This patch adds non-blocking event control for array words.
It also fixes a problem where the word used to put the
calculated delay for a non-blocking array assignment was
not being released. It also fixes the non-blocking array
assignments to correctly handle off the end/beginning part
selects.
Since some event control assignments can be skipped we need an
event control clear so that future %evctl statements do not fail
their assert. This patch adds %evctl/c and uses it in the compiler
as appropriate to keep the event control information in sync.
This patch adds full event control for vectors and parts of a
vector. It also fixes the other non-blocking part select code
to correctly handle a negative offset ([1:-2] of a [4:0] will
have an offset of -2).
The previous patch (commit 8b0ca902a6)
dealing with the possibilities of (unsigned long) and (vvp_time64_t)
being either the same or different managed to redefine UL_AND_TIME64_DIFF
in the 64-bit case. This does, of course, trigger a compiler warning.
That warning is repeated on every .cc file with a #include "config.h",
which is to say, just about every file.
This patch inverts the sense of the preprocessor conditional, calling
it UL_AND_TIME64_SAME. No more warnings!
A variable that is used to set the delay of a .delay statement
must be scaled to match the local units and for real values
rounded using the precision. This value is then converted to
the simulation precision.
This patch pushes the non-blocking event control information to
the code generator. It adds the %evctl statements that are used
to put the event control information into the special thread
event control registers. The signed version (%evctl/s) required
the implementation of %ix/getv/s to load a signed value into
an index register. It then adds %assign/wr/e event control based
non-blocking assignment for real values. It also fixes the other
non-blocking real assignments to use Transport instead of inertial
delays.
Nothing to do with tab width! Eliminates useless
trailing spaces and tabs, and nearly all <space><tab>
pairings. No change to derived files (e.g., .vvp),
non-master files (e.g., lxt2_write.c) or the new tgt-vhdl
directory.
Low priority, simple entropy reduction. Please apply
unless it deletes some steganographic content you want
to keep.
The right shift of vvp_vector2_t needs to
account for and mask off shifted bits. Otherwise
there will be unexpected results after
a vvp_vector2_t::trim method.
Assume that anything that is strength aware already handles a
recv_vec8_pv and make the default function convert the bits
to a vec4 and then call recv_vec4_pv with this new value.
This patch adds support for vpiScaledRealTime and vpiSuppressTime
to VPI callbacks. It also fixes a bug where the callback data
object was not being copied correctly and adds support for
vpiSuppressVal. This requires adding vpiSuppressVal to a few
other routines. It adds the ability for a callback to return
more than vpiScalarVal (all values supported by the
vpip_vec4_get_value() procedure). It also fixes a bug where
vpip_vec4_get_value() would incorrectly return vpiZ for a BIT4_X
scalar value. It also comments the potentially dangerous
vpiScalarVal calculation in the vvp_fun_signal::get_value()
procedure.
This patch causes a thread that is created to evaluate a function
to be executed immediately. The parent thread is resumed when the
function thread terminates.
When multiply is done in native words, the conversion to words from the
vp_vector4_t vectors must be done signed. This only matters if the
input operands are different sizes (and themselves signed) but will
not hurt even if we want an unsigned result.
Logical (in)equality needs to look at all the bits of both operands,
and cannot short circuit the test unless defined bits differ. If there
are undefined bits, the equality is undefined at that point, but return
x only if there are not other bits that make the results clearly
unequal.
The double to vvp_vector4_t constructor was not using the correct
declaration for the bit words. This worked as long as unsigned and
unsigned long were the same size (usually).
The new real to int conversion was incorrectly setting the
bits for minus infinity to all ones. This is incorrect in a
two's complement encoding where the largest negative number
would be a leading 1 followed by an infinite number of zeros.
This patch updates the %cvt/vr command to use the new double to vector
constructor. This allows the resulting bit pattern to be larger than
a long. The old method was producing incorrect results without a
warning for large bit values.
This patch makes .part/pv strength aware, resolv vec8_pv
aware. vvp_net_fun_t adds vec8_pv as a virtual function
with an appropriate error default. vvp_fun_signal should
full support vec8_pv (not tested and may not be needed).
This patch adds .cast/int and updates .cast/real to act as a local
(temporary) net and to support either a signed or unsigned input.
The vvp_vector4_t class not can convert an arbitrarily sized double
to a vector value. This removes the restriction of lround().
Also document the new statements.
