| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
| |
- move code to more appropriate locations
- apply clang-format to modified files
- fixed some cspell errors
- Applied emu.h rule.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- added support parameter FORCE_TRISTATE_LOGIC on mk28000
- add system device SYS_PULSE to generate pulse based on input changes
parameters are delay, pulse width, input polarity and output polarity
- add USE_SPEED_HACKS define to gtrak10 to enable the use of the above
Brings gtrak10 from 18% to 120%.
- Improved interlaced video support in fixfreq.cpp
Added two interlace modes: progressive and interlaced which can
be selected in machine configuration.
- more clang format work
- rename link to connection
- introduced typed aliases. This information may be used later to
rewrite netlists.
- added a three terminal base device
This will allow the analog code to be more easily comparable to
literature about MNA (Modal node analysis).
BJT code is now a lot better readable.
- fix doxygen consistency
- added an example for a cmos inverter based on NMOS/PMOS fets.
- TRUTHTABLE_START was renamed to TRUTH_TABLE
- truth tables are now enclosed in braces
- netlists are now enclosed in braces
- TRUTHTABLE_END() and NETLIST_END() removed from files
- Fixed static solver script
- Add file path to includes in nltool
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* netlist: fix bug, prepare for future changes and improve readability
- fix a bug where a net processing error may trigger a nullptr access
- applied some clang-tidy recommendations
- add no_return to plib::terminate
- properly encapsulate dynamic_cast usage
- more review of noexcept
- added a clang-format file. Over time, all source files will be
processed with clang-format
- Used clang format on a number of files
- Rewrote 74174
- all device constructors now use a struct to pass data on
to base classes. Neither netlist state nor the name are intended
to be used in a constructor. After the base class was
constructed, they can be accessed by state() and name().
- The device construction macros can now be removed. Changes to
the core will not need to be reflected in constructors.
- Change truth table macros so that going forward NETLIST_END and
TRUTH_TABLE_END can be replaced by a closing curly brace. netlists can
than use curly braces enclosed blocks.
- more clang-format
- removed some macros completely
- all derived classes from base_device_t now don't use macros
any longer.
- as a result, delegator_t was removed. This class was only used
to support macros :-(
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* netlist: fix bugs and more cpp instead of macros
- C-style comments converted to c++
- Fix crash in state saving code when an abort queue processing event
is pending.
- Fix a bug where a net could be twice in the queue.
- Convert more macros to c++
- fixed SUBTARGET=nl build
- fixed potential bugs which would allow a terminal to belong to more
than one net. This is not possible even for a short time.
- moved some member function definitions out-of-class.
- moved code out-of-class
- added constexpr where appropriate
- fixed mamenl build
- Cleanup and indentation
|
| |
|
|
|
|
|
|
|
|
|
| |
- nltool: Added support for building individual static solver files.
- Added pre-built documentation files for nltool and nlwav.
- plib: Extended typed_version to have patchlevel as well.
- plib: Compile more of ppmf.h with nvcc.
- create_devinc.py: Add include guards to the generated files.
- makefile: Recreate generated files when python code has changed.
- Reduced nld_base.h usage in the core.
- <oved extern template class declarations to where they belong.
- Generally improved code readability and fixed a lot of issues cspell reported.
|
| |
|
|
|
|
|
|
|
| |
- Added more documentation to palloc.h
- Made the use of allocation arenas more transparent throughout code
- palloc now uses c++17 alignment aware operator new.
This required further changes to the arena_deleter object to
track alignment and size of allocated objects to avoid the use
of alignment and size of base classes in case a unique_ptr was
cast to a unique_ptr of the base class.
|
| |
|
|
|
| |
- added cstddef to includes in ppmf.h
- more explicit comments in ppmf.h
- Fixed more typos and improved readiability
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- More c++, less macros
* Significantly reduced the use of unused_var and replaced it with
[[maybe_unused]]
* use enum class in ppmf.h
- Changes to testing code in ptest.h
* Catch exceptions in more places
* The verbosity of the output can now be controlled
* Display of test stats totals
- added support for cspell
- fixed various typos
- fixed SUBTARGET=nl build
- fixed more file permissions
- srcclean and add fix_permissions target to netlist makefile
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
netlist: pointer to member function code
- Address PR #9752 and issue #8590
- change comments in ppmf.h
- add support for function desciptors (IA64)
- sizeof(size_t) may be different on build targets
make plib::hash use uint64 to ensure consistent static solver
hashes across platforms (EMSCRIPTEN)
- Moved constructor and member functions outside the class declaration
with the exception of "call" member functions.
