[HIPIFY][rocBLAS] 64-bit functions support - Step 7

+ `rocblas_(s|d|c|z)syr_64` support + `rocblas_(c|z)her_64` support + Updated synthetic tests, the regenerated `hipify-perl`, and `BLAS` `CUDA2HIP` documentation
emankov · Oct 4, 2024 · 2eed94c · 2eed94c
1 parent 6b2f06c
commit 2eed94c
Show file tree

Hide file tree

Showing 5 changed files with 96 additions and 48 deletions.
diff --git a/bin/hipify-perl b/bin/hipify-perl
@@ -1588,7 +1588,9 @@ sub rocSubstitutions {
     subst("cublasCher2_v2", "rocblas_cher2", "library");
     subst("cublasCher2k", "rocblas_cher2k", "library");
     subst("cublasCher2k_v2", "rocblas_cher2k", "library");
+    subst("cublasCher_64", "rocblas_cher_64", "library");
     subst("cublasCher_v2", "rocblas_cher", "library");
+    subst("cublasCher_v2_64", "rocblas_cher_64", "library");
     subst("cublasCherk", "rocblas_cherk", "library");
     subst("cublasCherk_v2", "rocblas_cherk", "library");
     subst("cublasCherkx", "rocblas_cherkx", "library");
@@ -1633,7 +1635,9 @@ sub rocSubstitutions {
     subst("cublasCsyr2_v2", "rocblas_csyr2", "library");
     subst("cublasCsyr2k", "rocblas_csyr2k", "library");
     subst("cublasCsyr2k_v2", "rocblas_csyr2k", "library");
+    subst("cublasCsyr_64", "rocblas_csyr_64", "library");
     subst("cublasCsyr_v2", "rocblas_csyr", "library");
+    subst("cublasCsyr_v2_64", "rocblas_csyr_64", "library");
     subst("cublasCsyrk", "rocblas_csyrk", "library");
     subst("cublasCsyrk_v2", "rocblas_csyrk", "library");
     subst("cublasCsyrkx", "rocblas_csyrkx", "library");
@@ -1741,7 +1745,9 @@ sub rocSubstitutions {
     subst("cublasDsyr2_v2", "rocblas_dsyr2", "library");
     subst("cublasDsyr2k", "rocblas_dsyr2k", "library");
     subst("cublasDsyr2k_v2", "rocblas_dsyr2k", "library");
+    subst("cublasDsyr_64", "rocblas_dsyr_64", "library");
     subst("cublasDsyr_v2", "rocblas_dsyr", "library");
+    subst("cublasDsyr_v2_64", "rocblas_dsyr_64", "library");
     subst("cublasDsyrk", "rocblas_dsyrk", "library");
     subst("cublasDsyrk_v2", "rocblas_dsyrk", "library");
     subst("cublasDsyrkx", "rocblas_dsyrkx", "library");
@@ -1933,7 +1939,9 @@ sub rocSubstitutions {
     subst("cublasSsyr2_v2", "rocblas_ssyr2", "library");
     subst("cublasSsyr2k", "rocblas_ssyr2k", "library");
     subst("cublasSsyr2k_v2", "rocblas_ssyr2k", "library");
+    subst("cublasSsyr_64", "rocblas_ssyr_64", "library");
     subst("cublasSsyr_v2", "rocblas_ssyr", "library");
+    subst("cublasSsyr_v2_64", "rocblas_ssyr_64", "library");
     subst("cublasSsyrk", "rocblas_ssyrk", "library");
     subst("cublasSsyrk_v2", "rocblas_ssyrk", "library");
     subst("cublasSsyrkx", "rocblas_ssyrkx", "library");
@@ -2023,7 +2031,9 @@ sub rocSubstitutions {
