Commit Graph

9 Commits

Author SHA1 Message Date
Matt Pharr
c14c3ceba6 Provide both signed and unsigned int variants of bitcode-based builtins.
When creating function Symbols for functions that were defined in LLVM bitcode for the standard library, if any of the function parameters are integer types, create two ispc-side Symbols: one where the integer types are all signed and the other where they are all unsigned.  This allows us to provide, for example, both store_to_int16(reference int a[], uniform int offset, int val) as well as store_to_int16(reference unsigned int a[], uniform int offset, unsigned int val). functions.

Added some additional tests to exercise the new variants of these.

Also fixed some cases where the __{load,store}_int{8,16} builtins would read from/write to memory even if the mask was all off (which could cause crashes in some cases.)
2011-07-04 12:10:26 +01:00
Matt Pharr
eb22fa6173 Generalize FunctionEmitContext::PtrToIntInst and IntToPtrInst to
do the right thing if given a varying lvalue (i.e. an array of
pointers).  Fixes issue #34.
2011-06-29 12:38:12 +01:00
Matt Pharr
6b153566f3 Simplify a bunch of code by using CollectionType to collect struct
codepaths in with array/vector codepaths. (Issue #37).
2011-06-29 07:59:43 +01:00
Matt Pharr
214fb3197a Initial plumbing to add CollectionType base-class as common ancestor
to StructTypes, ArrayTypes, and VectorTypes.  Issue #37.
2011-06-29 07:42:09 +01:00
Matt Pharr
ce7978ae74 Align stack-allocated arrays of uniform types to the target vector alignment (they will often be accessed in programCount-sized chunks and this should make that a bit more efficient in the common case). Fixes issue #15 2011-06-28 20:42:18 -07:00
Matt Pharr
865e430b56 Finished updating alignment issues for vector types; don't assume pointers
are aligned to the natural vector width.
2011-06-23 18:51:15 -07:00
Matt Pharr
b84167dddd Fixed a number of issues related to memory alignment; a number of places
were expecting vector-width-aligned pointers where in point of fact,
there's no guarantee that they would have been in general.

Removed the aligned memory allocation routines from some of the examples;
they're no longer needed.

No perf. difference on Core2/Core i5 CPUs; older CPUs may see some
regressions.

Still need to update the documentation for this change and finish reviewing
alignment issues in Load/Store instructions generated by .cpp files.
2011-06-23 18:18:33 -07:00
Pete Couperus
af435e52c1 Minor mods to build on Fedora 15, LLVM 2.8 2011-06-21 22:57:36 -07:00
Matt Pharr
18af5226ba Initial commit. 2011-06-21 12:48:50 -07:00