Implementing outlining has greatly increased the amount of debug
information in a typical build, and consequently exposed performance
issues in C++ compilers, which are similar for both GCC and Clang;
the compile time of Minerva SoC SRAM increased almost twofold.
Although one would expect the slowdown to be caused by the increased
use of templates in `debug_eval()`, it is actually almost entirely
attributable to optimizations and codegen for `debug_items()`.
Fortunately, it is neither possible nor desirable to optimize
`debug_items()`: in most cases it is called exactly once, and its
body is a linear sequence of calls with unique arguments.
This commit turns off optimizations for `debug_items()` on GCC and
Clang, improving -Os compile time of Minerva SoC SRAM by ~40% (!)
Before this commit, if a sequence of wires assigned in a chain would
terminate on a cell, none of the wires would get marked as aliases,
and typically all of the public wires would get outlined. The reason
for this behavior is that alias analysis predates outlining and in
fact runs before it.
After this commit, alias analysis runs after outlining and considers
outlined wires valid aliasees. More importantly, if the chained wires
contain any valid aliasees, then all of the wires are aliased to
the one that is topologically deepest.
Aliased wires incur virtually no overhead for the VCD writer, unlike
outlined wires that would otherwise take their place. On Minerva SoC
SRAM, size of the full VCD dump is reduced by ~65%, and throughput
is increased by ~55%.
Aggressive wire localization and inlining is necessary for CXXRTL to
achieve high performance. However, that comes with a cost: reduced
debug information coverage. Previously, as a workaround, the `-Og`
option could have been used to guarantee complete coverage, at a cost
of a significant performance penalty.
This commit introduces debug information outlining. The main eval()
function is compiled with the user-specified optimization settings.
In tandem, an auxiliary debug_eval() function, compiled from the same
netlist, can be used to reconstruct the values of localized/inlined
signals on demand. To the extent that it is possible, debug_eval()
reuses the results of computations performed by eval(), only filling
in the missing values.
Benchmarking a representative design (Minerva SoC SRAM) shows that:
* Switching from `-O4`/`-Og` to `-O6` reduces runtime by ~40%.
* Switching from `-g1` to `-g2`, both used with `-O6`, increases
compile time by ~25%.
* Although `-g2` increases the resident size of generated modules,
this has no effect on runtime.
Because the impact of `-g2` is minimal and the benefits of having
unconditional 100% debug information coverage (and the performance
improvement as well) are major, this commit removes `-Og` and changes
the defaults to `-O6 -g2`.
We'll have our cake and eat it too!
"Elision" in this context is an unusual and not very descriptive term
whereas "inlining" is common and straightforward. Also, introducing
"inlining" makes it easier to introduce its dual under the obvious
name "outlining".
Before this commit, a cell's input was always assigned like:
p_cell.p_input = (value...);
If `p_input` is buffered (e.g. if the design is built at -O0), this
is not correct. (In practice, this breaks clocking.) Unfortunately,
the incorrect design was compiled without diagnostics because wire<>
was move-assignable and also implicitly constructible from value<>.
After this commit, cell inputs are no longer incorrectly assumed to
always be unbuffered, and wires are not assignable from values.
RTL contract violations and C++ contract violations are different:
the former depend on the netlist and will never violate memory safety
whereas the latter may. When loading a CXXRTL simulation into another
process, RTL contract violations should generally not crash it, while
C++ contract violations should.
Although it is always possible to destroy and recreate the design to
simulate a power-on reset, this has two drawbacks:
* Black boxes are also destroyed and recreated, which causes them
to reacquire their resources, which might be costly and/or erase
important state.
* Pointers into the design are invalidated and have to be acquired
again, which is costly and might be very inconvenient if they are
captured elsewhere (especially through the C API).
* backends/blif: Remove unused vector of strings
For reasons that are unclear to me, this was being used to store every
result of `cstr` before returning them. The vector was never accessed otherwise,
resulting in a huge unnecessary memory sink when emitting to BLIF.
