In particular, this gives us desired behavior for NaNs (all compares
involving a NaN evaluate to true). This in turn allows writing the
canonical isnan() function as "v != v".
Added isnan() to the standard library as well.
Previously, we were trying to take a uniform seed and then shuffle that
around to initialize the state for each of the program instances. This
was becoming increasingly untenable and brittle.
Now a varying seed is expected and used.
These tests all fail with generic-16/c++ output currently; however, the
output indicates that it's just small floating-point differences.
(Though the question remains, why are those differences popping up?)
Now a declaration like 'struct Foo;' can be used to establish the
name of a struct type, without providing a definition. One can
pass pointers to such types around the system, but can't do much
else with them (as in C/C++).
Issue #125.
The decl.* code now no longer interacts with Symbols, but just returns
names, types, initializer expressions, etc., as needed. This makes the
code a bit more understandable.
Fixes issues #171 and #130.
Once we're down to something that's not another nested expr list, use
TypeConvertExpr() to convert the expression to the type we need. This should
allow simplifying a number of the GetConstant() implementations, to remove
partial reimplementation of type conversion there.
For now, this change finishes off issue #220.
Previously, the compiler would crash if e.g. the program passed a
temporary value to a function taking a const reference. This change
fixes ReferenceExpr::GetValue() to handle this case and allocate
temporary storage for the temporary so that the pointer to that
storage can be used for the reference value.
Closer compatibility with C++: given a non-reference type, treat matching
to a non-const reference of that type as a better match than a const
reference of that type (rather than both being equal cost).
Issue #224.
Implicit conversion to function types is now a more standard part of
the type conversion infrastructure, rather than special cases of things
like FunctionSymbolExpr immediately returning a pointer type, etc.
Improved AddressOfExpr::TypeCheck() to actually issue errors in cases
where it's illegal to take the address of an expression.
Added AddressOfExpr::GetConstant() implementation that handles taking
the address of functions.
Issue #223.
We were incorrectly trying to type convert the varying offset to a
uniform value, which in turn led to an incorrect compile-time error.
Fixes issue #201.
In particular, we 1. weren't setting the function mask to 'all on', such that
any mixed function mask would in turn apply inside the foreach loop, and 2.
weren't always setting the internal mask to 'all on' before doing any additional
masking based on the iteration variables.
In particular, LLVMVectorIsLinear() and LLVMVectorValuesAllEqual() are able
to reason a bit about the effects of the shifts and the ANDs that are
generated from SOA indexing calculations, so that they can detect more cases
where a linear sequence of locations are in fact being accessed in
the presence of SOA data.
Now, the pointed-to type is always uniform by default (if an explicit
rate qualifier isn't provided). This rule is easier to remember and
seems to work well in more cases than the previous rule from 6d7ff7eba2.
Now, if a struct member has an explicit 'uniform' or 'varying'
qualifier, then that member has that variability, regardless of
the variability of the struct's variability. Members without
'uniform' or 'varying' have unbound variability, and in turn
inherit the variability of the struct.
As a result of this, now structs can properly be 'varying' by default,
just like all the other types, while still having sensible semantics.
Now, if rate qualifiers aren't used to specify otherwise, varying
pointers point to uniform types by default. As before, uniform
pointers point to varying types by default.
float *foo; // varying pointer to uniform float
float * uniform foo; // uniform pointer to varying float
These defaults seem to require the least amount of explicit
uniform/varying qualifiers for most common cases, though TBD if it
would be easier to have a single rule that e.g. the pointed-to type
is always uniform by default.
There are two related optimizations that happen now. (These
currently only apply for gathers where the mask is known to be
all on, and to gathers that are accessing 32-bit sized elements,
but both of these may be generalized in the future.)
First, for any single gather, we are now more flexible in mapping it
to individual memory operations. Previously, we would only either map
it to a general gather (one scalar load per SIMD lane), or an
unaligned vector load (if the program instances could be determined
to be accessing a sequential set of locations in memory.)
Now, we are able to break gathers into scalar, 2-wide (i.e. 64-bit),
4-wide, or 8-wide loads. Further, we now generate code that shuffles
these loads around. Doing fewer, larger loads in this manner, when
possible, can be more efficient.
Second, we can coalesce memory accesses across multiple gathers. If
we have a series of gathers without any memory writes in the middle,
then we try to analyze their reads collectively and choose an efficient
set of loads for them. Not only does this help if different gathers
reuse values from the same location in memory, but it's specifically
helpful when data with AOS layout is being accessed; in this case,
we're often able to generate wide vector loads and appropriate shuffles
automatically.
? : now short-circuits evaluation of the expressions following
the boolean test for varying test types. (It already did this
for uniform tests).
Issue #169.
We now follow C's approach of evaluating these: we don't evaluate
the second expression in the operator if the value of the first one
determines the overall result. Thus, these can now be used
idiomatically like (index < limit && array[index] > 0) and such.
For varying expressions, the mask is set appropriately when evaluating
the second expression.
(For expressions that can be determined to be both simple and safe to
evaluate with the mask all off, we still evaluate both sides and compute
the logical op result directly, which saves a number of branches and tests.
However, the effect of this should never be visible to the programmer.)
Issue #4.
We now do a single atomic hardware swap and then effectively do
swaps between the running program instances such that the result
is the same as if they had happened to run a particular ordering
of hardware swaps themselves.
Also cleaned up __atomic_swap_uniform_* built-in implementations
to not take the mask, which they weren't using anyway.
Finishes Issue #56.
Previously, when we had a switch statement with a uniform switch condition
but a 'break' statement that was under varying control flow inside the
switch, we'd promote the switch condition to be varying so that the
break would work correctly.
Now, we leave the condition as uniform and are thus able to use the
more-efficient LLVM switch instruction in this case.
Issue #156.
