TM1 Forum

Has anyone tried running TM1 on GPU, with Linux OS?

I don't believe IBM support GPU execution - so I can't think how you would make that happen. It's been used by other OLAP products eg
https://knowledgebase.jedox.com/knowled ... rator-faq/

You can't.
If you don't want the code executed, either remove the mappings etc which generated the code, or wrap the code block as follows:

TM1PY could be useful for this eg:
https://github.com/cubewise-code/TM1py- ... equests.py

Re: Exporting data from view to file consuming memory Post by vladino » Thu Apr 12, 2018 3:17 pm David Usherwood wrote: ↑ Thu Apr 12, 2018 3:06 pm If memory is really constrained you could clear the cache periodically by changing a value somewhere and changing it back again What do you mean by clea...

If memory is really constrained you could clear the cache periodically by changing a value somewhere and changing it back again. But I suspect your export view includes many more data points than you really need, especially consolidations. Depending (geddit?) on your dependences, the cellputn may be...

I just hit this moving a server between PA Local versions, specifically from 10.3.10200.60377 back to 10.3.10000.60286 . Only the new objects were kicked out.

I'm prepared to be corrected, but my belief is that if you wrap DBRWs or DBSWs in other functions, the performance benefit from using Stargate Views(via VIEW()) does not apply.

I agree that DBSW still has a part to play - but with your testing logic I don't believe you will get the Stargate VIEW() performance win. Given that you don't want any data to go unless your complicated tests are passed, I think you would be better off changing your cubename cell ($C$4) to be a for...

Turn on Performance Monitor and study }StatsbyCube and }StatsforServer - that should tell you what's eating memory (and thus your pagefile).

Marius Wirtz' TM1py has some sample code which appears to address what you want: http://tm1py.readthedocs.io/en/latest/api.html#tm1-objects Extracts from the docs: class TM1py.ChoreService(rest) Service to handle Object Updates for TM1 Chores activate(chore_name) create(chore) create_task(chore_name...

Finally!
Must have a look at how they've done it.

You can certainly use a sequential dimension to capture your data. But since your content doesn't seem to work well in TM1, can you explain what you are going to do with it?

I raised this with IBM in Feb 17 about PA Local 2.0 and received nothing of use. I suspect that the FIPSOperationMode parameter may be relevant but (again) the documentation is of no help.

Short answer is yes. In PM you can connect to and build applications from multiple servers. Possibly stating the obvious, the applications should have different names. I haven't used the application migration facility as it had a poor reputation (possibly been fixed) and I found that rebuilding an a...

That looks like a grid, not an active form. If you use an active form, you can put conditional formatting rules in the format area at the top and they will apply to the current output.

The current version is supported under Docker for Windows (Server 2016) but I have not heard positive reports about it so you are better off installing on Linux (IBM push Red Hat but Ubuntu is fine) I have now heard from the InfoCat team that the latest release of PAW for Windows is stable and reco...

PAW is a separate SKU. Also, it requires Docker to run. The current version is supported under Docker for Windows (Server 2016) but I have not heard positive reports about it so you are better off installing on Linux (IBM push Red Hat but Ubuntu is fine). This can be virtualised. It is a new set of ...

There isn't any. (Well, only the extension )
For strange historical reasons TM1 exports files as CMA. You can parse them into Excel or just rename them to .CSV .

I am planning to wrap rest api with JAVA and possibly publish it via github too.

What a good idea
Hubert Heijkers, who 'owns' the REST API, has some routines for Google's Go language, but when I last looked they weren't as comprehensive as TM1PY:
https://github.com/Hubert-Heijkers

Re: View Extract with TM1 API by macsir » Tue Oct 24, 2017 11:38 pm David Usherwood wrote: ↑ Tue Oct 24, 2017 3:13 pm Re: View Extract with TM1 API Yes, that's good idea but basically TM1py is still built on rest api. If you are experienced with programming and want to learn things, better to start...