3
0
Fork 0
mirror of https://github.com/YosysHQ/yosys synced 2025-04-22 08:35:32 +00:00
Commit graph

1893 commits

Author SHA1 Message Date
Eddie Hung
7cfdf4ffa7 verilog: fix $specify3 check 2020-02-13 12:42:04 -08:00
N. Engelhardt
e069259a53
Merge pull request #1679 from thasti/delay-parsing
Fix crash on wire declaration with delay
2020-02-13 12:01:27 +01:00
whitequark
c34d7b13f4 ast: avoid intermediate wires/assigns when lowering to AST_MEMINIT.
Before this commit, every initial assignment to a memory generated
two wires and four assigns in a process. For unknown reasons (I did
not investigate), large amounts of assigns cause quadratic slowdown
later in the AST frontend, in processAst/removeSignalFromCaseTree.
As a consequence, common and reasonable Verilog code, such as:
  reg [`WIDTH:0] mem [0:`DEPTH];
  integer i; initial for (i = 0; i <= `DEPTH; i++) mem[i] = 0;
took extremely long time to be processed; around 80 s for a 8-wide,
8192-deep memory.

After this commit, initial assignments where address and/or data are
constant (after `generate`) do not incur the cost of intermediate
wires; expressions like `mem[i+1]=i^(i<<1)` are considered constant.
This results in speedups of orders of magnitude for common memory
sizes; it now takes merely 0.4 s to process a 8-wide, 8192-deep
memory, and only 5.8 s to process a 8-wide, 131072-deep one.

As a bonus, this change also results in nontrivial speedups later
in the synthesis pipeline, since pass sequencing issues meant that
all of these intermediate wires were subject to transformations such
as width reduction, even though they existed solely to be constant
folded away in `memory_collect`.
2020-02-07 00:41:54 +00:00
Rodrigo Alejandro Melo
da485dc007 Modified $readmem[hb] to use '\' or '/' according the OS
Signed-off-by: Rodrigo Alejandro Melo <rmelo@inti.gob.ar>
2020-02-06 10:10:29 -03:00
Eddie Hung
fc33287bc1 verilog: instead of modifying localparam size, extend init constant expr 2020-02-05 17:19:42 -08:00
Stefan Biereigel
b844b078db correct wire declaration grammar for #1614 2020-02-03 21:29:40 +01:00
Rodrigo Alejandro Melo
313a425bd5 Merge branch 'master' of https://github.com/YosysHQ/yosys
Solved a conflict into the CHANGELOG

Signed-off-by: Rodrigo Alejandro Melo <rmelo@inti.gob.ar>
2020-02-03 10:56:41 -03:00
Rodrigo Alejandro Melo
71f3afb9a2 Replaced strlen by GetSize into simplify.cc
As recommended in CodingReadme.

