Page 1 of 1
Too long visualization time
Posted: Tue Jul 17, 2012 7:58 am
by ioscat
When I ask my TM1 to build 3D view it takes two steps. First - it shows picture in "fogged' mode; and Second - it takes a long time before showing it again (in normal, contrast mode).
Is it bug or our developing mistake somewhere?
Re: Too long visualization time
Posted: Tue Jul 17, 2012 8:03 am
by Martin Ryan
Sounds like you might have a rule that hasn't been fed properly, or at all. Can you post the rule for this cube?
Also a screenshot of the cube structure, showing which dimensions make up the cube, would be handy.
Martin
Re: Too long visualization time
Posted: Tue Jul 17, 2012 9:16 am
by ioscat
1) there no feeders at all
2) I'm not sure it is possible to paste here client data.
Here is combination of screens I see when making 3D chart:
- 3d_tm1.png (171.71 KiB) Viewed 5696 times
While making this chart my browser asks me for several times to stop script. I ckick decline to get result picture, but it compiles (as it seems to me) too long.
Re: Too long visualization time
Posted: Tue Jul 17, 2012 10:16 am
by lotsaram
I suggest you (or more likely whoever is specifying report and chart types on your project) to go read up on Tufte, Hichert and Few (a new version of Show Me the Numbers is out ...) and make an effort to present data in a more clear and meaningful way. BUT the god awful choice of junk chart isn't likely to be your problem but rather a poorly written calculation. If you are serious about getting help then you need to post the rule and not the chart.
Re: Too long visualization time
Posted: Wed Jul 18, 2012 12:35 am
by Gregor Koch
Really don't mean to crush the chart and I agree with lotsa that it most likely is not the issue. But I can't help myself, this looks like a flight game from the 80s or a mixture between Super Saxxon and Root Beer Tapper.
Re: Too long visualization time
Posted: Wed Jul 18, 2012 1:06 am
by Martin Ryan
ioscat wrote:
2) I'm not sure it is possible to paste here client data.
If you can't paste the data then we're unlikely to be able to assist.
Re: Too long visualization time
Posted: Wed Jul 18, 2012 7:04 am
by Olivier
this looks like a flight game from the 80s or a mixture between Super Saxxon and Root Beer Tapper
These games were great
Re: Too long visualization time
Posted: Wed Jul 18, 2012 7:30 am
by David Usherwood
@ioscat, we don't need the data, just the rules.
But I'll stick my neck out and surmise that you aren't using skipcheck (and feeders) so you aren't taking advantage of TM1's sparsity handling. This will not work.
Re: Too long visualization time
Posted: Wed Jul 18, 2012 10:33 am
by ioscat
Hello to all. Thanks for your comments, spent a lot of time to translate all your "But I'll stick my neck out and surmise" and "But I can't help myself" and other. =)
Just asked my Boss, he didn't approved publication. But I can answer for question
David Usherwood wrote:But I'll stick my neck out and surmise that you aren't using skipcheck (and feeders) so you aren't taking advantage of TM1's sparsity handling
SkipCheck is used.
In addition I can say that calculations and recalculations goes fast enough (17 MB memory used). Maybe it is not properly fed, but is it possible to estimate feed rightness? An example by comparing two cell statictics numbers.
Maybe if I repeat this problem on sample model I'll post here rules. So on today topic should be temporaly stopped to not be trashed.