This handles the general case of a non-real operand to a real-valued
division. This can turn up if only 1 operand of a divide is real. In
this case the division as a whole is real and the other operand must
be cast to real.
This method creates an extra node, but it should be a very compact
node and this node does no evaluation tricks so in the run time should
be no more expensive then folding the cast into the .arith/div.r itself.
This patch changes the base oct to string converter to correctly
pad the top digit. x or xx should display as a single lower case
x when they are located in the top bits. Before these were being
interpreted as 00x or 0xx and displayed X. Also modified the hex
conversion to use this same scheme instead of a loop.
First, handle the trivial (but possibly common) resolution cases in
inlined code, and only call the complete function for the complicated
cases. Then clean up the complex function for readability, and account
for the constraints that the front-end function established.
Arrays of vvp_vector4_t values redundantly store some fields in every
word. Create a special type that stores vvp_vector4_t values in a form
that does not duplicate the width of all the items. This can save a lot
of space when big memories are simulated.
The schedule_assign_plucked_vector is a better way to implement the
schedule_assign_vector, or at least no worse, so remove the now
redundent schedule_assign_vector.
It is legal (though worthy of a warning, I think) for the part select
of an l-value to me out of bounds, so replace the error message with
a warning, and generate the appropriate code. In the process, clean
up some of the code for signal l-values to divide out the various kinds
of processing that can be done. This cleans things up a bit.
The load-and-add for vectors %load/vp0/s can be combined with the
load-and-add for array words, and the %load/avp0/s added to round
out the combinations. This can make for fewer instructions when
words are padded in arithmetic expressions.
The functor_ref_lookup() function fills its argument in with the
vvp_net_t* pointer that matches the var name, so there is no need
to create the vvp_net_t object before then.
The vpiIndex is really just a different view into the same object,
so implement the trickery needed to support a vpiIndex with the
absolute minimum memory cost.
The %load/vp0 instruction adds a signed value to the signal value being
loaded, but it doesn't allow for a signed source vector. Add the
%load/vp0/s instruction that pads the loaded vector, and add the code
generator details to properly use it.
The vvp_net_fun_t objects, and derived objects, are small, and are
created in large quantities. Tightly pack them into permanently
allocated space in order to save on system allocation overhead, and
thus save overall on memory.
The part select of a vector is converted by the compiler during
elaboration to a 0-based canonical address. But since it is legal
to address bits below the LSB, the canonical address can be negative.
So make the part select base for selecting from signals work with
signed arithmetic and make the code generator generate negative
indices when needed.
Scheduler cells are small objects that come and go in great quantities.
Even though they are allocated and deallocated a lot, they tend to a
steady state quantity, so put together a heap that is unique for each
cell type.
This heap actually saves memory overall because cells are allocated in
chunks, thus eliminating allocator overhead, and they are pulled/pushed
from/to a heap very quickly so that what overhead remains is slight and
bounded.
The vvp_net_t objects are never deleted, so overload the new operator
to do a more space efficient permanent allocation.
The %assign/v instruction copied the vvp_vector4_t object needlessly
on its way to the scheduler. Eliminate that duplication.(cherry picked from commit d0f303463d)
The vvp_vector8_t constructor and destructor involve memory allocation
so it is best to pass these objects by reference as much as possible.
Also have the islands take more care not to perform resolution if the
inputs aren't really different.
NOTE: This is a port of commit 2f4e5bf5b6
from the "performance" branch, without the resolver scheduling changes.
This was causing test suite variances with pr1820472.v. It looks like
there might be a race in that program anyhow, but for now leave out the
resolver scheduling changes so that the rest of this commit can go in.
The __vpiPV objects express themselves as vpiPartSelect objects.
Add support for value change callbacks by attaching the callback
to the signal that we part select from.
The %load/v instruction was doing some spurious resizes of the vector
that comes from the signal. Eliminate those resizes that can be
removed, and optimize some that remain.
This is not a solution to all the problems, but is a better catch-all
then what is currently there. Allow the index field to be a T<> that
accesses the thread to get the address index.
Note that the lexor.lex currently returns the T<> as a T_SYMBOL, and the
users of T_SYMBOL objects need to interpret the meaning. This is
probably not the best idea, in light of all the other *<> formats that
now exist.
If a memory word was accessed before it was defined the
code was returning a zero width vector result. Now it
returns an appropriately sized vector of 'x'.
The memory opcodes %assign/mv, %load/mv and %set/mv
were removed by a previous patch. This one removes
the documentation from opcodes.txt. It also removes
the documentation for the .mem* statements for the
same reason.