     subst("cublasZher2_v2", "rocblas_zher2", "library");
     subst("cublasZher2k", "rocblas_zher2k", "library");
     subst("cublasZher2k_v2", "rocblas_zher2k", "library");
+    subst("cublasZher_64", "rocblas_zher_64", "library");
     subst("cublasZher_v2", "rocblas_zher", "library");
+    subst("cublasZher_v2_64", "rocblas_zher_64", "library");
     subst("cublasZherk", "rocblas_zherk", "library");
     subst("cublasZherk_v2", "rocblas_zherk", "library");
     subst("cublasZherkx", "rocblas_zherkx", "library");
@@ -2058,7 +2068,9 @@ sub rocSubstitutions {
     subst("cublasZsyr2_v2", "rocblas_zsyr2", "library");
     subst("cublasZsyr2k", "rocblas_zsyr2k", "library");
     subst("cublasZsyr2k_v2", "rocblas_zsyr2k", "library");
+    subst("cublasZsyr_64", "rocblas_zsyr_64", "library");
     subst("cublasZsyr_v2", "rocblas_zsyr", "library");
+    subst("cublasZsyr_v2_64", "rocblas_zsyr_64", "library");
     subst("cublasZsyrk", "rocblas_zsyrk", "library");
     subst("cublasZsyrk_v2", "rocblas_zsyrk", "library");
     subst("cublasZsyrkx", "rocblas_zsyrkx", "library");
@@ -12541,8 +12553,6 @@ sub warnRocOnlyUnsupportedFunctions {
         "cublasZsyrkx_64",
         "cublasZsyrk_v2_64",
         "cublasZsyrk_64",
-        "cublasZsyr_v2_64",
-        "cublasZsyr_64",
         "cublasZsyr2k_v2_64",
         "cublasZsyr2k_64",
         "cublasZsyr2_v2_64",
@@ -12559,8 +12569,6 @@ sub warnRocOnlyUnsupportedFunctions {
         "cublasZherkx_64",
         "cublasZherk_v2_64",
         "cublasZherk_64",
-        "cublasZher_v2_64",
-        "cublasZher_64",
         "cublasZher2k_v2_64",
         "cublasZher2k_64",
         "cublasZher2_v2_64",
@@ -12610,8 +12618,6 @@ sub warnRocOnlyUnsupportedFunctions {
         "cublasSsyrkx_64",
         "cublasSsyrk_v2_64",
         "cublasSsyrk_64",
-        "cublasSsyr_v2_64",
-        "cublasSsyr_64",
         "cublasSsyr2k_v2_64",
         "cublasSsyr2k_64",
         "cublasSsyr2_v2_64",
@@ -12767,8 +12773,6 @@ sub warnRocOnlyUnsupportedFunctions {
         "cublasDsyrkx_64",
         "cublasDsyrk_v2_64",
         "cublasDsyrk_64",
-        "cublasDsyr_v2_64",
-        "cublasDsyr_64",
         "cublasDsyr2k_v2_64",
         "cublasDsyr2k_64",
         "cublasDsyr2_v2_64",
@@ -12823,8 +12827,6 @@ sub warnRocOnlyUnsupportedFunctions {
         "cublasCsyrkEx",
         "cublasCsyrk3mEx_64",
         "cublasCsyrk3mEx",
-        "cublasCsyr_v2_64",
-        "cublasCsyr_64",
         "cublasCsyr2k_v2_64",
         "cublasCsyr2k_64",
         "cublasCsyr2_v2_64",
@@ -12847,8 +12849,6 @@ sub warnRocOnlyUnsupportedFunctions {
         "cublasCherkEx",
         "cublasCherk3mEx_64",
         "cublasCherk3mEx",
-        "cublasCher_v2_64",
-        "cublasCher_64",
         "cublasCher2k_v2_64",
