LCA Commons ecoinvent 3.5 linked database giving errors when characterising using python IPC.

Question

I have been characterising the LCA commons ecoinvent 3.5  linked database using olca-ipc. The issue is when applied on random datasets selected from this database, it works fine. However, when executed on the whole database it gives error. I suspect the issue is with calculation for some of the datasets. The error log files can be found here ERROR_LOGFILE

If you could offer priority support, can you provide price info to my email. 

Answer 1

Hi, indeed, this is a known issue, we will release the new LCA Commons linked to newer ecoinvent, where we spent quite some effort to check and correct data from the different repositories, also taking the latest release from LCA Commons - you can write to the contact email of openLCA but maybe it is good to wait for the newer version. Thank you!

Comments

Hi Andreas, thanks for the response.
-   I solved it by increasing the ram size from default 3000+ to 15000 MB.
- Also, when I execute this in parallel, and when set to maximum cores, this issue occurs. So, I set to 80% of the cores and the execution works flawlessly without any resource constraints.
-  No issues with datasets it seem. I guess for some datasets , the calculation requires more resources and when it is not available, these issues occur.
-  Regarding parallel, It is not  working as expected . In many cases serialized OLCA calculations are faster than parallel executions . I do not how know  it works in the background

---

thank you for the comment - are you using the Blas libraries? You can download these from the welcome page in openLCA. They increase the performance a lot.  And if you are using them, these are mentioned in the logfile.