- Updated comment on accessing RDX
- Added more test cases
* Removed execution bit on certain files
I have separated out the raw processing code into `ppmf.cpp`. I have
also taken first steps to clean the abi identification.
The code submitted now supports:
a) The unknown virtual inheritance case with specific code for MSVC for
simple (i.e. scalar, void, reference, pointer) return types using the
optimized code. This is the test case only supported under MSVC and
which should return 7. Please have a look at the code. Basically
m_vptr_offset is an index into a pointer table. It needs to be
multiplied by `sizeof(void *)`.
b) If you enable PPMF_EXPERIMENTAL complex return type member functions
for MSVC will also use the optimized code path. This is explained in
detail in the code in ppmf, including caveats and possible solutions.
c) Enabling PPMF_USE_MAME_DELEGATES will use `src/lib/util/delegate.h`
as a replacement for `ppmf.h`. This allows to run the same tests
(`nltool -c tests` from standalone build) as for the ppmf code.
d) The code now also supports `clang-cl`
e) sizeof(size_t) may be different on build targets.
Make plib::hash use uint64 to ensure consistent static solver hashes
across platforms. This fixes loading static solvers (if enabled) on
EMSCRIPTEN.
|
| |
|
|
| |
nvcc 11.3 (latest cuda tools) has an issue with some auto x(some
variable) declarations.
|
| |
|
|
|
|
| |
- update netlist makefile to more recent compilers
- prototypes for NETLISTs in macros are now created create_devinc.py
- apply ctidy recommendations
- Python None in create_device.py
|
| |
|
|
| |
- Reduce the use of macros by converting defines into c++ constants.
- more minor code clean up.
|
| |
|
|
|
| |
- Added a first version of a FAQ.
- Use better error messages in pfunction.
- Made member functions static where appropriate in nld_solver.
|
| |
|
|
|
|
|
|
|
|
| |
copyright owners. Much thanks for that Couriersud, the main creator
and contributor, but also to Jonathan Gevaryahu and Sergey Svishchev.
There are small remnants in machine/NL_*, specifially breakout, pong
doubles and rebound that are also copyrighted by the DICE team,
whoever that means. They're not critical since they only concern
these drivers and not an important core subsystem.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
* These are used during creation and reset only. No need to
maintain them on net_t level.
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
* Under cirtum circumstances the splitter would create "ghost" solvers
consisting of terminals already used in another and complete
solver. This may impact all netlist which use opamps and thus
is committed early in the cycle.
* This commit adds functionality to instruct the splitter code to
include terminals which will not create matrix elements into
the parsing of net groups for solvers.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
* Devices like BJTs and FETs which are defined but not used will now
cause an error.
* An unused device will create an additional solver with a singular
matrix.
- This is adding unnecessary performance overhead.
- It complicates debugging because the unused device will cause an
arithmetic signal if used with --fperr (nltool).
* Fixed all validation errors.
|
| |
|
|
| |
* No longer added value.
|
| | |
|
| |
|
|
|
| |
- rename mat_cr.h to pmatrix_cr.h
- Optimization to the gmres solver.
- Simplifcation of vector operation code
|
| |
|
|
| |
* Consequently prefer to save on compile time and size.
|
| |
|
| |
* also fix an issue with netlist vs build
|
| |
|
|
|
|
|
|
|
|
|
| |
* rename some misleading type names
* remove callback_t and replace by better scalable approach
* hide implementations details
* move sources classes from putil.h to psources.h
* reduce code complexity
* improve parsing performance, parsing netlists now is twice as fast.
