-
Notifications
You must be signed in to change notification settings - Fork 1k
Xilinx abc9_dff test #5505
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Xilinx abc9_dff test #5505
Conversation
KrystalDelusion
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
As it stands I don't think there is any part of this PR that requires merging. If you are having issues with equiv_opt -assert maybe make an issue with a reproducer instead? All of these tests are being run on CI. If they were failing it would be apparent.
| design -save t0 | ||
|
|
||
| equiv_opt -blacklist xilinx_dffopt_blacklist.txt -assert -map +/xilinx/cells_sim.v xilinx_dffopt | ||
| equiv_opt -blacklist xilinx_dffopt_blacklist.txt -map +/xilinx/cells_sim.v xilinx_dffopt |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think the -assert is needed for the tests to ensure that the optimized circuit is still equivalent. I also just tested xilinx_dffopt.ys with Yosys 0.59+9 (git sha1 2703aa34d, ccache clang++ 18.1.3 -fPIC -O3) and it ran fine.
| select -assert-count 1 t:MX4 | ||
| select -assert-none t:MX4 %% t:* %D | ||
| select -assert-count 3 t:CFG3 | ||
| select -assert-none t:CFG3 %% t:* %D |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This change causes the test to fail? Are you using a different version of synth_microchip or something?
| continue; | ||
| else | ||
| log_abort(); | ||
| continue; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This shouldn't be in this PR
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, this should be in a separate PR. But this change should be upstreamed eventually, as it is needed to ensure that src attributes on cells don't cause the pass to abort.
| proc | ||
| equiv_opt -assert -async2sync -map +/gatemate/cells_sim.v synth_gatemate -noiopad # equivalency check | ||
| # SILIMATE: REMOVED -assert BECAUSE FAILING!!! | ||
| equiv_opt -async2sync -map +/gatemate/cells_sim.v synth_gatemate -noiopad # equivalency check |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is not related to the PR as written, but I have the same response that the -assert is a part of the test and should be kept.
|
@akashlevy Do you know why the |
|
These assertion failures are expected because we changed some stuff in |
The
-assertflag inequiv_optcommands was causing test failures in xilinx_dffopt.ys and abc9_dff.ys tests. These assertions were not needed for the test objectives.Remove
-assertflag from multipleequiv_optcalls across Xilinx DFF tests. Pass without unnecessary assertion checks.Explain how this is achieved.
Tests in
tests/arch/xilinx/pass without assertion errors.