Recursive branch resolution was scanning every branch end, even though
many branch ends share ports and need not be repeatedly scanned. Handle
marks and flags to cut off recursion where it is not needed so as to
save much run time.
Conflicts:
tgt-vvp/vvp_scope.c
Note that the draw_net_input.c takes in a lot of the codes that used
to be in vvp_scope.c, so some changes may have been lost.
This patch cleans up the dump routines and adds file and
line number information for errors. It also adds some of
the missing MemoryWord properties so they can now be
dumped and monitored correctly.
This patch adds $simparam and $simparam$str from Verilog-A.
The analog simulator parameters return 0.0 or N/A. The
vvp_cpu_wordsize system function has been moved into the
$simparam call and is now named CPUWordSize.
This patch also starts the factoring of common code in the
vpi directory. Some routines were renamed.
The priv.c file was renamed to sys_priv.c to match the
include file.
System functions can now have strings put to their output.
Fold the bi-directional part select into the pass switch (tran) support
so that it can be really bi-directional. This involves adding a new
tranvp device that does part select in tran islands, and reworking the
tran island resolution to handle non-identical nodes. This will be needed
for resistive tran devices anyhow.
The vvp_island classes are added, as well as support for tranif nodes
that use this concept. The result is a working implementation for
tranif0 and tranif1.
In the process, the symbol table functions were cleaned up and made
into templates for better type safety, and the vvp_net_ptr_t was
generalized so that it can be used by the branches in the island
implementation.
Also fix up the array handling to use the better symbol table support,
and to remember to clear its own table when linking is done.
The multiply runs does not need to do all the combinations of digit
products, because the higher ones cannot add into the result. Fix the
iteration to limit the scan.
Clarify that operands are typically 32bits, and have the code generator
make better use of this.
Also improve the %movi implementation to work well with marger vectors.
Add the %andi instruction to use immediate operands.
Use high radix long division to take advantage of the divide hardware
of the host computer. It looks brute force at first glance, but since
it is using the optimized arithmetic of the host processor, it is much
faster then implementing "fast" algorithms the hard way.
This instruction adds an integer value to the value being loaded. This
optimization uses subarrays instead of the += operator. This is faster
because the value is best loaded into the vector as a subarray anyhow.
The AND and OR operators for vvp_bit4_t are slightly tweaked to be
lighter and inlinable.
The vvp_vector4_t::set_bit is optimized to do less silly mask fiddling.
When processing wide vectors of these operations, it pays to process
them as vectors. This improves run-time performance. Have the run time
select vectorized or not based on the vector width.
Improve vvp_vector4_t methods copy_bits and the part selecting constructor
to make better use of vector words. Eliminate bit-by-bit processing by
these methods to take advantage of host processor words.
Improve vthread_bits_to_vector to use these improved methods and Update
the %load/av and %set/v instructions to take advantage of these changes.
The MinGW system() implementation appears to return the straight
return value instead of the waitpid() like result that more
normal systems return. Because of this just return the system()
result without processing for MinGW compilations.
Older version of the MinGW runtime (pre 3.14) just used the
underlying vsnprintf(). Which has some problems. The 3.14 version
has some nice improvements, but it has a sever bug when processing
"%*.*f", -1, -1, <some_real_value>. Because of this we need to use
the underlying version without the enhancements for now.
snprintf prints %p differently than the other printf routines
so use _snprintf to get consistent results.
Only build the PDF files if both man and ps2pdf exist.
MinGW does not know about the z modifier for %d, %u, etc.
Add some missing Makefile check targets.
These instructions can take advantage of the much optimized
vector_to_array function to do their arithmetic work quickly and
punt on X very quickly if needed. This helps some benchmarks.
Functions like $monitor need to attach callbacks to array words if
those words are to be monitored. Have the array hold all the callbacks
for words in the array, under the assumption that the monitored words
are sparse.
Array words don't have a vpiHandle with a label, so the %vpi_call
needs a special syntac for arguments that reference array words.
This syntax creates an array word reference that persists and can
be used at a VPI object by system tasks.
It is possible for the code generator to create .array/port objects
before the .array object that the port refereces, so use the resolv_list
to arrange for binding during cleanup.
Save tons of space per memory word by not creating a vpi handle for
each and every word of a variable array. (Net arrays still get a
vpiHandle for every word.) The consequence of this is that all
accesses to a variable array need to go through the indexing.
