The input to a shift operation is padded.
This reduced the final number of MUX cells
but during techmap it can create huge
temporary multiplexers in the log shifter.
This significantly increases runtime and resources.
A limit is added with a warning when it is used.
The `has_srst`` case was checking `sig_ce` instead of `sig_srst` due to
a copy and paste error.
This would crash when `has_ce` was false and could incorrectly determine
that an initial value is unused when `has_ce` and `has_srst` are both
set.
Rename formal cells in addition to witness signals. This is required to
reliably track individual property states for the non-smtbmc flows.
Also removes a misplced `break` which resulted in only partial witness
renaming.
Processes can contain `MemWriteAction` entries which are invisible to
most passes operating on memories but which will be lowered to write
ports later on by `proc_memwr`. For that reason we can get corrupted
RTLIL if we sequence the memory passes before `proc`. Address that by
making the affected memory passes ignore modules with processes.
This fixes an issue introduced in commit 26644ea due to which flip-flops
are inadvertently ignored when building up driver map. The mentioned
commit wasn't without functional change after all.
Depending on the WIN32 compilation mode, PathMatchSpec may expect a LPCSTR or
LPCWSTR argument. char* is only convertable to LPCSTR, so use that
implementation
Signed-off-by: Austin Rovinski <rovinski@nyu.edu>
This is an alternative to setting the dont_use property in lib. This brings
dfflibmap in parity with the abc pass for dont_use.
Signed-off-by: Austin Rovinski <rovinski@nyu.edu>
See the test case. PROC_ROM will consider this for evaluation, even
though -- without any actions -- lhs is empty (but still "uniform").
A zero-width memory is constructed, which later fails check with:
ERROR: Assert `width != 0' failed in kernel/mem.cc:518.
Ensure we don't proceed if there's nothing to encode.
Fixes a bug in the handling of the recently introduced $check cells.
Both $check and $print cells in clk2fflogic are handled by the same code
and the existing tests for that were only using $print cells. This
missed a bug where the additional A signal of $check cells that is not
present on $print cells was dropped due to a typo, rendering $check
cells non-functional.
Also updates the tests to explicitly cover both cell types such that
they would have detected the now fixed bug.
Checks to see if a cell is of type ff in the liberty,
and keeps track of an additional area value.
```
Chip area for module '\addr': 92.280720
Sequential area for module '\addr': 38.814720
```
Signed-off-by: Ethan Mahintorabi <ethanmoon@google.com>
This fixes hierarchy when used with cell libraries that were loaded with
-defer and also makes more of the hierarchy visible to the auto-top
heuristic.
This adds support for `$check` cells in chformal and adds a `-lower`
mode which converts `$check` cells into `$assert` etc. cells with a
`$print` cell to output the `$check` message.
If the offset is larger than the signal itself,
meaning the signal is completely shifted out,
it tried to extract a negative amount of bits from the old signal.
This RTL pattern is suspicious since it is a complicated way of
arriving at a constant value, so we warn the user.
This allows tools like SBY to capture the $display output independent
from anything else sim might log. Additionally it provides source and
hierarchy locations for everything printed.
The previous version could easily generate a large amount of padding
when the constant factor was significantly larger than the width of the
shift data input. This could lead to huge amounts of logic being
generated before then being optimized away at a huge performance and
memory cost.
Additionally and more critically, when the input width was not a
multiple of the constant factor, the input data was padded with 'x bits
to such a multiple before interspersing the 'x padding needed to align
the selectable windows to power-of-two offsets.
Such a final padding would not be correct for shifts besides $shiftx,
and the previous version did attempt to remove that final padding at the
end so that the native zero/sign/x-extension behavior of the shift cell
would be used, but since the last selectable window also got
power-of-two padding appended after the padding the code is trying to
remove got added, it did not actually fully remove it in some cases.
I changed the code to only add 'x padding between selectable windows,
leaving the last selectable window unpadded. This omits the need to add
final padding to a multiple of the constant factor in the first place.
In turn, that means the only 'x bits added are actually impossible to
select. As a side effect no padding is added when the constant factor is
equal to or larger than the width of the shift data input, also solving
the reported performance bug.
This fixes#4056
Remove duplicate %.pmg -> %_pm.h pattern. One of the duplicates overrode
the other, and in some conditions there were build races as to whether
the target directory for the generated header would exist. Instead have
a single rule which is properly generalized.
Generalize what was formerly the unsigned-only architecture to support
both signed and unsigned multiplication, use that as default, and set
aside the special low-power architecture that was formerly used for
signed multipliers.
- moved all selection and filtering logic to the match block
- applied less-verbose code suggestions
- removed constraint on number of bits in shift-amount
- added check for possible wrap-arround in the operation
Add a separate shiftmul pattern to match on left shifts which implement
demuxing. This mirrors the right shift pattern matcher but is probably
best kept separate instead of merging the two into a single matcher.
In any case the diff of the two matchers should be easily readable.
The `opt_expr` pass running before `peepopt` can interfere with the
detection of a shiftmul pattern due to some of the bottom bits of the
shift amount being replaced with constant zero. Extend the detection to
cover those situations as well.
`memory_nordff` has the advantage that it can be called just ahead of
the simulation step no matter whether the clocked read port has been
inferred or was explicitly instantiated in a flow.
In commit fedd12261 ("booth: Move away from explicit `Wire` pointers")
a bug was introduced when checking for vacant slots in arrays holding
some intermediate results. Non-wire SigBit values were taken to imply
a vacant slot, but actually a constant one can make its way into those
results, if the multiplier cell configuration is just right. Fix the
vacancy check to address the bug.
The main speedup comes from swithing from using a SHA1 hash to std::hash<std::string>. There is no need to use an expensive cryptographic hash for fingerprinting in this context.
This adds the -noinitstate option which is required to simulate
counterexamples to induction with yw-cosim. Also add handling for
$initstate cells for non-co-simulation.
For the basic single-bit operations, opt for gate cells (`$_AND_` etc.)
instead of the coarse cells (`$and` etc.). For the emission of cells
move to the conventional module methods (`module->addAndGate`) away
from the local helpers. While at it, touch on the surrounding code.
To represent intermediate signals use the `SigBit`/`SigSpec` classes as
is customary in the Yosys codebase. Do not pass around `Wire` pointers
unless we have special reason to.
* Speed up the autoname pass by 2x. This is accomplished by only constructing IdString objects for plain strings that have a higher score.
* Defer creating IdStrings even further. This increases the speedup to 3x.
This does not correctly handle an `$overwrite_tag` on a module output,
but since we currently require the user to flatten the design for
cross-module dft, this cannot be observed from within the design, only
by manually inspecting the signals in the design.
This is still missing a mode to rewrite $overwrite_tag and $original_tag
by injecting $set_tag and $get_tag in the right places. It's also
missing bit-precise propagation models for shifts and arithmetic and
requires the design to be flattened.
* Fixes a non-deterministic polarity error for $eqx/$nex cells
* Fixes a deterministic polarity error for $_NOR_ and $_ORNOT_ cells
* Generates hdlnames when xprop is run after flatten
ABC's read_lib command has a dont_use
cell list that is configurable by the user.
This PR exposes that option to Yosys.
See
5405d4787a/src/map/scl/scl.c (L285)
for documentation on this option.
Signed-off-by: Ethan Mahintorabi <ethanmoon@google.com>