somethings about panic file

Hi nidon

This means that your Virtuoso session has crashed. It might be due to your codes or it might be a bug. Please report the issue to your local Cadence support together with the CDS.log file that shows the crash stack.

Best regards
Quek

Hi quote:

Thanks for your reply.As what you said, I check the CDS.log file and the Virtuoso did crashed.

But I found other two problems:

1.The reason cased crashed is seems I generate too much layouts ont time and it will make Virtuoso crashed.And I reduce the number of layouts,the code executed well.

2.I found that,at the beginning of the code executed,the code running very fast,but after a few time,the speed is begin to slow.So,is there any way to keep the speed?

Hi nidon

It is good to know that you have a workaround for the crash. Perhaps you might want to examine your codes to make it more efficient (e.g. use dbClose to purge unused cellviews, etc) so that it does not take up more memory as it runs. You can use the "Profiler" in "ciw:Tools->SKILL Development" to check the amount of time/memory used up by various functions in your codes. It requires a SKILL development license.


Best regards
Quek

Hi,

Regarding #1 reason could be the memory - to much memory used and no posibility to alloc more.

I guess that if this is the case there you soubd see some message in CDS.log, before the message which referes to panic file.

Regarding #2, have a look on function  hiSetUndoLimit(n_undoLimit)

This will disable the Undo edit feature, but the procedures like creating/resizing/deleting shapes should run faster.

I found it very useful when doing thousands operations with shapes.

Best Regards,

Marcel

Thanks for Quek and  marcelpreda

I check my code,and I found I use dbOpenCellViewByType and don't use the dbClose to release the memory generated by dbOpenCellViewByType .

So,I think it's the key reason.

And for your suggestions to  improve the speed of script,I did a test for them.
I use the command groups dbOpenCellViewByType(fileA) ,dbOpenCellViewByType + dbClose(fileB),dbOpenCellViewByType +dbClose+hiSetUndoLimit(0)(fileC) in three files(each file total generate 60000 layouts) .And run them at one time.

And the result is running time of fileB > fileA  >= fileC.

I feel very insteresting here,theoretically speaking I use dbClose to release the memory,so the time of file A should be the longest time in the three.But the result is the longest file is fileB(I run these files in three computer).

Hi,

I think that is OK to have exec_time(fileB) > exec_time(fileA).

The idea is that dbClose() is also saving data on disk, and this opperation is time consuming .

Best Regards,

Marcel

Thanks for your advice,marcelperda.

It makes the file average executing time is reduced more than 30%.

Thanks for your kindly support,marcelperda and Quek.