Xeon 533 FSB or P4 800 FSB ? (Corrected post)

[kony]

Right, so please quote your benchmarks measuring things like the delay
between keypress and matching change on the screen. I want to see your
numbers on time to display a complex web page, either lots of images or
better yet a mix of images and JAVA.

LOL. The delay between keypresses on a properly running single-CPU
system, is lower than a human can discriminate. Is there some reason
you need the keys to work faster than any human can perceive, let
alone type?

Complex web page display is bottlenecked by network/internet
throughput, then HDD subsystem for the cache, not the CPU, not on a
single or dual CPU system.

It is certainly possible that if you could rig some benchmark to
display webpages over and over as fast as possible, the dual CPU
system would win. This wouldn't correspond to real-world usage.

It's not the climate control or ride which make riding in a Bentley
nicer than a Yugo, either, but how fast you get to the end of the
journey is not why people by Bentleys. It's just *nicer* to use an SMP
machine, less frustrating. Things do actually do get done sooner (as
above) but that's just not the point.

A grand theory except that the single CPU system could likewise be
considered better for the same BS comparison.

I've invited you to show any benchmarks indicating that responsiveness
is as good with uni as smp. Ball's in your court. You might see if you
can find the IBM study of productivity vs. response time, my (hard) copy
is in a box in another state, so I won't quote from memory.

It's quite simple, TODAY'S systems respond faster than a human. There
is no possible way productivity can be increased with a dual CPU in
situations were the human input is the bottleneck. This is a
different era, we're beyond the point where a large percentage of CPU
time was spent mucking around with the GUI and office apps, a modern
single CPU has plenty of time to spare.

Response time is not usually dictated by the number of processors
either, most often the network or HDD.

It would appear that advocates of dual CPU systems, haven't even had
the pleasure of working on a properly configured high-end single CPU
system. It has none of the delays being implied.

I asked a simple question in a prior POST, what is this resonably
common use of a system where there's any lag? You want me to believe
it's inherant in a single CPU system to see lag, yet there isn't any
I've seen in my own use or observing many other users. Give a
concrete example that would make it worthwhile beyond very atypical
usage patterns or workstation-like usage of applications specifically
designed for duals.


Dave

 

[kony]

If you can't notice the difference between uni and smp, please buy
something cheap and slow and don't try to stop others who are more
perceptive from benefitting from a more responsive system. You clearly
don't do more than one thing at a time, or never noticed that delay
between pressing a key and seeing a character echo.

On the contrary, I notice that often the "uni" is often faster and
smoother given same budget. Vaguely implying that there's a lag just
doesn't cut it, there are PLENTY of examples of users in this very
forum who have systems that don't have these significant lags, ie-
unresponsiveness.

I don't know if you can't tell the difference or just want to cast
aspersions on your betters (well, your computer's betters) out of envy,
but this tirade claiming that slow is beautiful is getting tiresome.

I suggest you learn a little system troubleshooting, so you'll better
understand why there are lags, when there are, and how to fix them
instead of jumping to conclusions about uniprocessor systems.

Once you learn how to do that, if you insist on using an SMP systems
they'll have even better performance.


Dave

 

[Nate Edel]

Systems are multi-tasking but almost always, only one user. That user
is occupied with a single application at a time,

Assuming facts not in evidence. A LOT of power users will start one
application in one window, and alt-tab to another window while it runs.
Doing that twice or three times in a row gets you a degree of task switching
where a 2xSMP system _may_ make sense.

 

[bill davidsen]

| On Sat, 27 Dec 2003 11:55:31 -0500, Bill Davidsen <[email protected]>
| wrote:
|
| >kony wrote:
| >
| >> "Real-world observations" are only as good as the data that supports
| >> them. Your word is not enough to overcome a staggering number of
| >> benchmarks of real applications, often running in these "half-a-dozen
| >> user applications, a dozen system services" environments that you're
| >> trying to differentiate. All those services and apps in the
| >> background "usually" don't need a lot of CPU time, they'd run on a
| >> Pentium 200 and certainly on a single CPU that's significantly faster
| >> than either one of two in a dual CPU system.
| >
| >Right, so please quote your benchmarks measuring things like the delay
| >between keypress and matching change on the screen. I want to see your
| >numbers on time to display a complex web page, either lots of images or
| >better yet a mix of images and JAVA.
| >
|
| LOL. The delay between keypresses on a properly running single-CPU
| system, is lower than a human can discriminate. Is there some reason
| you need the keys to work faster than any human can perceive, let
| alone type?

