Upcoming hardware (gfx12 and some future gfx9) will support the OCP
8-bit float formats for their matrix multiplication intrinsics and
conversion operations, retaining existing opcodes and compiler builtins.

This commit adds support for these types to the MLIR wrappers around
such operations, ensuring that the OCP types aren't used to generate
those builtins on hardware that doesn't expect that format and,
conversely, to ensure that the pre-OCP formats aren't used on new
hardware.

those operations were not being converted to the LLVM intrinsics they
correspond to because the rewrite patterns were still checking for
gfx940+.

As part of this, factor out tests for type-match isto isNativeFp8()
and isNativeBf8() functions in the AMDGPUToRocdl rewrites.

Also, fix a typo in isGfx940() that caused it to be true for gfx950.

Finally, test all these OCP format conversions by duplicating the
gfx940 tests.