https://open-source-silicon.dev logo
Channels
Powered by
#openlane
  • a

    aryap

    09/04/2020, 9:22 PM
    i feel like i'm running into common problems but the answers are scattered through this slack
  • a

    aryap

    09/04/2020, 9:22 PM
    it's also usually unclear whether i've designed something wrong, i've configured something wrong, or there's a bug
  • a

    aryap

    09/04/2020, 9:23 PM
    would there be appetite for such a thing? probably as a wiki?
    a
    m
    • 3
    • 9
  • a

    aryap

    09/05/2020, 7:11 PM
    has anyone else run into this sort of error opening the magic-output GDS in klayout?
    s
    • 2
    • 3
  • t

    Tim 'mithro' Ansell

    09/05/2020, 7:29 PM
    @User @User might know
  • t

    Tim Edwards

    09/05/2020, 7:33 PM
    @User: No idea. I would need more information on what data were surrounding byte position 525606954.
  • a

    aryap

    09/05/2020, 7:37 PM
    I mean, i can give you the corresponding chunk of the file, but would that even help? I'm trying to open the relatively vanilla output of an OpenLane run. Is there any wisdom to be received on how to start debugging this?
  • t

    Tim Edwards

    09/05/2020, 7:47 PM
    @User: I use a tool I have called stream-utils which dumps ASCII from a GDS file, but there are other tools that can do that, too. This seems to be more than just the fact that you don't have a setup for klayout; there is a difference between unknown layers and missing records. I am not aware of any recent instance where magic failed to generate at least a syntactically-correct GDS output file.
  • a

    aryap

    09/05/2020, 7:48 PM
    thanks @User. i wanted to try to open it in magic but that seems hard, so now i'm going to try and open it in virtuoso. that said, i'm only 85% sure this run completed without error
  • a

    aryap

    09/05/2020, 7:49 PM
    also indeed after setting up the sky130A tech klayout behaves no differently (but can read the floorplanned LEF fine)
  • t

    Tim 'mithro' Ansell

    09/05/2020, 7:54 PM
    @User Why would be opening in magic be hard?
  • a

    aryap

    09/05/2020, 7:55 PM
    because i have to learn how first 🤷 but virtuoso is handy 😉
  • a

    aryap

    09/05/2020, 8:02 PM
    ok the gds is probably busted for sure:
    Copy code
    ERROR (XSTRM-80048): Parser: A 'BOUNDARY' record in GDSII file '/users/aryap/riscv_top.gds' terminated without an ENDEL record near offset 525743052 The GDSII file could be corrupted.
  • a

    aryap

    09/05/2020, 8:02 PM
    argh this took 16 hours
  • a

    aryap

    09/05/2020, 8:05 PM
    ok the lesson here is to make sure the run actually completed, even if you don't see errors in the obvious places
  • a

    aryap

    09/05/2020, 8:05 PM
    Copy code
    25399357 Diagnostic:  Scale value is 0.005000
25399358 [INFO]: LEF Write Complete
25399359 [INFO]: Saving .mag view With BBox Values: 0 0 3139592 3824932
25399360 Warning: I/O error in writing file
25399361 Could not write file.  Cell not written.
25399362 [INFO]: MAGIC TAPEOUT STEP DONE
  • a

    aryap

    09/05/2020, 8:06 PM
    this was in the magic log
  • a

    aryap

    09/05/2020, 8:06 PM
    not errors.txt, because guess what, i ran out of disk space 🤦
  • t

    Tim Edwards

    09/05/2020, 8:25 PM
    I wondered that, given that it was at >500MB.
    ✔️ 2
    a
    • 2
    • 1
  • g

    Gaurav Kumar K

    09/06/2020, 2:45 PM
    Hi @User, I am facing issue while importing gds generated by open-lane into cadence virtuoso. The issue is that the pins are not getting imported into layout. I can see only labels in the layout. Please let me know how to resolve the issue. Regards, Gaurav
  • t

    Tim Edwards

    09/06/2020, 4:09 PM
    Are you using the SkyWater tech file for Virtuoso? Which purpose number does it use for labels, and which one for text? Unfortunately, SkyWater has made this into a big problem by confusing labels and pins in all of its cell libraries. When reading GDS, I have to treat both purpose 5 and 16 as pins. I probably need to reverse the numbers as I have them in the magic tech file when writing GDS output.
  • g

    Gaurav Kumar K

    09/06/2020, 5:06 PM
    Yes I am using skywater tech file.. I don't know how to find the purpose number. Please let me know where to find them I see that it is using met2 and met3 for the purpose of pins, i.e., met2 and met3 labels instead of pins
    t
    • 2
    • 2
  • g

    GitHub (Legacy)

    09/06/2020, 9:26 PM
    #34 How to determine die area from successful flow? Issue opened by SweeperAA I have successfully executed the top-level flow.tcl in the docker on one of the example designs (chacha in this case). Are there any results that provide the final die area required for the design (and what units the measurements are in)? Thank you for the help! efabless/openlane
    ✔️ 1
    👏 1
  • g

    GitHub (Legacy)

    09/07/2020, 2:02 PM
    #34 How to determine die area from successful flow? Issue closed by agorararmard efabless/openlane
  • a

    aryap

    09/07/2020, 8:28 PM
    before i go down this rabbit hole, can anyone confirm whether we can run OpenSTA's power and timing analysis on skywater-130 designs?
  • a

    aryap

    09/07/2020, 8:29 PM
    from previous chatter about this it looks to be difficult, and i can't find a blessed method in the docs
  • a

    aryap

    09/07/2020, 8:29 PM
    but it also seems like we do have the parts we need? (liberty files for sky130?)
  • t

    tnt

    09/07/2020, 8:37 PM
    openlane runs a pre-pnr STA analysis by default.
  • a

    aryap

    09/07/2020, 8:57 PM
    ah i see this is done in scripts/sta.tcl
  • a

    aryap

    09/07/2020, 8:57 PM
    thanks!
1...252627...131Latest