Small fixes for CUDA code emit (#1564)

* Small fixes for CUDA code emit

* Add a CUDA translation to `GroupMemoryBarrierWithWaveSync()`. We map this to `__syncwarp()` for CUDA (with no mask, implying a full-warp sync).

* Consistently use `SLANG_PRELUDE_ASSERT` for assertions introduced in code emit (rather than just using the bare `assert(...)` function, which is not included by our CUDA prelude by default)

* Add a new `SLANG_CUDA_STRUCTURED_BUFFER_NO_COUNT` flag to the CUDA prelude that allows the `count` field to be omitted from `(RW)StructuredBuffer<T>`. This is a bit of a hacky because the computed layouts will still assume the `count` field is present, but this feature is required by at least one client application for now. A better long-term fix will take more time to design and implement.

* fixup: CUDA prelude code fix for pedantic compilers

Co-authored-by: Tim Foley <tim.foley.is@gmail.com>
Co-authored-by: Yong He <yonghe@outlook.com>

1 files changed, 2 insertions, 2 deletions

diff --git a/source/slang/slang-emit-cpp.cpp b/source/slang/slang-emit-cpp.cpp
index 5359740b4..1e2482bb5 100644
--- a/source/slang/slang-emit-cpp.cpp
+++ b/source/slang/slang-emit-cpp.cpp
@@ -984,7 +984,7 @@ void CPPSourceEmitter::_emitGetAtDefinition(const UnownedStringSlice& funcName,
         {
             int vecSize = int(getIntVal(vectorType->getElementCount()));
 
-            writer->emit("assert(b >= 0 && b < ");
+            writer->emit("SLANG_PRELUDE_ASSERT(b >= 0 && b < ");
             writer->emit(vecSize);
             writer->emit(");\n");
             if (lValue)
@@ -997,7 +997,7 @@ void CPPSourceEmitter::_emitGetAtDefinition(const UnownedStringSlice& funcName,
             //int colCount = int(getIntVal(matrixType->getColumnCount()));
             int rowCount = int(getIntVal(matrixType->getRowCount()));
 
-            writer->emit("assert(b >= 0 && b < ");
+            writer->emit("SLANG_PRELUDE_ASSERT(b >= 0 && b < ");
             writer->emit(rowCount);
             writer->emit(");\n");