* fix issues around multi-byte string support
* moved psplit into pstrutil.h
|
| |
|
|
|
|
|
|
|
| |
* removed include directory src/lib/netlist from various genie files to
avoid potential issues.
* Code using netlist should use #include "netlist/*".
* Updated includes.
* Fixed standalone makefile depend target to properly deal with relative
paths.
|
| |
|
|
|
| |
* Exposing sub-device members is not best practice.
* The need for sub-devices is a clear indication that a netlist language
implementation would be a better solution.
|
| | |
|
| |
|
|
|
| |
* This is an infrastructure change to enable better error reporting
including file/source and line numbers in the future
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
* Example: PARAM(Solver.Solver_0.METHOD, "GMRES")
* Will use the GMRES solver instead of the default MAT_CR solver.
* Same applies to all Solver parameters.
* Please use with care. If you change your netlist (e.g. using
frontiers) the allocation of nets to solvers and the number of solvers
may change. Thus this type of tweaking should only be used after the
netlist completely works.
|
| | |
|
| |
|
|
|
|
|
| |
* The GMRES solver for larger matrices (>>100) can outperform Gaussian
elimination. Including it so it gets wider attention.
* As far as I know netlist is the only SPICE-like circuit simulation
providing a GMRES solver.
|
| | |
|
| | |
|
| |
|
|
|
|
| |
* This clean-up exercise will hopefully make it easier to navigate the
core code. Another long term goal is to further straighten the object
model.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* All solver scheduling is now handled by nld_solver.
* Previously, for dynamic timestepping the sub solvers would be
responsible for their scheduling themselves.
- This prevented any attemps to use parallel execution of solvers.
* Now the route is free towards experiments to use parallel execution of
solvers.
* Uses ptimed_queue_linear in solver scheduling
* Improved netlist queue implementation (template now)
* Added const delegates.
* Added subsolver stats
* Removed dead code.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
| |
* Removed update and NETLIB_UPDATE completely.
* Startup initialization order may change.
* This may cause regression tests to trigger. This is expected. Logic
TTL devices do not have a defined power-up state. That's why reset
circuits exist which create a reset signal *after* all power supply
lines are stable.
|
| |
|
|
|
| |
* This is a long term transition goal. Documented in source (see
NL_USE_BACKWARD_EULER).
|
| |
|
| |
* all inputs now explicitly specify a signal handler.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* No changes for well-behaved netlists.
* Netlists reporting "newton loops exceeded" messages should now report
a lot less of these messages.
* In case newton-raphson does not converge, netlist will reset the state
back to the beginning of the timestep. It will than use a number of very
small timesteps to hopefully deal with the non-convergence due to a too
big timesteps. Afterwards solving continues using dynamic timestepping
until the time-slice is complete.
* This is a lot better than the previous approach.
* This is meant as a performance optimization helper. Ideally, your
netlists never produce "newton loops exceeded" messages.
* Any events for connected analog and digital inputs will happen after
the timeslice is complete. Thus this approach - as the previous one -
are not suited for high-frequency (such as video outputs) situations. In
these case, only dynamic timestepping or reducing the fixed timestep are
a suitable solution.
|
| |
|
|
|
|
|
|
| |
* Electronic circuits and base components like resistors or capacitors
do not have a reset line. You can use them to create reset circuits.
There is thus no point to support soft reset, the equivalent to pressing
the reset button.
* Fixed some bugs around reset and start up logic.
* This also fixes the "scramble F3" crash.
|
| |
|
|
|
|
|
|
|
|
|
| |
* palloc.h/pmatrix2d.h: Fix static_assert warnings at the origin.
* Rework hints to broaden their use and fix NC hint.
* 74377: use NC hint
* plists.h: Fix debugging in MSVC
* Include cleanup: Move everything not needed by netlists from
nl_setup.h into core/setup.h
* Fix some clang tidy warnings
* srcclean
|
| |
|
|
|
|
|
|
| |
The purpose of this ongoing exercise is to remove unnecessary
dependencies in header files. netlist implementations should only have
access to what they need. The same applies to device implementations.
Core stuff will be moved to the core subdirectory going forward.
|