This commit handles the most common places where this impacts, but
there are still problems.
vpi_handle_by_name() was assuming it was always given a valid scope
object. In the context of vpi_chk_error() this is not required and
some users use/abuse the interface by calling the function with invalid
objects expecting a 0 return value. This patch adds an explicit check
for the supported types vpiScope and as an extension vpiModule.
Anything else should be flagged as an error once we have vpi_chk_error()
implemented, but for now it just returns 0.
The abs() function needs to be able to turn -0.0 into 0.0. This proved
to be too clunky (and perhaps impossible) to do with tests and jumps,
so add an %abs/wr opcode to do it using fabs().
The min/max functions need to take special care with the handling
of NaN operands. These matter, so generate the extra code to handle
them.
This patch adds file and line information for parameters and
local parameters. It also adds file/line stubs for signals in
the tgt-* files. It adds the pform code needed to eventually
do genvar checks and passing of genvar file/line information.
It verifies that a genvar does not have the same name as a
parameter/local parameter.
This patch makes sure that objects either support vpiFile
and vpiLineNo or adds dummy code so that a runtime error
will not occur when accessing these properties. It also
returns 1 for the size of real variables and adds a
simplified vpiIndex that matches the Memory interface.
This patch adds code to push the file and line information
for scope objects (modules, functions, tasks, etc.) to the
runtime. For modules, this includes the definition fields.
This patch adds ifnone functionality. It does not produce an
error when both an ifnone and an unconditional simple module
path are given. For this case the ifnone delays are ignored.
Fix handling of cases where multiple specified delays are activated
for a given output. Need to apply the standard selection criteria
that gets the minimum value.
The %sub instruction didn't have the efficent implementation that
the %add instructions used. Update subtraction to use the array
method, so that it gets the same performance benefits.
The vvp_vector4_t often receives the results of vector arithmetic.
Add an optimized method for setting that data into the vector. Take
into account that arithmetic results have no X/Z bits, etc.
On some systems, 1UL<<X will make a mess if X is the size of
an unsigned long. This especially seems to be a problem on i386
systems. Protect those shifts in the vvp_net.cc.
By slightly altering the vvp_bit4_t encoding, a few simple
optimizations become possible. By making Z==2 and X==3, the
conversion from X/Z to X is a simple shift-or, and this can
be used to reduce the size of some of the bit4 operators.
In previous incarnations of the vvp runtime, bit vectors were passed
around as arrays of unsigned char that charried bit4 vectors. That
is no longer used. Remove the last vestiges of that dead code.
Remove dependencies on vvp_bit4_encoding outside of the vvp_net
core types. The table_functor_s class was the worst offfender and
was barely used, so it is now removed completely. There are a few
opcodes in vhtread.cc that also make vvvp_bit4_t encoding
assumptions (and used casts) and those have been fixed. There
were also various VPI interface functions that are fixed.
The vvp_vector4_t holds 4-value logic. This patch changes the encoding
of 4-value bits in the vector to use separate A- and B bit vectors,
with the B- vector signaling the A- bits that are not 0/1. This
allows rapid conversion to 2-value logic, and rapid tests for X
and Z values.
This patch fixes deassign to allow it to unlink from a driver.
It also zeros the cassign_link and force_link pointers after
they have been unlinked. Not doing this will cause an assert
if deassign/release are called multiple times (variable only).
This patch adds the ability to assign/deassign a bit or part select.
It also cleans up the code and fixes some problem in the forcing of
strength aware nets.
Rework the MUXZ and MUXR code to use an enum instead of plain
integers for the select input state. This makes it more obvious
what is actually going on.
This patch adds a flag to the MUXZ object to make sure that it will
run at time zero if needed. If this is not done the default Z result
may not be overridden by an X result.
This patch adds a %assign/av/d opcode. This is a version of %assign/av
that allows a delay expression. Ultimately this allows a dynamically
indexed array to have a delay expression (non-constant delay value).
This patch add (I believe back in) the < and > characters to the
general T_SYMBOL regular expression. The match was also moved after
the special T_SYMBOL cases so that flex would match them correctly.
Flex matching rules are longest or if a tie the first.
vvp did not have the ability to handle real parameters.
This patch fixes that omission. Parameters are only used
by vpi calls to get compile time information.
Threads used to be deleted when they finished processing code.
The problem with this is that some of the code could be
rescheduled to run at rosync ($strobe, etc.). This allowed the
thread data the code depended on to be reaped too soon. This
patch uses a new queue to schedule thread deletion. The queue
is processed after rosync has finished.