         "cublasCher2k_64",
         "cublasCher2_v2_64",

diff --git a/docs/tables/CUBLAS_API_supported_by_HIP_and_ROC.md b/docs/tables/CUBLAS_API_supported_by_HIP_and_ROC.md
@@ -751,9 +751,9 @@
 |`cublasCher2_64`|12.0| | | |`hipblasCher2_v2_64`|6.2.0| | | | | | | | | | |
 |`cublasCher2_v2`| | | | |`hipblasCher2_v2`|6.0.0| | | | |`rocblas_cher2`|3.5.0| | | | |
 |`cublasCher2_v2_64`|12.0| | | |`hipblasCher2_v2_64`|6.2.0| | | | | | | | | | |
-|`cublasCher_64`|12.0| | | |`hipblasCher_v2_64`|6.2.0| | | | | | | | | | |
+|`cublasCher_64`|12.0| | | |`hipblasCher_v2_64`|6.2.0| | | | |`rocblas_cher_64`|6.2.0| | | | |
 |`cublasCher_v2`| | | | |`hipblasCher_v2`|6.0.0| | | | |`rocblas_cher`|3.5.0| | | | |
-|`cublasCher_v2_64`|12.0| | | |`hipblasCher_v2_64`|6.2.0| | | | | | | | | | |
+|`cublasCher_v2_64`|12.0| | | |`hipblasCher_v2_64`|6.2.0| | | | |`rocblas_cher_64`|6.2.0| | | | |
 |`cublasChpmv`| | | | |`hipblasChpmv_v2`|6.0.0| | | | |`rocblas_chpmv`|3.5.0| | | | |
 |`cublasChpmv_64`|12.0| | | |`hipblasChpmv_v2_64`|6.2.0| | | | | | | | | | |
 |`cublasChpmv_v2`| | | | |`hipblasChpmv_v2`|6.0.0| | | | |`rocblas_chpmv`|3.5.0| | | | |
@@ -775,9 +775,9 @@
 |`cublasCsyr2_64`|12.0| | | |`hipblasCsyr2_v2_64`|6.2.0| | | | | | | | | | |
 |`cublasCsyr2_v2`| | | | |`hipblasCsyr2_v2`|6.0.0| | | | |`rocblas_csyr2`|3.5.0| | | | |
 |`cublasCsyr2_v2_64`|12.0| | | |`hipblasCsyr2_v2_64`|6.2.0| | | | | | | | | | |
-|`cublasCsyr_64`|12.0| | | |`hipblasCsyr_v2_64`|6.2.0| | | | | | | | | | |
+|`cublasCsyr_64`|12.0| | | |`hipblasCsyr_v2_64`|6.2.0| | | | |`rocblas_csyr_64`|6.2.0| | | | |
 |`cublasCsyr_v2`| | | | |`hipblasCsyr_v2`|6.0.0| | | | |`rocblas_csyr`|1.7.1| | | | |
-|`cublasCsyr_v2_64`|12.0| | | |`hipblasCsyr_v2_64`|6.2.0| | | | | | | | | | |
+|`cublasCsyr_v2_64`|12.0| | | |`hipblasCsyr_v2_64`|6.2.0| | | | |`rocblas_csyr_64`|6.2.0| | | | |
 |`cublasCtbmv`| | | | |`hipblasCtbmv_v2`|6.0.0| | | | |`rocblas_ctbmv`|3.5.0| | | | |
 |`cublasCtbmv_64`|12.0| | | |`hipblasCtbmv_v2_64`|6.2.0| | | | | | | | | | |
 |`cublasCtbmv_v2`| | | | |`hipblasCtbmv_v2`|6.0.0| | | | |`rocblas_ctbmv`|3.5.0| | | | |
@@ -839,9 +839,9 @@
 |`cublasDsyr2_64`|12.0| | | |`hipblasDsyr2_64`|6.2.0| | | | | | | | | | |
 |`cublasDsyr2_v2`| | | | |`hipblasDsyr2`|3.5.0| | | | |`rocblas_dsyr2`|3.5.0| | | | |
 |`cublasDsyr2_v2_64`|12.0| | | |`hipblasDsyr2_64`|6.2.0| | | | | | | | | | |
-|`cublasDsyr_64`|12.0| | | |`hipblasDsyr_64`|6.2.0| | | | | | | | | | |
+|`cublasDsyr_64`|12.0| | | |`hipblasDsyr_64`|6.2.0| | | | |`rocblas_dsyr_64`|6.2.0| | | | |
 |`cublasDsyr_v2`| | | | |`hipblasDsyr`|3.0.0| | | | |`rocblas_dsyr`|1.7.1| | | | |
-|`cublasDsyr_v2_64`|12.0| | | |`hipblasDsyr_64`|6.2.0| | | | | | | | | | |
+|`cublasDsyr_v2_64`|12.0| | | |`hipblasDsyr_64`|6.2.0| | | | |`rocblas_dsyr_64`|6.2.0| | | | |