* backends/blif: Remove CSTR macro
* backends/blif: Actually call str()
In most cases, a CXXRTL simulation would use a top module, either
because this module serves as an entry point to the CXXRTL C API,
or because the outputs of a top module are unbuffered, improving
performance. Taking this into account, the CXXRTL backend now runs
`hierarchy -auto-top` if there is no top module. For the few cases
where this behavior is unwanted, it now accepts a `-nohierarchy`
option.
Fixes#2373.
This can be useful to determine whether the wire should be a part of
a design checkpoint, whether it can be used to override design state,
and whether driving it may cause a conflict.
Before this commit, the meaning of "sync def" included some flip-flop
cells but not others. There was no actual reason for this; it was
just poorly defined.
After this commit, a "sync def" means that a wire holds design state
because it is connected directly to a flip-flop output, and may never
be unbuffered. This is not affected by presence of async inputs.
This can be useful to distinguish e.g. a combinatorially driven wire
with type `CXXRTL_VALUE` from a module input with the same type, as
well as general introspection.
The only difference between "RTLIL" and "ILANG" is that the latter is
the text representation of the former, as opposed to the in-memory
graph representation. This distinction serves no purpose but confuses
people: it is not obvious that the ILANG backend writes RTLIL graphs.
Passes `write_ilang` and `read_ilang` are provided as aliases to
`write_rtlil` and `read_rtlil` for compatibility.
This commit adds support for real-valued parameters in blackboxes. Additionally,
parameters now retain their types are no longer all encoded as strings.
There is a caveat with this implementation due to my limited knowledge of yosys,
more specifically to how yosys encodes bitwidths of parameter values. The example
below can motivate the implementation choice I took. Suppose a verilog component
is declared with the following parameters:
parameter signed [26:0] test_signed;
parameter [26:0] test_unsigned;
parameter signed [40:0] test_signed_large;
If you instantiate it as follows:
defparam <inst_name> .test_signed = 49;
defparam <inst_name> .test_unsigned = 40'd35;
defparam <inst_name> .test_signed_large = 40'd12;
If you peek in the RTLIL::Const structure corresponding to these params, you
realize that parameter "test_signed" is being considered as a 32-bit value
since it's declared as "49" without a width specifier, even though the parameter
is defined to have a maximum width of 27 bits.
A similar issue occurs for parameter "test_unsigned" where it is supposed to take
a maximum bit width of 27 bits, but if the user supplies a 40-bit value as above,
then yosys considers the value to be 40 bits.
I suppose this is due to the type being defined by the RHS rather than the definition.
Regardless of this, I emit the same widths as what the user specifies on the RHS when
generating firrtl IR.
Previous blackbox components were just emitted with their interface ports,
but their generic parameters were never emitted and it was therefore
impossible to customize them.
This commit adds support for blackbox generic parameters, though support
is only provided for INTEGER and STRING parameters. Other types of
parameters such as DOUBLEs, ..., would result in undefined behavior here.
This allows the emission of custom extmodule instances such as the following:
extmodule fourteennm_lcell_comb_<instName>:
input cin: UInt<1>
output combout: UInt<1>
output cout: UInt<1>
input dataa: UInt<1>
input datab: UInt<1>
input datac: UInt<1>
input datad: UInt<1>
input datae: UInt<1>
input dataf: UInt<1>
input datag: UInt<1>
input datah: UInt<1>
input sharein: UInt<1>
output shareout: UInt<1>
output sumout: UInt<1>
defname = fourteennm_lcell_comb
parameter extended_lut = "off"
parameter lut_mask = "b0001001000010010000100100001001000010010000100100001001000010010"
parameter shared_arith = "off"
Refer to the SMT-LIB specification, section 4.1.7. According to the spec, some options can only be specified in `start` mode. Once the solver sees `set-logic`, it moves to `assert` mode.
whitequark
Miodrag Milanovic
Sahand Kashani
William Woodruff
Marcelina Kościelnicka
Jakob Wenzel
N. Engelhardt
Xiretza
Noah Moroze
Andy Knowles
Alberto Gonzalez
Claire Wolf