This patch makes the muxz and muxr functors schedule events
instead of directly calling vvp_send_*(). This prevents the
code from going into an infinite loop when the output feeds
back to the select.
This patch adds functionality to do a bit or part select release
when a constant value is forced to the net/register. It also adds an
error message when the user tries to force a signal to a bit/part
select. This is not currently handled by the run time, so is now
caught in the compiler (tgt-vvp). Where when this functionality is
needed, it will be easy to know what to do instead of trying to track
down some odd runtime functionality.
What this all means is that you can force a signal to an entire
signal or you can force a constant to any part of a signal (bit,
part or entire) and release any of the above. Technically the
release of a constant value does not have to match the force.
The runtime verifies that if you are releasing a signal driver
it is being done as a full release. I don't see an easy way to
check this in the compiler.
To fix the signal deficiencies we need to rework the force_link
code to allow multiple drivers and partial unlinking. Much of
this is in the runtime, but the %force/link operator may also
need to be changed like I did to the %release opcode.
vpi_put_value can mimic force and release with vpiForceFlag and
vpiReleaseFlag flags to the vpi_put_value call. With this patch,
the infrastructure is added to allow the flags argument to be passed
to the dispatched put_value function, and for signals handle those
flags as force/release of a net.
This patch reworks much of the ternary code to short circuit when
possible and supports real values better. It adds a blend operator
for real values that returns 0.0 when the values differ and the value
when they match. This deviates slightly from the standard which
specifies that the value for reals is always 0.0 when the conditional
is 'bx. There are also a couple bug fixes.
These fixes have not been ported to continuous assignments yet.
Ternary operators used at compile time and in procedural assignments
should be complete (short circuit and support real values).
draw_number_bool64() in tgt-vvp/eval_bool.c was using %ix/load to
load immediate values into registers greater than three. The problem
was that of_IX_LOAD() in vvp/vthread.cc was masking off the upper
bits. This was putting the results in the wrong register. This patch
removes the bit masking from of_IX_LOAD() and updates the %ix/load
documentation.
This patch fixes two problems. The first is that thr_check_addr()
was being used inconsistently. It should be passed a real address,
but the resize of the vector should be at least one more than this
address. The extra and unneeded CPU_WORD_BITS was also removed
from the routine.
The second problem involved an invalid memory access in
vvp_vector4_t::set_vec() when the vector being copied was an integer
multiple of the machine word width. Under this condition there would
be no remaining bits that needed to be copied but the routine was always
trying to copying some remaining bits. This code is now only executed
when there is a remainder.
Neither of these appear to be causing runtime problems. The second one
was found with valgrind. The first were found while tracking down the
second problem.
This patch adds the power operator for signed bit based values
in a continuous assignment. It also fixes a few other power
expression width problems. The expression width is still not
calculated correctly, since the correct method can produce huge
possible bit widths. The result is currently limited to the width
of the native long. This is because lround() is used to convert
from a double to an integer. A check in the code generator protects
the runtime from this limitation.
This patch adds the power operator for unsigned bit based values
in a continuous assignment. It also refactors the power code for
normal expressions and continuous assignments.
This patch adds bit based power support to normal expressions.
It also pushes the constant unsigned bit based calculation to
the runtime until the bit based method can be copied to the
compiler. Continuous assignments also need to use this type
of calculation.
vvp/main.cc was including asm/page.h on GNU/Linux
systems, though that file does not often exist and
is not necessary.
Signed-off-by: Michael Witten <mfwitten@mit.edu>
This patch makes sure the delay is calculated correctly when only some
of the bits change and you are comparing against the initial (t0) value.
Basically you have to check the initial value against all the bits in
the new signal not just the first bit since the order that bits change
is not deterministic.
Allow user defined functions to take real value arguments and return
real value results in net contexts. Use the data type of the nets
attached to the ports to define the data types of the arguments and
return value.
When conditional delays are in use, it is sometimes possible for there
to be no delays available for a given input event. In that case, skip
the delay processing for that case instead of crashing.
Rework the encoding of a real value in the Cr<> label to be similar to
the format used by the %loadi/wr instruction. This mantissa-expoment
format better carries all the bits of the desired real value with
plenty of fidelity and range.
Cary R
Larry Doolittle
Stephen Williams
Martin Whitaker
Holger Wächtler
Stephen Williams
J Varshney
Michael Witten