[KPWalker]

Hi,

Just wanted to say thanks for the new version of inst::Flatten() – works a treat!

Cheers,

Keith W.

[KPWalker]

Hi, that sounds great thanks.

I wouldn't add an extra flag to these functions to control flattening instances hierarchically, or by some level. It's easy enough to do by cycling through by as many levels as required (or until they're aren't any instances or arrays left) so I'd leave the functions nice and simple on the toolbox vs. Swiss army knife principle.

Look forward to using the new functions when they're available!

Cheers,

Keith W.

[KPWalker]

Hi Keith,

Ah, OK, I think I understand. I'd call that something like MergeAll.

To my mind 'Flatten' refers to the hierarchy of cells and 'Merge' to layer primitives, i.e. Flatten leaves primitives alone and flattens instances (one layer of cell hierarchy only) whereas Merge operates on Rectangles, Polygons, Paths, etc.

It does leave the question of what, if anything, either type of function should do with PCells. My take would be Flatten would operate on them not Merge but…

… of course these are just my opinions – others may beg to differ!

Cheers,

Keith W.

[KPWalker]

Hmmmm, this is even more complicated than I thought – the version of Python I'm now running is 2.7, i.e. I've just downgraded by 0.0.3 ! Still, the latest and greatest Glade is running again so I'm happy enough.

Thanks again for your support.

Cheers,

Keith W.

[KPWalker]

Hi,

I couldn't find a python27.dll in the system32 directory of my machine but there was one in SysWOW64 – 2.19MB 11-Apr-12

So I re-installed Python from the link you give and that's cured the problem – Glade 4.3.27 for Win7 64bit is now running happily – thanks for the great support.

Interesting quirk – the python27.dll now the system32 directory is reported to be 2.83MB 04-Jul-10

Cheers,

Keith W

[KPWalker]

Hi, I think there may be an issue flattening an instance using Flatten() in a script. It works fine from the GUI but when I try to do this in a script I get a second copy of instance (with a new instance name) on top of the (unchanged) original.

I think the following fragment will demonstrate the issue (and/or my incompetence)

inst_list = cv.getInsts()
for inst in inst_list :
name = inst.instName()
print "nFlattening ", name
inst.Flatten(cv, dbTransform())

cv.update()

inst_list = cv.getInsts()
for inst in inst_list :
name = inst.instName()
print name

This is using Glade 4.3.25 Win7 64bit and Python 2.7.3

Any ideas?

Cheers,

Keith W.

[KPWalker]

Hi, I have a problem with the latest release (4.3.27) – when I try to run it all I get is a brief flicker of the screen then it's as if the program hadn't been run. In my case it's the 64bit Win7 version but a friend has the same results with the 32bit version under both Vista and Win7. My back-up of 4.3.25 still works just fine (thanks for the warning!)

If the information is of any help I'm running Python version 2.7.3 and if there are any other things you need to know that would help pleasesay and I'll do my best to answer.

Cheers,

Keith W.

[KPWalker]

Yup, a hex dump shows 0x0D 0x0A – apologies for describing them as blank in the first place.

I'd also note that there is the same thing in the Cadence technology file (.tf) itself. For instance in the one I have there are 'blank' lines in the definitions of the routing directions and min spacings amongst others. It looks like they come from where the file is created for specific process options, e.g the number of metal layers, so it's probably quite common. It doesn't crash Glade but I don't know if it's an issue or not as I haven't investigated these features yet.

Sorry I can't supply the files – they come from a foundry's PDK.

Rgds,

Keith W.

[KPWalker]

Hi,

Looks like there's still an issue in Glade 4.3.24 (64 bit under Win7) with reading Cadence display files. Glade no longer crashes after the latest fix but it doesn't read in/recognize the stipple definitions (except the first). Per previously removing the blank lines between the stipple definitions cures the problem.

Rgds,

Keith W.

[KPWalker]

Hi Keith,

I've just found the issue in my case – there are blank lines between each of the stipple definitions in my Cadence display.drf and according to the message window the Glade parser only sees one stipple definition (presumably the first). Now I've manually deleted these blank lines from the .drf Glade 4.1.7 reads it (and the .tf) just fine.

I guess Python isn't quite the same Skill!

Cheers,

Keith W.

[KPWalker]

Hi, I saw a similar issue with a Cadence PDK I received recently. The crash seems to be caused by the techDisplays section of the .tf rather than the .drf itself (which read in OK).

techDisplays(
;( LayerName Purpose Packet Vis Sel Con2ChgLy DrgEnbl Valid )
;(

---

---

---

— —

---

---

---

)
( L1 P1 L1_P1 t t nil t t )
( L1 P2 L1_P2 t t nil t t )
( L2 P1 L2_P1 t t nil t t )
( L2 P3 L2_P3 t t nil t t )
( …

) ;techDisplays

When I edited the .tf and removed the 'Packet' column it no longer crashed Glade 4.1.7 although it didn't seem to work as a tech file.

It may be a different issue or just a symptom but…

Rgds,

Keith W.

[Author]

Posts