Signed-off-by: Rodrigo Alejandro Melo <rmelo@inti.gob.ar>
2020-02-03 10:44:09 -03:00
David Shah
4bfd2ef4f3 sv: Improve handling of wildcard port connections
Signed-off-by: David Shah <dave@ds0.me>
2020-02-02 16:12:33 +00:00
David Shah
5df591c023 hierarchy: Resolve SV wildcard port connections
Signed-off-by: David Shah <dave@ds0.me>
2020-02-02 16:12:33 +00:00
David Shah
50f86c11b2 sv: Add lexing and parsing of .* (wildcard port conns)
Signed-off-by: David Shah <dave@ds0.me>
2020-02-02 16:12:33 +00:00
David Shah
9f5613100b
Merge pull request #1647 from YosysHQ/dave/sprintf
ast: Add support for $sformatf system function
2020-02-02 14:53:46 +00:00
Rodrigo Alejandro Melo
b4c30cfc8d Fixed a bug in the new feature of $readmem[hb] when an empty string is provided
Signed-off-by: Rodrigo Alejandro Melo <rodrigomelo9@gmail.com>
2020-02-01 17:03:56 -03:00
Rodrigo Alejandro Melo
d74b9604e3 Modified the new search for files of $readmem[hb] to be backward compatible
Signed-off-by: Rodrigo Alejandro Melo <rodrigomelo9@gmail.com>
2020-01-31 22:10:51 -03:00
Rodrigo Alejandro Melo
7b3fe404ab $readmem[hb] file inclusion is now relative to the Verilog file
Signed-off-by: Rodrigo Alejandro Melo <rodrigomelo9@gmail.com>
2020-01-31 18:20:22 -03:00
Claire Wolf
2ce7a0d369
Merge pull request #1667 from YosysHQ/clifford/verificnand
Add Verific support for OPER_REDUCE_NAND
2020-01-30 19:55:53 +01:00
Claire Wolf
60876ce183
Merge pull request #1503 from YosysHQ/eddie/verific_help
`verific` pass to print help message when command syntax error
2020-01-30 18:05:16 +01:00
Claire Wolf
ffadaddab5
Merge pull request #1654 from YosysHQ/eddie/sby_fix69
verific: unflatten struct ports
2020-01-30 18:03:35 +01:00
Claire Wolf
23c44afaed Add Verific support for OPER_REDUCE_NAND
Signed-off-by: Claire Wolf <clifford@clifford.at>
2020-01-30 18:01:13 +01:00
Eddie Hung
6d27d43727 Add and use SigSpec::reverse() 2020-01-28 10:37:16 -08:00
Eddie Hung
ce6a690d27 xilinx/ice40/ecp5: undo permuting LUT masks in lut_map
Now done in read_aiger
2020-01-27 13:30:27 -08:00
Eddie Hung
f443695a38 Merge remote-tracking branch 'origin/master' into eddie/verific_help 2020-01-27 10:34:10 -08:00
Eddie Hung
d730bba6d2 verific: no help() when no YOSYS_ENABLE_VERIFIC 2020-01-27 10:32:18 -08:00
Eddie Hung
7b445121cc verific: also unflatten for 'hierarchy' flow as per @cliffordwolf 2020-01-27 10:15:22 -08:00
Eddie Hung
c7fbe13db5 read_aiger: set abc9_box_seq attr 2020-01-24 13:11:43 -08:00
Eddie Hung
cccc0ae112 verific: unflatten struct ports 2020-01-24 10:12:52 -08:00
Eddie Hung
73526a6f10 read_aiger: also parse abc9_mergeability 2020-01-22 14:21:25 -08:00
Eddie Hung
cd093c00f8 read_aiger: discard LUT inputs with nodeID == 0; not < 2 2020-01-21 11:56:30 -08:00
Eddie Hung
7f728bc116 read_aiger: ignore constant inputs on LUTs 2020-01-21 11:16:50 -08:00
David Shah
22c967e35e ast: Add support for $sformatf system function
Signed-off-by: David Shah <dave@ds0.me>
2020-01-19 21:20:17 +00:00
Jeff Wang
98c6bd7630 fix bug introduced by not taking all of PeterCrozier's changes in 16ea4ea6
The if(str == node->str) is in fact necessary (otherwise causes generate
for in Multiplier_2D in tests/simple/multiplier.v to fail with error
message "Right hand side of 3rd expression of generate for-loop is not
constant!"). Note: in PeterCrozier's implementation, the break only
breaks out of the switch-case, not the outer for loop.
2020-01-17 02:07:42 -05:00
Jeff Wang
549deb6373 fix enum in generate blocks 2020-01-16 18:13:30 -05:00
Jeff Wang
41a0a93dcc allow enums to be declared at toplevel scope 2020-01-16 18:13:14 -05:00
Jeff Wang
cc2236d0c0 lexer doesn't seem to return TOK_REG for logic anymore 2020-01-16 18:08:58 -05:00
Jeff Wang
5ddf84d430 allow enum typedefs 2020-01-16 17:17:42 -05:00
Jeff Wang
16ea4ea61a partial rebase of PeterCrozier's enum work onto current master
I tried to keep only the enum-related changes, and minimize the diff. (The
original commit also had a lot of work done to get typedefs working, but yosys
has diverged quite a bit since the 2018-03-09 commit, with a new typedef
implementation.) I did not include the import related changes either.

Original commit:
"Initial implementation of enum, typedef, import.  Still a WIP."
881833aa73
2020-01-16 13:51:47 -05:00
Eddie Hung
03ce2c72bb Merge remote-tracking branch 'origin/master' into eddie/abc9_refactor 2020-01-15 16:42:16 -08:00
Eddie Hung
05c8858a90 read_aiger: $lut prefix in front 2020-01-15 14:31:32 -08:00
Eddie Hung
53a99ade9c Merge remote-tracking branch 'origin/master' into eddie/abc9_refactor 2020-01-14 11:46:56 -08:00
Eddie Hung
f63f76c372 read_aiger: also rename "$0" 2020-01-14 09:01:53 -08:00
Eddie Hung
2c65e1abac abc9: break SCC by setting (* keep *) on output wires 2020-01-13 21:45:27 -08:00
Eddie Hung
ee95fa959a read_aiger: uniquify wires with $aiger<autoidx> prefix 2020-01-13 21:28:27 -08:00
Eddie Hung
766e16b525 read_aiger: make $and/$not/$lut the prefix not suffix 2020-01-13 17:34:37 -08:00
Eddie Hung
d979648b7a read_aiger: more accurate debug message 2020-01-09 10:02:19 -08:00
Eddie Hung
943ea4bf9e read_aiger: do not double-count outputs for flops 2020-01-09 08:55:36 -08:00
Eddie Hung
2ca8c10e7a Merge remote-tracking branch 'origin/master' into eddie/abc9_refactor 2020-01-07 15:43:22 -08:00
Eddie Hung
2ac36031d4 read_aiger: consistency between ascii and binary; also name latches 2020-01-07 13:30:31 -08:00
Eddie Hung
8f5388ea5b read_aiger fixes 2020-01-07 11:59:57 -08:00
Eddie Hung
b94cf0c126 read_aiger: connect identical signals together 2020-01-07 11:43:28 -08:00
Eddie Hung
baba33fbd3 read_aiger: cope with latches and POs with same name 2020-01-07 11:22:48 -08:00