Cannot find a legal location....

 Hi Trigger,

Check the LEF for the BUX20 cell. What SITE is defined for it? Is that SITE defined in your tech LEF? And does your floorplan have rows made from that type of SITE? If so, is there open room to place cells (meaning, the rows are not covered by blockages, macros, etc.)?

- Kari

 Hi Kari,

 thanks for the reply...

The SITE defined is 'core'...it looks like:

MACRO BUX20
    CLASS CORE ;
    FOREIGN BUX20 0 0 ;
    ORIGIN 0.000 0.000 ;
    SIZE 18.200 BY 13.000 ;
    SYMMETRY X Y ;
    SITE core ;

     ..... 

 ...but at the technology LEF file nothing is mentioned about SITE core...Does it has to be..? I get a warning about this:

'Physical library does not contain description of any core site' 

At the first look it seems that at the floorplan there isn't rows from that kind - SITE...about the open room place....You mean is there overlaping between rows and blockages for example..? If so...no..there is no overlaping

During synthesis when LEF file is read I get another one info message :

Not all cells of logical library exist in physical library.
<PHY-291>.
 All cells of physical library exist in logical library. <PHY-294>

I do not if this could be part of the problem..cause I get this message on the two machines...

I'll be glad on any help from Your side...

Thanks.... 

 Yes, you ABSOLUTELY have to have your SITE definitions in the tech lef. Sometimes they are in the macro lef along with the cells - it can be either place, but they do have to be defined somewhere. That's why you have no rows - Encounter doesn't have a SITE to build the rows with.

The other warning means that you are reading in cells either in the design or in the .lib files that you are not reading in LEF for. The log file should list what cells these are. You will need to find the LEF for these cells. It's a separate problem from the row/SITE issue, but one that still needs to be fixed.

Hope that helps,

- Kari

 Hi,

yes...as You are saying  it is defined at the library .lef file and not at the technology .lef file...about the rows...I have them..cause after executing the script I open the encounter graphically and I can see them...I have checked the cells height...I thought that it may be dfferent than that defined at the core..but it is the same....it is something with the cells that I'm missing cause I do not get it for every cell....is it possible that the space is not enought...I mean physically? 

Abou the second one...I'm still investigating....

Thanks a lot...

Reagrds,Martin  

Strictly speaking, the SITE definintions do not have to be in the technology LEF, the first LEF file in the list of lefFiles.  But they must be defined before being used in any of the LEF files.  The site definition (from memory) would look like:

SITE core

CLASS CORE ;  # can be also be PAD or COVER, blocks nolonger need sites

SIZE <x> <y> ;  # placer expects that x dimension will be smaller then y for horizonal row sites

SYMMETRY Y ;   # legal values are X Y and R90, but only use Y for horizontal row libraries

END core

Note the SYMMETRY for sites has a different meaning then it has for cells.  Normally only use SYMMETRY Y, which would allow flipping (mirroring) in the Y dimension in rows.  If you specify X it would allow the cell to flip upside down in a North or R0 row , which can cause power shorts. 

For cells it is a list of allowed orientations, so if you want cells to be placed in upside down rows, be sure to allow X  for cells.  If you want to allow mirroring of cells, then allow Y for cell symmetry.

I know this is confusing, the developer of LEF want to save a token, so reused the SYMMETRY statement.  Not the best decision when viewed as a user.

Regards,

- Tom Belpasso

Integrated-Vision.com