PC compare question

B

Brian

I know this is probably a completely ignorant question but....I'm going
to ask it anyway.


I have three pcs that I'm running a large model on:

PC1 2.4 GHz processor - 2 G RAM - not hyperthreaded P-4 chip
PC2 1.2 GHz processor - 1 G RAM - not hyperthreaded Pentium M chip
PC3 2.8 GHz processor - 1 G RAM - hyperthreaded - P-4 chip


My run times are

PC1 - 22 Min
PC2 - 12 Min
PC3 - 12 Min

Q: Why is PC1 twice the run time? and why isn't PC3 faster than PC2?

What else do I need to consider?

Thanks for your help.
 
Y

Yousuf Khan

Brian said:
I know this is probably a completely ignorant question but....I'm going
to ask it anyway..

Well, you haven't given us all of the information necessary, so we're
left to assume and guess.
I have three pcs that I'm running a large model on:

PC1 2.4 GHz processor - 2 G RAM - not hyperthreaded P-4 chip
PC2 1.2 GHz processor - 1 G RAM - not hyperthreaded Pentium M chip
PC3 2.8 GHz processor - 1 G RAM - hyperthreaded - P-4 chip

The Hyperthreading should be largely irrelevant. Without knowing what
software you're running, we'll assume that it is not multithreaded.

Now the 2.4Ghz P4 system might be an older system running older memory
types, such as SDR rather than DDR memory. So that would slow it down.
It's also possible that that if it's not running dual-channel memory
regardless of whether it is DDR or SDR.

The Pentium-M system is an updated processor architecture that is based
off of the older Pentium 3 architecture, which was a much more efficient
architecture than the later Pentium 4 one. The new Core 2 Duo is an even
newer iteration of the Pentium-M and thus Pentium 3. Most people know
that the Core architecture just slaps the Pentium 4 architecture up and
down.

Yousuf Khan
 
T

Tony Hill

I know this is probably a completely ignorant question but....I'm going
to ask it anyway.


I have three pcs that I'm running a large model on:

PC1 2.4 GHz processor - 2 G RAM - not hyperthreaded P-4 chip
PC2 1.2 GHz processor - 1 G RAM - not hyperthreaded Pentium M chip
PC3 2.8 GHz processor - 1 G RAM - hyperthreaded - P-4 chip


My run times are

PC1 - 22 Min
PC2 - 12 Min
PC3 - 12 Min

Q: Why is PC1 twice the run time? and why isn't PC3 faster than PC2?

What else do I need to consider?

What else to consider?

1. Processor architecutre: This is *KEY* and it is entirely the
reason why PC2 is so much faster, clock for clock, than PC3. The
Pentium M acheives MUCH more on a per-clock basis than a Pentium 4,
while the P4 can clock to much higher speeds on a given process.

2. Processor cache memory: All modern computer chips contain some
fast on-board memory to use as a cache for the comparatively slow
system memory. More and faster cache is better. Given that
processors can execute 200+ instructions while waiting for one bit of
data from main memory, having data in your fast cache memory is very
important.

3. System bus speed: Modern Intel chips use bus speeds ranging from
400MT/s up to 1333MT/s. Faster bus speed helps transfer data between
the processor and memory more quickly.

4. Memory speed: As should be obvious from this point, getting data
to and from memory is an important factor in the speed of a computer.
The systems you mentioned above could have anything from
single-channel 133MHz SDRAM (1.06GB/s) up to dual-channel 800MT/s DDR2
SDRAM (12.8GB/s).

5. Motherboard and motherboard chipset: This is closely related to #3
and #4 above, however even with the same on-paper specs you can get
different performance depending on the quality of the motherboard and
it's chipset. There are dozens of chipsets that could be used on any
one of the systems above, and for each of those dozens of chipsets
there are dozens of individual motherboard models. Performance
between them might not vary too much, but certainly a 10% difference
between two motherboards with the same on-paper specs would not be
unheard of.

6. Every other component that might be involved: Things like hard
drive, video card, sound card, NIC or damn near any other piece of
hardware in the system could potentially have some effect here.

And of course, last but not least:

7. Software: Small differences in software can sometimes result in
HUGE differences in performance. On some complex models it wouldn't
be completely unreasonable to see an order of magnitude difference
from software optimizations alone.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top