Monday, 2019-06-10

*** tpb has joined #symbiflow00:00
duck2latest symbiflow-arch-defs and conda vpr prints ~140K warnings to the output while building buttons_basys3_bin. am i doing something wrong?00:10
*** space_zealot has joined #symbiflow00:17
*** space_zealot has quit IRC00:53
*** proteusguy has quit IRC02:40
*** Bertl is now known as Bertl_zZ03:01
*** citypw has joined #symbiflow03:16
*** citypw has quit IRC03:32
mithroduck2: You are right about the output be hilarious!
mithroduck2: No, you should log a bug about that06:11
mithroduck2: (Regarding the vtr errors)06:15
*** proteusguy has joined #symbiflow06:34
*** proteusguy has quit IRC06:40
*** OmniMancer has joined #symbiflow07:15
*** futarisIRCcloud has joined #symbiflow08:00
sf-slack2<acomodi> duck2, mithro: There is an open PR I am dealing with to make the error/warnings optional:
tpbTitle: [WIP] vpr: added optional vpr_throw to demote errors to warnings by acomodi · Pull Request #66 · SymbiFlow/vtr-verilog-to-routing · GitHub (at
*** futarisIRCcloud has quit IRC10:10
*** citypw has joined #symbiflow10:23
*** citypw has quit IRC10:28
*** proteusguy has joined #symbiflow10:41
*** proteusguy has quit IRC11:04
*** futarisIRCcloud has joined #symbiflow11:06
*** Bertl_zZ is now known as Bertl11:30
*** Miyu has joined #symbiflow11:36
*** hackkitten has quit IRC11:38
*** Miyu is now known as hackkitten11:57
*** proteusguy has joined #symbiflow12:13
*** futarisIRCcloud has quit IRC14:06
sf-slack2<mkurc> @mithro I checked Verilator for attribute support and it seems that all cases relevant to us (parameters and port connections) are supported.15:17
sf-slack2<mkurc> @mithro There is still no feedback on my PR to the Icarus Verilog.15:19
sf-slack2<mkurc> @mithro Have you consulted with Yosys devs the idea of having wires that mock parameters? (for having attributes on them and preserving default assigned value) Are they willing to accept such a solution?15:26
*** OmniMancer has quit IRC15:48
mithromkurc: Do you want to chat with Eddie directly?16:34
sf-slack2<mkurc> @mithro It makes sense. But unfortunately I don't have time for it today.18:00
mithromkruc: I sent an introduction18:04
sf-slack2<mkurc> @mitho: Saw it. Thanks.18:06
hackerfooIt looks like no IOI bits need to set to pass the signal through. I built "buttons" with Vivado, then converted to FASM, and then back to bitstream without the ROI bitstream, and it works. There are no IOI features in the FASM, just IOB. So I can ignore the IO logic for now.22:50
litghosthackerfoo: Yes22:52
litghosthackerfoo: Make sure when you do those kind of experiments you run bit2fasm with "--verbose" to ensure unknown bits are reported22:53
litghosthackerfoo: There should be 0 of them, but best to check22:53
hackerfoolitghost: I did, and there were no unknown bits.22:54
litghosthackerfoo: There is one IOI bit you do need:
tpbTitle: prjxray-db/design.json at master · SymbiFlow/prjxray-db · GitHub (at
litghosthackerfoo: Otherwise ILOGIC will invert the signal22:55
litghosthackerfoo: It is worth noting that there is no way to actually invert IBUF signals without the use of an IDDR, which is not going to be the baseline ROI breakout22:57
hackerfooWhy is it IOB_* instead of IOI_*?22:59
litghosthackerfoo: Frankly there is not a strong way to seperate the too from a bitstream perspective, so the initial IOB/IOI fuzzers emit a mix of both23:01
litghosthackerfoo: I don't know if tilegrid even has an entry for IOI tiles yet, as they overlap strongly with the IOB bits23:02
litghostIOI -> RIOB33.IOB_Y0.OSERDESE.DATA_WIDTH.2 30_12723:03
litghostIOB -> RIOB33.IOB_Y0.INOUT 30_6723:03
litghosthackerfoo: So we can restructure the fuzzers and tilegrid, but the bits are strongly interwoven23:03

Generated by 2.13.1 by Marius Gedminas - find it at!