[llvm] [SPIR-V] Access-chain aggregate pointers to element 0 for cooperative matrix load/store (PR #202050)

[Julian Klappenbach via llvm-commits]

================
@@ -0,0 +1,79 @@
+; cajeta-gpu cooperative-matrix increment CM3: a spirv-val-clean GLCompute kernel
+; that runs a cooperative-matrix matmul tile against DESCRIPTOR-BOUND storage
+; buffers under the Vulkan flavor.
----------------
jklappenbach wrote:

Sorry about that.  I'm creating a language that's utilizing this API, and that's what leaked through.  You have a good question above that I'm looking into.  So, this particular PR might be moot.

https://github.com/llvm/llvm-project/pull/202050