#Limiting Datastructures
8 messages in this thread
I did some tests to determine what memory gains could be made by adjusting the
shaper and lofter settings in 3DS.SET. I came to these results:
A) Standard Version 3DS.SET with default settings
B) Slimmed 3DS.SET with:
SPLINE-MAX = 1
SHAPE-MAX = 1
MOD-MAX = 1
PATH-MAX = 2
DEFORM-MAX = 2
C) Bloated 3DS.SET with:
SPLINE-MAX = 9999
SHAPE-MAX = 9999
MOD-MAX = 500
PATH-MAX = 9999
DEFORM-MAX = 9999
There was only about 50byte difference between the slimmed version and the
standard one, but there was a 4M difference between the bloated version and the
slim version. So unless you are using the maximum settings you aren't going to
get much of a memory reduction by changing these values.
Jonas: In the docs, it states the minimum for both PATH-MAX and DEFORM-MAX to
be 1, but 3DS won't run with this setting, instead they must be set to a
minimum of 2. Just needs a document correction.
David "Setting=infinity" Gould
Hi David,
<< doc bug.. 1 to 2 >>
thanks. I'll forward it to QA.
jonas[adesk]
David:
Thanks for taking the time to do this test, I've been meaning to for years and
never got around to it! :^)
Greg Pyros
Hiya Sparky!
Thanks for the test results, they were PERZAKLY the answer to my question.
Occasionally, I need to import or load a huge .DXF or .SHP file to use as a
template in the shaper, and since 3DS only tells me the SHAPE-MAX is too small,
but not by how much, I just top it out and import away. Not being particularly
judicious in my house keeping, I have been known to allow that setting (and the
PATH-MAX setting) to remain maxed out until I need to change something else in
3DS.SET later. I haven't ever noticed any problem from leaving them at 9999
(currently with 44 MB of RAM), but I was curious as to the size of the buffers
allocated for each parameter.
Now I know I won't need two .SET files for importing/loading and editing. 😀
BTW, did you ever upgrade to WinCIM 1.3? After visiting the WinCIM Support
forum for a few days, I'm a little hesitant to upgrade to another flavor just
yet. Lotsa whinin' about every iteration BUT 1.3 makes me think they did this
one right for the most part. I think its still available in the one of the
libs, so you might want to look (its FREE!).
BTW2, did you get your UL problem fixed?
Thanks again for the test report …B.
RE: WinCim 3 and "free."
At 2400 baud it takes many hours, then you get charged $10, I think.
FWIW.
Paul S.
Paul S.
<< At 2400 baud it takes many hours, then you get charged $10, I think…>>
Going by the sheer volume of D "verbose-r-us" G's posts, I think its
safe to assume he's not communicating at 2400 baud. When I downloaded
WinCIM 1.3, it was two disks, took maybe 5-10 minutes at 9600 (just a
guess??), and my CI$ account was credited $10 which put the cost at
$0 (excluding connect charges during downloading). I use it
exclusively for CI$, and have no complaints with it.
FWIW, I tried OzCIS (DOS version) for awhile, but found the pull-down
menus and pretty pitchurs on the little buttons in WinCIM made CI$
more fun and much easier to use.
Bruce,
>> Going by the sheer volume of D "verbose-r-us" G's posts, I think its safe to
assume he's not communicating at 2400 baud. <<
<g>. When I'm in Minneapolis I can hook up at 14.4, which is 6x faster
than 2400 in Fargo. WinCim 1.3.1 estimated dl time was 28 minutes @ 14.4, vs.
3 hours @ 2400. I forgot about the $10 credited back to your account. 🙂
>> FWIW, I tried OzCIS (DOS version) for awhile, but found the pull-down menus
and pretty pitchurs on the little buttons in WinCIM made CI$ more fun and much
easier to use. <<
I heard that OzCIS did have a GUI, I assume there is a Windows version?
Any idea (anyone) where I can get a hold of it?
Paul "Faster is Better" S.
Paul:
>> I heard that OzCIS did have a GUI, I assume there is a Windows version?
yup.
>> Any idea (anyone) where I can get a hold of it?
GO OZCIS.
Greg (short and to the point for a change!) Pyros