i know it sounds blasé, but shouldn't Next Limit be *fixing* the code
to take advantage of multiple cores?
Adrian Wyer
Fluid Pictures
33 Glasshouse Street
London
W1B 5DG
T: +44 (0) 207 434 5133
www.fluid-pictures.com <http://www.fluid-pictures.com>
----- Original Message -----
*From:* Kim Aldis <mailto:XSI(at)kim-aldis.co.uk>
*To:* XSI(at)Softimage.COM <mailto:XSI(at)Softimage.COM>
*Sent:* Monday, February 12, 2007 1:34 PM
*Subject:* RE: realflow test
Check using the performance monitor, see if it’s only ever using
one cpu or if it’s swapping threads between the two. You can run
two traces, one for each core. If they’re inverted versions of
each other then it’s using both cores, just not at the same time.
I’m not sure I remember exactly what that means but it may just be
that RF isn’t being too efficient.
*From:* owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM]
*On Behalf Of *Duncan Burch
*Sent:* 12 February 2007 13:29
*To:* xsi(at)Softimage.COM
*Subject:* RE: realflow test
Hey
Initially was a 2cpu dual core thingy.
Never used more than 40% or my machine.
Now I use 1 cpu dual core, never gets above 50%.
Ive increased the number of threads, doesnt help, but is this the
same as cores?
ta
Dunk
------------------------------------------------------------------------
From: adrian.wyer(at)fluid-pictures.com
To: XSI(at)Softimage.COM
Subject: Re: realflow test
Date: Mon, 12 Feb 2007 12:27:58 +0000
what machine you running on?
im assuming you upped the number of threads in prefs?
Adrian Wyer
Fluid Pictures
33 Glasshouse Street
London
W1B 5DG
T: +44 (0) 207 434 5133
www.fluid-pictures.com <http://www.fluid-pictures.com>
----- Original Message -----
*From:* Duncan Burch <mailto:duncan_burch(at)hotmail.com>
*To:* xsi(at)Softimage.COM <mailto:xsi(at)Softimage.COM>
*Sent:* Monday, February 12, 2007 11:49 AM
*Subject:* RE: realflow test
I have to completly agree here.
I love using RF, at work and in my spare time.
The tests i do in my sparetime look kinda good, but arent very
complex, from a scene point of view.
But when I use RF at work, it can be a pain to tweak things
100 times for the client.
Drives me mad.
On a side note, anyone know why RF never uses all my cores?
TA
Dunk
------------------------------------------------------------------------
From: forum(at)virtualrepublic.org
To: XSI(at)Softimage.COM
Subject: RE: realflow test
Date: Mon, 12 Feb 2007 12:22:49 +0100
Kim hit the spot of pain and frustration.
Most Realslow examples are fluid sims in a simple and limited
environment.
If you want to spread it to bigger scenes you will figure out
that RF is a great and faszinating tool but totally useless
for a production ruled by short deadlines, many changes and
limited budget.
Another killer is that you can't simulate with a low-res setup
and just use high-res settings to get a final »perfect«
result. Changing the settings will cause a total different
output of course and all you showed to the client in a fast
pre-viz version is history. You are forced to figure out the
best looking simulation first and then you end up in an
incredible simulation process which could just be done by a
single number cruncher. Means: you never will have enough RAM
or procs. At the end of a week a 128GB RAM with 8 procs is on
your wishlist and even this is not enough I believe.
I am soooo negative ... I know ... but I have been spanked by
RF so often and in Germany we have a famous sentence for that:
Die Hoffnung stirbt zuletzt! [the hope dies at last].
------------------------------------------------------------------------
*From:* owner-xsi(at)Softimage.COM
[mailto:owner-xsi(at)Softimage.COM] *On Behalf Of *Kim Aldis
*Sent:* Monday, February 12, 2007 11:22 AM
*To:* XSI(at)Softimage.COM
*Subject:* RE: realflow test
The real killer is when you want to flood a street. The level
of detail you need requires huge numbers of particles. It’s
not actually working out how to get the effect that’s the
problem, it’s how to get enough particles into the shot to
achieve the effect.
*From:* owner-xsi(at)Softimage.COM
[mailto:owner-xsi(at)Softimage.COM] *On Behalf Of *VR XSI Forces
*Sent:* 12 February 2007 10:04
*To:* XSI(at)Softimage.COM
*Subject:* RE: realflow test
Yeah ... I also recommend the DigitalTutor DVD's ... scene
scale is everything ... right ... but at the end it wouldn't
help us here.
We all start with an example like this where a tube of 10
litre water fills a little glass or a box. Fine and fascinating.
Then you get a storyboard where the client wants to have hot
chocolate floating over a waffle in highspeed.
The client now expects: less costs, fast and flexible
production, control over fluid behaviour, shapes and drops and
... a perfect looking result like the EUR 40.000,00 highspeed
shooting.
------------------------------------------------------------------------
*From:* owner-xsi(at)Softimage.COM
[mailto:owner-xsi(at)Softimage.COM] *On Behalf Of *Andi Farhall
*Sent:* Monday, February 12, 2007 8:35 AM
*To:* XSI(at)Softimage.COM
*Subject:* RE: realflow test
what scale is this?
-----Original Message-----
*From:* owner-xsi(at)Softimage.COM
[mailto:owner-xsi(at)Softimage.COM]*On Behalf Of *adrian
*Sent:* Sunday, February 11, 2007 3:43 PM
*To:* XSI(at)Softimage.COM
*Subject:* OT: realflow test
a few hours of noodling......
http://www.fluid-pictures.com/PUBLIC/FLUID.MOV
i recommend the Gnomon realflow training dvd, even though
it's a bit tedious in parts
there are some insights into settings that really help,
and of course, scene scale is EVERYTHING!
A
Adrian Wyer
Fluid Pictures
33 Glasshouse Street
London
W1B 5DG
T: +44 (0) 207 434 5133
www.fluid-pictures.com <http://www.fluid-pictures.com>
______________________________________________________________________
This email has been scanned by the MessageLabs Email
Security System.
For more information please visit
http://www.messagelabs.com/email
______________________________________________________________________
------------------------------------------------------------------------
Get connected - Use your Hotmail address to sign into Windows
Live Messenger now. Connect now!
<http://get.live.com/messenger/overview>
------------------------------------------------------------------------
Get connected - Use your Hotmail address to sign into Windows Live
Messenger now. Connect now! <http://get.live.com/messenger/overview>