I can perceive 1500-2000ms nicely, which is bout the delay in Word
running realtime spell check with some other stuff going in the
background. No it doesn't happen all the time, yes it is a reasonably
typical user thing to do, at least in a business environment.
|
| Complex web page display is bottlenecked by network/internet
| throughput, then HDD subsystem for the cache, not the CPU, not on a
| single or dual CPU system.

Don't have JAVA applets, do you?
|
| It is certainly possible that if you could rig some benchmark to
| display webpages over and over as fast as possible, the dual CPU
| system would win. This wouldn't correspond to real-world usage.

Using CPU bound application is most definitely a real-world occurence.
Think Word, think JAVA, think photoshop. Now think print driver in the
background sending graphical data (web page with pictures, etc) to the
printer. You can see and feel the benefit.
|
|
| >> It is not a fluid-like sequence of execution that gets the job done
| >> faster, it's actual performance at each and every application, when
| >> needed, one at a time then switching tasks faster than the user can
| >> even perceive it. If all you want is a smooth experience regardless
| >> of the actual performance, I suggest you take some valium. On the
| >> other hand, if you want peak performance for the actual jobs you're
| >> running, no subjective opinion can have more weight than actual
| >> benchmarks of that application, in the same environment in which it's
| >> to be running.
| >
| >It's not the climate control or ride which make riding in a Bentley
| >nicer than a Yugo, either, but how fast you get to the end of the
| >journey is not why people by Bentleys. It's just *nicer* to use an SMP
| >machine, less frustrating. Things do actually do get done sooner (as
| >above) but that's just not the point.
|
| A grand theory except that the single CPU system could likewise be
| considered better for the same BS comparison.

For definitions of better including jerky, unresponsive, and similar? I
fail to find a definition of better which does favor smooth response. I
can't think of any way uni would be "better" other than "cheaper."

| >I've invited you to show any benchmarks indicating that responsiveness
| >is as good with uni as smp. Ball's in your court. You might see if you
| >can find the IBM study of productivity vs. response time, my (hard) copy
| >is in a box in another state, so I won't quote from memory.
|
| It's quite simple, TODAY'S systems respond faster than a human. There
| is no possible way productivity can be increased with a dual CPU in
| situations were the human input is the bottleneck. This is a
| different era, we're beyond the point where a large percentage of CPU
| time was spent mucking around with the GUI and office apps, a modern
| single CPU has plenty of time to spare.

If you can't perceive delays they for you they don't exist. That doesn't
mean others don't notice them.

| I asked a simple question in a prior POST, what is this resonably
| common use of a system where there's any lag? You want me to believe
| it's inherant in a single CPU system to see lag, yet there isn't any
| I've seen in my own use or observing many other users. Give a
| concrete example that would make it worthwhile beyond very atypical
| usage patterns or workstation-like usage of applications specifically
| designed for duals.

I've listed a number of examples, you obviously don't find them common,
so I'm bowing out because I think you're just trolling.

--
Bill Davidsen <[email protected]> CTO, TMR Associates
As we enjoy great advantages from inventions of others, we should be
glad of an opportunity to serve others by any invention of ours; and
this we should do freely and generously.
-Benjamin Franklin (who would have liked open source)

 

[kony]

| LOL. The delay between keypresses on a properly running single-CPU
| system, is lower than a human can discriminate. Is there some reason
| you need the keys to work faster than any human can perceive, let
| alone type?

I can perceive 1500-2000ms nicely, which is bout the delay in Word
running realtime spell check with some other stuff going in the
background. No it doesn't happen all the time, yes it is a reasonably
typical user thing to do, at least in a business environment.

I'd say spell-checking in Word is pretty typical in any environment,
and it doesn't cause 1500ms delay on any systems I've seen in the past
2 years, that is, 2 yr old or newer systems. Fix the problem, it
isn't the number of CPUs.

| Complex web page display is bottlenecked by network/internet
| throughput, then HDD subsystem for the cache, not the CPU, not on a
| single or dual CPU system.

Don't have JAVA applets, do you?

Is it even possible to browse the internet without running across a
zillion of them these days?

I don't understand why it isn't sinking in, but these things do not
cause desitation on a properly running, modern uniprocessor system.

|
| It is certainly possible that if you could rig some benchmark to
| display webpages over and over as fast as possible, the dual CPU
| system would win. This wouldn't correspond to real-world usage.

Using CPU bound application is most definitely a real-world occurence.
Think Word, think JAVA, think photoshop. Now think print driver in the
background sending graphical data (web page with pictures, etc) to the
printer. You can see and feel the benefit.

Perhaps there is some horrific Winprinter somewhere with a driver that
chokes on images, but I've never seen it, and print plenty of
webpages/documents/etc. If you have software-driver hardware that's
THAT bad, I'll grant that a 2nd CPU will help, but I'd go for a better
printer myself.

