[rocm-libraries] ROCm/rocm-libraries#4302 (commit e62bd8a)

[CK_TILE] add tf32 support
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

## Proposed changes

TF32 is added in CK on gfx942 and gfx950. This PR is to initiate tf32 in
CK_TILE on gfx942 and gfx950.

## Checklist

Please put an into the boxes that apply. You can also fill these out
after creating the PR. If you're not sure, please don't hesitate to ask.

- [ ] I have added tests relevant to the introduced functionality, and
the unit tests are passing locally
- [ ] I have added the test to REGRESSION_TESTS list defined at the top
of CMakeLists.txt in tests/CMakeLists.txt, **IF** the test takes more
than 30 seconds to run.
- [ ] I have added inline documentation which enables the maintainers
with understanding the motivation
- [ ] I have removed the stale documentation which is no longer relevant
after this pull request
- [ ] (If this change is user-facing) I have added release notes which
provide the end users with a brief summary of the improvement from this
pull request
- [x] I have run  on all changed files
- [ ] Any dependent changes have been merged

## Discussion

include/ck_tile/ops/fmha/pipeline/block_fmha_fwd_v3_pipeline.hpp

@@ -229,15 +229,6 @@ CK_TILE_DEVICE fp16x2_t cvt_pk_fp16_f32(float a, float b)
     return result;
 }
 
-CK_TILE_DEVICE bf16x2_t cvt_pk_bf16_f32(float a, float b)
-{
-    bf16x2_t result;
-    asm volatile("v_cvt_pk_bf16_f32 %[result], %[a], %[b]"
-                 : [result] "=v"(result)
-                 : [a] "v"(a), [b] "v"(b));
-    return result;
-}
-
 CK_TILE_DEVICE fp32x2_t pk_mul_f32(fp32x2_t lhs, fp32x2_t rhs)
 {
     fp32x2_t result;
@@ -856,7 +847,7 @@ struct BlockFmhaFwdV3Pipeline
                 }
                 else
                 {
-                    auto casted                           = detail::cvt_pk_bf16_f32(x, y);
+                    auto casted                           = ck_tile::cvt_pk_bf16_f32(x, y);
                     sp(sp_reg_idx).p.thread_buf_[idx]     = casted.x;
                     sp(sp_reg_idx).p.thread_buf_[idx + 1] = casted.y;
                 }