01:03 <mithro> litghost: Dunno if it makes sense to use something like https://github.com/xmlrunner/unittest-xml-reporting to output the QoR results....

01:03 <tpb> Title: GitHub - xmlrunner/unittest-xml-reporting: unittest-based test runner with Ant/JUnit like XML reporting. (at github.com)

13:42 <sf-slack2> <mkurc> @litghost I identified through minitests that we are probably missing 2 more bits related to DOUTMUX.AMC31 input for SLICEM. See https://github.com/SymbiFlow/prjxray/issues/927

13:42 <tpb> Title: Documents bits for DOUTMUX.AMC31 input · Issue #927 · SymbiFlow/prjxray · GitHub (at github.com)

14:04 <sf-slack2> <mkurc> @litghost As for DI1MUXes I have a hypothesis that they are controlled also by the RAM bit. When the associated LUT is in SRL32 mode then the mux selects between eg. CI and DMC31. But when in RAM mode it selects between eg. CI and DI.

14:04 <sf-slack2> <mkurc> @litghost I tried to route to DI1 of C6LUT (in SRL32 mode) through DI manually in vivado. It did not allowed me to do that.

19:50 <mithro> litghost: ITS ZYNQ NOT ZYNC !!!!!

21:28 <mithro> litghost / hackerfoo: Andre pushed some examples of the stuff he is doing to https://github.com/dehon/prflow_leaf_bench

21:28 <tpb> Title: GitHub - dehon/prflow_leaf_bench: Benchmarkset of PRflow Leaf logic for Rosetta Benchmarks (at github.com)

22:57 <sf-slack2> <butta> @me1 @litghost What is the current status of the clock network support? We're running into over congestion that involves clocks that vpr is routing on general interconnect.

22:58 <litghost> E.g. more than one clock or a divided clock? Basically none at this time

22:58 <litghost> It is a goal next month or two to add first clock network support

22:59 <sf-slack2> <butta> 2 clocks. They are independent.

23:00 <litghost> If you need that in the short term, you'll need to create a new ROI

23:00 <sf-slack2> <butta> Well, that explains what we're seeing :slightly_smiling_face:

23:01 <litghost> It is a known limitation at this time

23:01 <sf-slack2> <butta> So should it work with a single clock?

23:02 <litghost> Single clock is supported, you must tied to the clock pin for the ROI

23:02 <litghost> The clock within the ROI after a BUFH

23:03 <litghost> Which is the CMT clock fanout

23:04 <sf-slack2> <butta> So with a single clock is it routing through general interconnect or separate clock resources?

23:05 <litghost> Clock resources