For definitions of better including jerky, unresponsive, and similar? I
fail to find a definition of better which does favor smooth response. I
can't think of any way uni would be "better" other than "cheaper."

Just for the heck of it I'll mention it one last time:
These jerky, unresponsive systems you've seen, are misconfigured.

You can't think of any ways because you're ignoring the benchmarks
that prove a uniprocessor system can and does have higher performance
for many uses.

| >I've invited you to show any benchmarks indicating that responsiveness
| >is as good with uni as smp. Ball's in your court. You might see if you
| >can find the IBM study of productivity vs. response time, my (hard) copy
| >is in a box in another state, so I won't quote from memory.
|
| It's quite simple, TODAY'S systems respond faster than a human. There
| is no possible way productivity can be increased with a dual CPU in
| situations were the human input is the bottleneck. This is a
| different era, we're beyond the point where a large percentage of CPU
| time was spent mucking around with the GUI and office apps, a modern
| single CPU has plenty of time to spare.

If you can't perceive delays they for you they don't exist. That doesn't
mean others don't notice them.

You claim delays that are clearly large enough to be noticed by myself
and others,, yet they don't exist on a properly working system. The
answer is to correctly diagnose the problem then fix the problem, not
throw another CPU at it and jump to conclusions.

| I asked a simple question in a prior POST, what is this resonably
| common use of a system where there's any lag? You want me to believe
| it's inherant in a single CPU system to see lag, yet there isn't any
| I've seen in my own use or observing many other users. Give a
| concrete example that would make it worthwhile beyond very atypical
| usage patterns or workstation-like usage of applications specifically
| designed for duals.

I've listed a number of examples, you obviously don't find them common,
so I'm bowing out because I think you're just trolling.

No, you never did give an example that showed these delays you imply
are inherant in a uniprocessor system. As soon as you give an example
that is both reproducible and remotely likely usage pattern, I'll
easily agree, at least for that specific use. The key here is
"reproducible"... you can claim whatever you like, but if it can't
even be reproduced on most of not all uniprocessor systems, it's
pretty clear that the number of CPUs wasn't the commonality, not the
cause of the problems you report.

 

[Nate Edel]

| LOL. The delay between keypresses on a properly running single-CPU
| system, is lower than a human can discriminate. Is there some reason
| you need the keys to work faster than any human can perceive, let
| alone type?

I can perceive 1500-2000ms nicely, which is bout the delay in Word
running realtime spell check with some other stuff going in the
background. No it doesn't happen all the time, yes it is a reasonably
typical user thing to do, at least in a business environment.

1500-2000ms is 1 1/2 to 2 _seconds_ and if word is lagging for that long
except at start up, something is not working right with your copy of Word or
with your system overall.

Don't have JAVA applets, do you?

Properly functioning Java applets very rarely use close to 100% of CPU. I
seem them barf and hang pretty often, which can eat up 100% of CPU, but
duals aren't going to fix that.

 

[Nate Edel]

I'd say spell-checking in Word is pretty typical in any environment,
and it doesn't cause 1500ms delay on any systems I've seen in the past
2 years, that is, 2 yr old or newer systems. Fix the problem, it
isn't the number of CPUs.

Spell check in the background didn't cause delays like that on Word 2000 on
my Pentium-MMX 233 laptop, so it certainly shouldn't on a modern system as
you say.

You can't think of any ways because you're ignoring the benchmarks
that prove a uniprocessor system can and does have higher performance
for many uses.

A better examples of things that duals can help with is software
development. Do a big compilation job, and try to do editing or work in
word or even browsing while it's building. HT helps, duals help more, with
keeping the second application responsive while the build is going on.

That's an everyday occurrence at my job. Is that going to be true for
average home users? No, but average home users tend to be fine with Durons
and Celerons to begin with.

Another example: doing research, I'll often use "Open link in a new window"
to leave a trail of papers or sites I'm looking at, and I can end up with
60-70 windows open in various stages of rendering. Memory is the key
bottleneck there, but the number of IE windows (and thus threads) that can
be rendering at a given time without a visible performance degredation is
significantly higher on a dual.

Another really classic example is "try burning a DVD while doing real work."
Normally a non-issue, if you do hit something that causes a spike of
CPU-intensive activity, you end up with either Burn-proof kicking in (fine
for data, unacceptable for video) or a coaster. A dual makes a noticeable
difference in that.

As a last example, although it's only one relavent for Linux users
typically: try running games on a system that's also running live services.
See how often load on the services makes the game stutter. Duals are MUCH
better.

Are any of these typical home use situations? No. Are any of these THAT out
of the ordinary for power users? Not that far gone, no.