These compute the average of two given values, rounding up and down,
respectively, if the result isn't exact. When possible, these are
mapped to target-specific intrinsics (PADD[BW] on IA and VH[R]ADD[US]
on NEON.)
A subsequent commit will add pattern-matching to generate calls to
these intrinsincs when the corresponding patterns are detected in the
IR.)
On a target with a 16-bit mask (for example), we would choose the type
of an integer literal "1024" to be an int16. Previously, we used an int32,
which is a worse fit and leads to less efficient code than an int16
on a 16-bit mask target. (However, we'd still give an integer literal
1000000 the type int32, even in a 16-bit target.)
Updated the tests to still pass with 8 and 16-bit targets, given this
change.
For varying int8/16/32 types, divides by small constants can be
implemented efficiently through multiplies and shifts with integer
types of twice the bit-width; this commit adds this optimization.
(Implementation is based on Halide.)
All of these pass the current mask to FunctionEmitContext::SetBlockEntryMask()
so that when a break/continue/return is encountered, it can test to see if all
lanes have followed that path and then return; this in turn ensures that we never
run statements with an all-off execution mask.
These functions were passing the function internal mask, not the full mask, and
thus could end up executing code with the mask all off if some lanes were
disabled by an outer function. (The new tests test this case.)
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.
