| *** tpb has joined #yosys | 00:00 | |
| *** _whitelogger has quit IRC | 02:33 | |
| *** _whitelogger has joined #yosys | 02:36 | |
| *** perryprog has joined #yosys | 03:33 | |
| *** citypw has joined #yosys | 04:33 | |
| *** nrossi has joined #yosys | 06:23 | |
| *** proteusguy has quit IRC | 06:35 | |
| *** proteusguy has joined #yosys | 06:48 | |
| *** svenn4 has quit IRC | 06:58 | |
| *** _whitelogger has quit IRC | 07:57 | |
| *** _whitelogger has joined #yosys | 08:00 | |
| *** dys has joined #yosys | 08:06 | |
| *** Jybz has joined #yosys | 08:12 | |
| *** _whitelogger has quit IRC | 11:42 | |
| *** _whitelogger has joined #yosys | 11:45 | |
| *** lutsabound has joined #yosys | 14:17 | |
| *** gorbak25 has quit IRC | 15:39 | |
| *** X-Scale has quit IRC | 16:21 | |
| *** X-Scale has joined #yosys | 16:24 | |
| *** X-Scale has quit IRC | 16:38 | |
| *** X-Scale` has joined #yosys | 16:41 | |
| *** X-Scale` is now known as X-Scale | 16:43 | |
| *** Boko_dabce has joined #yosys | 16:48 | |
| *** Boko_dabce has quit IRC | 16:49 | |
| *** Boko_debac has joined #yosys | 16:53 | |
| *** Boko_debac has quit IRC | 16:59 | |
| *** citypw has quit IRC | 17:03 | |
| *** Boko_edeea has joined #yosys | 17:55 | |
| ZirconiumX | So, for the $alu cell, what would be a cell that would change BI at runtime as opposed to it being constant? | 18:00 |
|---|---|---|
| *** Boko_edeea has quit IRC | 18:01 | |
| *** Jybz has quit IRC | 18:10 | |
| *** Jybz has joined #yosys | 18:11 | |
| *** Boko_bbbde has joined #yosys | 18:12 | |
| *** Boko_bbbde has quit IRC | 18:15 | |
| ZirconiumX | daveshah, maybe? | 18:16 |
| daveshah | Hypothetically an adder/subtractor | 18:16 |
| daveshah | I don't know if Yosys maps these yet (from add sub and mux) | 18:17 |
| daveshah | But it's something I've wanted to look at at some point | 18:17 |
| ZirconiumX | I'm wondering if trying to specialise based on CI and/or BI is worth it or not | 18:17 |
| *** Boko_aeadc has joined #yosys | 18:19 | |
| daveshah | BI=1 is a common case (subtractor) | 18:19 |
| daveshah | Fixed BI and CI is much more common than variable | 18:19 |
| ZirconiumX | I've been inspecting the output of Quartus and it's interesting (to me) that comparisons are not done with carry chains | 18:20 |
| ZirconiumX | Which is actually probably cmp2lut | 18:20 |
| *** Boko_aeadc has quit IRC | 18:28 | |
| *** rombik_su has joined #yosys | 18:35 | |
| daveshah | Not just cmp2lut, we should probably have a mode in Yosys not to convert any compare to soft logic | 18:44 |
| mwk | tbh I think we shouldcompletely dump convering compares to sub | 19:20 |
| mwk | it's completely bonkers | 19:21 |
| *** develonepi3 has quit IRC | 19:32 | |
| daveshah | Agreed | 19:57 |
| daveshah | If nothing else because it prevents any serious downstream optimisation | 19:58 |
| mwk | well it's also horribly inefficient | 20:16 |
| mwk | at least on xilinx | 20:16 |
| mwk | you could reduce area 2× by using the carry chain in a better way for the variable inputs case (two pairs of input bits per LUT) | 20:17 |
| mwk | and at least 5× when one of the inputs is const | 20:17 |
| *** Jybz has quit IRC | 20:40 | |
| *** nrossi has quit IRC | 21:12 | |
| *** lutsabound has quit IRC | 21:17 | |
| *** lutsabound has joined #yosys | 22:33 | |
| *** rombik_su has quit IRC | 22:47 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!