12/28/2022 0 Comments Xilinx ise 14.7 closes on open fileIt produced a large number of metavalue warnings from package numeric_std indicating a lack of initial values and updates (to_integer warnings per clock) for signed and/or unsigned objects. I analyzed, elaborated and simulated CHIP8 using a rudimentary testbench providing 25 MHz clock and reset event, static '0's or '1's for other inputs. The extended identifiers haven't been a problem with previous output from the same VHDL generator (fed into the same XST version), so I'll look into the block statements, thanks! – Cactus 23 hours yeah it seems CLaSH converts everything to 64-bit Ints for some reason, but as long as everything gets converted back to its original small width at its consumption, it shouldn't be a problem should it? – Cactus 23 hours I can't try it out right now, but will give it a shot if it can help: are you saying the unguarded block statements don't make any semantic difference and just flattening their contents into the main begin.end scope would/could help? – Cactus 20 hours ago Some deleted comments from about how inefficient the generated VHDL is, and how all those unnecessary conversions to/from Int64 are putting extra pressure on the about MCV: well my basic problem is that I can't start to make it minimal until I know where to even start, since the error message is not pointing me to any meaningful location. There are already several comments gone, so I'm copying all his comments from the SO question to here: A very helpful StackOverflow user has managed to simulate the CLaSH-generated VHDL, so maybe this is more a performance problem (as in, CLaSH generates unnecessarily complex and inefficient HDL) than a bug, which turns into a bug when combined with the ISE scalability problem that causes it to get confused on the over-complicated HDL.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |