Reviewing the gds output logs, it appears that there may only be 2 times when the GDS_FILE property is used to copy the gds data instead of converting the mag data.
when LEFview is true
when the mag file is completely empty of layers and instances like small cells that have been expanded (M3_M2*).
I tried the following
Opening libs.ref/gf180mcu_fd_ip_sram/mag/gf180mcu_fd_ip_sram__sram512x8m8wm1.mag (has GDS_FILE property but no LEFview property), and expanding before gds write writes all mag data without accessing the GDS_FILE.
Copying that mag file to a local directory and opening, expanding and attempting gds write gives
Failure to read entire subtree of the cell.
I/O error in writing file sram_test.gds.
File may be incompletely written.
Comparison of console log files is exactly the same except one file opens the cells from
There are no missing cells messages.
This may be an unrelated problem.
Deleting all use and associated statements from the local file, adding LEFview true and the running gds write will produce a gds file with a copy of the specified gds as desired.
Reviewing the gds output logs, it appears that there may only be 2 times when the
GDS_FILEproperty is used to copy the gds data instead of converting the mag data.LEFviewistrueM3_M2*).I tried the following
Opening
libs.ref/gf180mcu_fd_ip_sram/mag/gf180mcu_fd_ip_sram__sram512x8m8wm1.mag(hasGDS_FILEproperty but noLEFviewproperty), and expanding before gds write writes all mag data without accessing theGDS_FILE.Copying that mag file to a local directory and opening, expanding and attempting gds write gives
Comparison of console log files is exactly the same except one file opens the cells from
and the other opens from
There are no missing cells messages. This may be an unrelated problem.
Deleting all use and associated statements from the local file, adding
LEFview trueand the runninggds writewill produce a gds file with a copy of the specified gds as desired.