Celeron and Solidworks?

[wgates666]

Hi.

We have P][-350's here at work, which are limited to a FSB of 66MHz. We are
not going to upgrade entire systems at this time, and I think upgrading the
motherboards is probably just going to be more of a pain than it's worth.
But I do want to upgrade the CPUs. The highest P][ is 450MHz, which is only
an increase of roughly 20%. The MB's are dual-processor capable, but again,
you generally only get roughly a 20% (at best) performance increase by
adding a second processor.

Here's what I'm thinking about: Ditch the P]['s, and toss in one or two 733
or 766MHz celerons (dual celerons is relatively easy with a little
modification), since I'm limited to a FSB of 66MHz. And, of course, get a
socket-to-slot1 converter.

Now, the celeron has 128k of cache vs 512k for the P][. Would this affect
performance very much, despite more than doubling the (current) CPU speed? I
would think the celeron766 would still be much better than a P][-450.

Does anybody have any experience with Solidworks and the Celeron? Any major
problems/slowdowns/etc?

We're running NT4.0, SW2000, and 256M RAM (i think). (I'd also like to get
some more memory when I upgrade.)

Thanks,

john

[wgates666]

Actually, never mind mostly. my max speed for the mb is 450Mhz. I was misled
(unintentionally) by some tech support folks. I also have a 100MHz FSB with
the 350 CPU.

oh well.

but still, is the celeron even worth considering for NT and solid works? at
any speed.

thx,
john

[kellne...@my-deja.com]

I would try overclocking before going to the Celeron. My PII 400 will do
550 without too much trouble. See how fast you can run up the system
clock. That will speed up the bus and graphics. Then see if you can jack
up the cpu clock multipliers.

This is a short term fix as overclocking may shorten the life of your
system. Also, make sure you dust your fans and heatsinks to like new.
And maybe put a bigger fan on the cpu.

Stay away from the Celeron. With the small cache you are just
exacerbating the problem you are having with the 66MHz FSB, by limiting
memory bandwidth. Cache is three or four times faster than memory. I
remember the day I got my first 300Mhz PIII with 100MHz FSB. What a
difference that made!

In article <93hrbu$gdg$1...@slb1.atl.mindspring.net>,

Sent via Deja.com
http://www.deja.com/

[Glen Ditchburn]

We just went through this dilema. What we did was upgrade m/b's and
processors to PIII 800 with 100 mhz bus to suit the pc100 ram we already had
and have been extremely happy with the outcome. Please understand we already
had an average of 384mb of pc100 ram for each machine that we upgraded
that's why we stayed with a 100mhz cpu. The machines are much more
responsive than before and we passed our motherboards and cpu's to the
number and word crunchers in the other departments. All in all we made a lot
of people happy for a small amount of money. (compared to new systems)
Glen Ditchburn
"wgates666" <jjabl...@deluxebostitch.com> wrote in message
news:93hrbu$gdg$1...@slb1.atl.mindspring.net...

[kellne...@my-deja.com]

The point I was making about the Celeron comes from an article on
memory bandwidth on the Anandtech site. Seems that on any given machine
memory bandwidth rises rapidly till the L2 cache can't handle the
data chunk being written out.After that it drops like a rock in a pond.
The bigger the L2 cache the bigger the data size can be before you fall
off the cliff. After that it is up to memory bandwidth as determined by
FSB frequency among other things. The point is the Celeron's small L2
cache means you will be depending on the memory bandwidth more.

In article <93i3ji$4rv$1...@nntp9.atl.mindspring.net>,