
slow and not worth the money spent for the desktop regardless of what is
said about them.

Lane

 

[SIOL]

Still no benchmarks
Still no real world tests
Still no stats
Still no reviews

Just your anecdotal evidence, which judging from your insulting childish
behavior is worth about nothing.

Lane

I don't feel like making too much effort for you.
You'll have to live just "with my anecdotal evidence".
If you find it "worth nothing" then please ignore it and find someone
willing to do tests for you...

Good luck.

 

[Lane Lewis]

When your knowledge level is so low, anything is beyond you.

What kind of proof do you want from me ?
I have single CPU and dual CPU systems. Latter work much better for me,
especially under Linux.
Difference is so obvious I don't even bother quantifying it. Besides, what
benchmark should I use ?
And why ? Real experience is what counts for me.

I need benchmarks just for the things that I can not evaluate from real
experience. And even then, I have to be carefull which benchmark to use as a
guide.

So, what would be proof for you ? Mpeg file showing that I really can do all
that on this system at the same time ?

I am aware of pale results of dual CPU systems on classic benchmarks.
Judging the system by them, one would never opt for dual CPU system. It just
doesn't make sense. Usually very expensive board, expensive SMP CPUs that
are often a step or more behind leading single CPUs AND much more expensive
etc...

But just try it once (O.K. for a few days or so ) on desk and you would
never want to leave it for uniCPU machine.

My next personal choice would be an early Tyan board fitted with two Bartons
(if I could get BArtons to work with it) or XPs. BArton would be much better
due to bigger L2, but Xp would be nice also. Note that early Tyans can work
with XPs and MPs alike...

But since Opterons are around corner, I'll better wait for one decent board
and stick two 24x in it...

Still no benchmarks
Still no real world tests
Still no stats
Still no reviews

Just your anecdotal evidence, which judging from your insulting childish
behavior is worth about nothing.

Lane

 

[John-Paul Stewart]

This one here is blatantly false, nothing runs twice as fast on a dual
machine due to overhead. At best you might see a 10 percent improvement and
a 1.2 gig single will smoke a dual 600 in loading and processing Photoshop
filters.

If you're so adamant that others post evidence, it is high time for you
to do likewise. You keep making claims with *nothing* to back them up.
Prove your case (thereby disproving ours) by posting *your* evidence.

As to my own evidence that dual CPU's are faster than a single one, let
us take a look at kernel compile times on a dual 3.06GHz Xeon system.
Using only one of the two CPUs it took 2mins 42sec to compile my test
kernel. Using both CPUs, the compile time dropped to 1min 29sec.
That's about an 80% speedup on that task. Get it?

 

[Lane Lewis]

I don't feel like making too much effort for you.
You'll have to live just "with my anecdotal evidence".
If you find it "worth nothing" then please ignore it and find someone
willing to do tests for you...

Good luck.

Already did the tests.

Lane

 

[Michael Squires]

My two centiEuros go with Steve's perspective. I have several machines in my
workshop and I can tell you that dual PIII- 1 GHz feel much more responsive
than one Tualatin at 1.7 GHz. Yeah, there could be overhead, but dual CPU is
certainly worth it.

My experience with a home machine based on the SuperMicro P6DGH MB
with dual PIII/850's compared to my office machine with PIV 1.4GZ is the
same. The home system also uses a SCSI disk system, which was much
cheaper when using surplus disks when it was put together two years ago.

The P6DGH is an example of a motherboard built for Katmai CPUs but which
can, in some versions, run Coppermines. The differences aren't very obvious.

Mike Squires

 

[Steve Wolfe]

The computers I own are none of your business.

Back up your statements or retract them its that simple. If you can't back
up what you say then don't say it to begin with or at least have the honesty
to say its your opinion.

Please follow your own advice, and tell me how many dual-CPU desktops you
use.

None, right?

steve