 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I am putting this up as a new topic because it will likely be interesting to
anyone. I asked earlier today about the best way to use a float to find a
vector.
example: Var = 3 ------> vector <Var*2, Var, Var/2>
JRG gave the solution which I was trying to avoid, using a separate function
for x, y, and z.
ABX scripted a complex function do deal with the problem
news:jltmaukpci64s7497l1qdr79asqpl6tpln@4ax.com
I ran each (plus a macro) through a loop 1 million times and recorded the
parse times:
*Macro*
Macro was the big loser at 1:31
*<Calculate_x(), Calculate_y(), Calculate_z()>*
*ABX pigment function*
These both parsed in 50 seconds.
Functions as expected are much faster than macros, but no difference in
parse times over 1 million iterations between three very simple functions
and one complex function.
-Shay
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Thu, 4 Apr 2002 07:42:00 -0800, "Shay" <shi### [at] houston rrcom> wrote:
> I am putting this up as a new topic because it will likely be interesting to
> anyone. I asked earlier today about the best way to use a float to find a
> vector.
I believed you perform such a test. I was relly interested in results.
> I ran each (plus a macro) through a loop 1 million times and recorded the
> parse times:
>
> *Macro*
> Macro was the big loser at 1:31
1 minute 30 sec or 1 hour 30 minutes ?
> *<Calculate_x(), Calculate_y(), Calculate_z()>*
> *ABX pigment function*
> These both parsed in 50 seconds.
>
> Functions as expected are much faster than macros, but no difference in
> parse times over 1 million iterations between three very simple functions
> and one complex function.
Have you considered conversion from 5D to 3D in your test ?
One complex function version has an adventage that it can be passed as one
parameter for manipulation macro. It can speed up macro call I think.
ABX
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
news:k22oaucfog1al5e38660hg1s3h23jub2qf@4ax.com...
>
> 1 minute 30 sec or 1 hour 30 minutes ?
>
1 minute 30 seconds
>
> Have you considered conversion from 5D to 3D in your test ?
>
Removing conversion showed no measureable difference with 200,000
iterations. I did not test this at 1 million
> One complex function version has an adventage that it can be passed as one
> parameter for manipulation macro. It can speed up macro call I think.
>
I think that one complex function would have a speed advantage if it were
rearranged to fit a specific use. The individual functions which I used to
find x,y,&z contained operations which were common to each function. One
complex macro could perform the repetitive operations once while 3 simple
functions has to perform the operations for each vector.
-Shay
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
