Help setting up a Kick Ass Machine with 1 Terrabyte of Storage P4 or G5

  • Thread starter Thread starter L.S
  • Start date Start date
Marc de Vries said:
Well, that is just your word against mine.

I know from experience that it does not do that. Whatever you claim
otherwise in a ng doesn't change my real-world experience with the
software.


Again your word against mine.
SOME programs does this, MOST do not.


Your guess is as good as mine.

But since it is quesswork I have never claimed that that MUST be his
botteleneck. I have also never claimed that I know what would be the
best solution for his problem, because it is impossible to determine
that without knowing his bottlenecks.

Instead I have given several likely scenarios and told him to determine
which scenario fits in his case, and what that means for the hardware he
needs to buy.

What is so terrible about that?


And how would you know that?

You are also guessing. But you claim that better disk I/O will solve
every problem.

LOL! Bravo, excellent self knowledge.
Come on Rita, be realistic.
ROTFLOL!


Knowing nothing about the current bottlenecks and then claiming that
scsi will solve all his problems, THAT is giving intentionally wrong
information.

I have given him several possible scenarios, so that he can determine
which scenario applies to him, and make the best hardware choice
without having to come back to this ng for every single step in making
that choice.
It suprises me that you don't see that.

ROTFLOL!
What IS surprising is that you still don't know what Rita is.
I have found from experience that in most such cases disk I/O is not
the main culprit.

Photoshop relies on memory, memory and more memory. (and cpu)


It depends on the situation. He is not just doing videoediting, but also
photoediting. Those applications have completely different demands.

Most videoediting is done streaming, which means you don't need much
memory at all. But in the vast majority of cases that means that the
CPU is the bottleneck.

Only when you do some really simple editing which doesn't require much
cpu will the disk I/O become the bottleneck.

That might be exactly what the OP is doing, in which case he needs to
buy a SCSI array. But neither you nor I know that.
He might just have a CPU bottleneck in which case it would be wasted
money to buy SCSI.

I really don't understand why you don't want to acknowledge that.
Please explain it to me.

Please hold your breath!
 
Well, that is just your word against mine.


Isn't this always the case?

I know from experience that it does not do that. Whatever you claim
otherwise in a ng doesn't change my real-world experience with the
software.



I never said that it did. Maybe your experience is different from mine?

Your guess is as good as mine.



Now we're cooking with gas. You're finally catching on.

But since it is quesswork I have never claimed that that MUST be his
botteleneck. I have also never claimed that I know what would be the
best solution for his problem, because it is impossible to determine
that without knowing his bottlenecks.



The whole point is that it should never be guesswork. But, we both know
that the disk I/O is the bottleneck.

Instead I have given several likely scenarios and told him to
determine which scenario fits in his case, and what that means for the
hardware he needs to buy.



Which is great, while generic in nature, it does give him some points to
ponder.

What is so terrible about that?



While not terrible, it's not productive either.

And how would you know that?



A side-by-side comparison with a single drive vs. an array, both SCSI of
course.

You are also guessing. But you claim that better disk I/O will solve
every problem.



I never claimed, "better disk I/O will solve every problem". But, I did say
that he would definitely have better overall performance, which we both know
is indisputable.

Come on Rita, be realistic.



I am.

Knowing nothing about the current bottlenecks and then claiming that
scsi will solve all his problems, THAT is giving intentionally wrong
information.



And claiming that he has bottlenecks that most likely don't exist and
discounting SCSI as an option to boost his performance is intentionally
giving out wrong information

I have given him several possible scenarios, so that he can determine
which scenario applies to him, and make the best hardware choice
without having to come back to this ng for every single step in making
that choice.



Which may or may not apply to him at all.

It suprises me that you don't see that.



I do see a lot.

That might be exactly what the OP is doing, in which case he needs to
buy a SCSI array. But neither you nor I know that.
He might just have a CPU bottleneck in which case it would be wasted
money to buy SCSI.



The same can be said about the Mem/CPU. I guess both of us will really
never know.

I really don't understand why you don't want to acknowledge that.
Please explain it to me.

I did explain it, many times, in fact.





Rita
 
On Tue, 27 Jan 2004 18:03:02 -0500, "Rita Ä Berkowitz"

The whole point is that it should never be guesswork. But, we both know
that the disk I/O is the bottleneck.

I'm sorry, but I am really not convinced that disk I/O is the
bottleneck.

I sincerely think that his most likely bottleneck for photoshop is RAM
and CPU, and for videoediting it is CPU.


But unfortunately we have still not gotten more information on that
from the OP.

Which is great, while generic in nature, it does give him some points to
ponder.


While not terrible, it's not productive either.

It's not? It gives him the oppurtunity to figure out himself, what
hardware he needs to buy.
How can that not be productive?
I never claimed, "better disk I/O will solve every problem". But, I did say
that he would definitely have better overall performance, which we both know
is indisputable.

It is not indisputable at all.

If his current bottelneck is CPU, then better disk I/O will give him a
performance gain of 0.0%
I am.


And claiming that he has bottlenecks that most likely don't exist and
discounting SCSI as an option to boost his performance is intentionally
giving out wrong information

I think you had better read the thread again, because I think you have
confused some reactions from other people, with my reactions.

I have never discounted SCSI as an option. I have stated several times
that SCSI is the way to go if he has a disk I/O bottleneck.

But since we still don't know what bottlenecks he has, it is not a
good idea to advice scsi in every situation and discount RAM and CPU
as an option.
Which may or may not apply to him at all.

Correct. It's up to him to decide what applies.

The same can be said about the Mem/CPU.

Absolutely true.
That's why I advice that the OP first determines the bottlenecks for
his applications.
If he has a disk I/O bottleneck, then buying RAM or a faster CPU is
also wasted money.
I guess both of us will really
never know.

I guess you could be right. Haven't seen the OP reacting on what
bottlenecks he sees on his current machines.
 
Don't forget about the weekend sale at OfficeMax for the 120GB WD drives
at about $40-50 each after rebate!

You can get a bunch to RAID, dirt-cheap!
 
Back
Top