There is no practical benefit from using `const memory` for ROMs;
it uses an std::vector internally, which prevents contemporary
compilers from constant-propagating ROM contents. (It is not clear
whether they are permitted to do so.)
However, there is a major benefit from using non-const `memory` for
ROMs, which is the ability to dynamically fill the ROM for each
individual simulation.
This adds the new rewrite rule. But it's still missing a check that makes
sure the new rewrite rule is actually a valid substitute in the always
block being processed. Therefore the new rewrite rule is just disabled
for now.
Signed-off-by: Claire Wolf <claire@symbioticeda.com>
This brings the documented behavior for these cells in line with
$_DFFSR_* and $_DLATCHSR_*, which is that R has priority over S.
The models were already reflecting that behavior.
Also get rid of sim-synth mismatch in the models while we're at it.
This pass is a proper subset of opt_rmdff, which is called by opt, which
is called by every synth flow in the coarse part. Thus, it never
actually does anything and can be safely removed.
Before this commit, enum values were serialized as attributes of form
\enum_<width>_<value>
where <value> was a decimal signed integer.
This has multiple drawbacks:
* Enums with large values would be hard to process for downstream
tooling that cannot parse arbitrary precision decimals. (In fact
Yosys also did not correctly process enums with large values,
and would overflow `int`.)
* Enum value attributes were not confined to their own namespace,
making it harder for downstream tooling to enumerate all such
attributes, as opposed to looking up any specific value.
* Enum values could not include x or z, which are explicitly
permitted in the SystemVerilog standard.
After this commit, enum values are serialized as attributes of form
\enum_value_<value>
where <value> is a bit sequence of the appropriate width.
By operating at a layer of abstraction over the rather clumsy Intel primitives,
we can avoid special hacks like `dffinit -highlow` in favour of simple techmapping.
This also makes the primitives much easier to manipulate, and more descriptive
(no more cyclonev_lcell_comb to mean anything from a LUT2 to a LUT6).
In particular inside class declarations, a static const
assignment is technically not a definition, while constexpr is.
Signed-off-by: Henner Zeller <h.zeller@acm.org>
Detect the places in the $alu where the carry bit is constant (due to
const A[i] == B[i] ^ BI) and split it into smaller $alu at these points.
Also, make the existing const-carry detection for low bits more generic
(now handles cases where both BI and CI are constant, but not equal to
one another).
Fixes#1912.
whitequark
David Shah
Eddie Hung
Claire Wolf
Marcelina Kościelnicka
Alberto Gonzalez
Miodrag Milanović
N. Engelhardt
Dan Ravensloft
Claire Wolf
Henner Zeller