 |`cublasDtbmv`| | | | |`hipblasDtbmv`|3.5.0| | | | |`rocblas_dtbmv`|3.5.0| | | | |
 |`cublasDtbmv_64`|12.0| | | |`hipblasDtbmv_64`|6.2.0| | | | | | | | | | |
 |`cublasDtbmv_v2`| | | | |`hipblasDtbmv`|3.5.0| | | | |`rocblas_dtbmv`|3.5.0| | | | |
@@ -903,9 +903,9 @@
 |`cublasSsyr2_64`|12.0| | | |`hipblasSsyr2_64`|6.2.0| | | | | | | | | | |
 |`cublasSsyr2_v2`| | | | |`hipblasSsyr2`|3.5.0| | | | |`rocblas_ssyr2`|3.5.0| | | | |
 |`cublasSsyr2_v2_64`|12.0| | | |`hipblasSsyr2_64`|6.2.0| | | | | | | | | | |
-|`cublasSsyr_64`|12.0| | | |`hipblasSsyr_64`|6.2.0| | | | | | | | | | |
+|`cublasSsyr_64`|12.0| | | |`hipblasSsyr_64`|6.2.0| | | | |`rocblas_ssyr_64`|6.2.0| | | | |
 |`cublasSsyr_v2`| | | | |`hipblasSsyr`|3.0.0| | | | |`rocblas_ssyr`|1.7.1| | | | |
-|`cublasSsyr_v2_64`|12.0| | | |`hipblasSsyr_64`|6.2.0| | | | | | | | | | |
+|`cublasSsyr_v2_64`|12.0| | | |`hipblasSsyr_64`|6.2.0| | | | |`rocblas_ssyr_64`|6.2.0| | | | |
 |`cublasStbmv`| | | | |`hipblasStbmv`|3.5.0| | | | |`rocblas_stbmv`|3.5.0| | | | |
 |`cublasStbmv_64`|12.0| | | |`hipblasStbmv_64`|6.2.0| | | | | | | | | | |
 |`cublasStbmv_v2`| | | | |`hipblasStbmv`|3.5.0| | | | |`rocblas_stbmv`|3.5.0| | | | |
@@ -959,9 +959,9 @@
 |`cublasZher2_64`|12.0| | | |`hipblasZher2_v2_64`|6.2.0| | | | | | | | | | |
 |`cublasZher2_v2`| | | | |`hipblasZher2_v2`|6.0.0| | | | |`rocblas_zher2`|3.5.0| | | | |
 |`cublasZher2_v2_64`|12.0| | | |`hipblasZher2_v2_64`|6.2.0| | | | | | | | | | |
-|`cublasZher_64`|12.0| | | |`hipblasZher_v2_64`|6.2.0| | | | | | | | | | |
+|`cublasZher_64`|12.0| | | |`hipblasZher_v2_64`|6.2.0| | | | |`rocblas_zher_64`|6.2.0| | | | |
 |`cublasZher_v2`| | | | |`hipblasZher_v2`|6.0.0| | | | |`rocblas_zher`|3.5.0| | | | |
-|`cublasZher_v2_64`|12.0| | | |`hipblasZher_v2_64`|6.2.0| | | | | | | | | | |
+|`cublasZher_v2_64`|12.0| | | |`hipblasZher_v2_64`|6.2.0| | | | |`rocblas_zher_64`|6.2.0| | | | |
 |`cublasZhpmv`| | | | |`hipblasZhpmv_v2`|6.0.0| | | | |`rocblas_zhpmv`|3.5.0| | | | |
 |`cublasZhpmv_64`|12.0| | | |`hipblasZhpmv_v2_64`|6.2.0| | | | | | | | | | |
 |`cublasZhpmv_v2`| | | | |`hipblasZhpmv_v2`|6.0.0| | | | |`rocblas_zhpmv`|3.5.0| | | | |
@@ -983,9 +983,9 @@
 |`cublasZsyr2_64`|12.0| | | |`hipblasZsyr2_v2_64`|6.2.0| | | | | | | | | | |
 |`cublasZsyr2_v2`| | | | |`hipblasZsyr2_v2`|6.0.0| | | | |`rocblas_zsyr2`|3.5.0| | | | |
 |`cublasZsyr2_v2_64`|12.0| | | |`hipblasZsyr2_v2_64`|6.2.0| | | | | | | | | | |
-|`cublasZsyr_64`|12.0| | | |`hipblasZsyr_v2_64`|6.2.0| | | | | | | | | | |
+|`cublasZsyr_64`|12.0| | | |`hipblasZsyr_v2_64`|6.2.0| | | | |`rocblas_zsyr_64`|6.2.0| | | | |
 |`cublasZsyr_v2`| | | | |`hipblasZsyr_v2`|6.0.0| | | | |`rocblas_zsyr`|1.7.1| | | | |
-|`cublasZsyr_v2_64`|12.0| | | |`hipblasZsyr_v2_64`|6.2.0| | | | | | | | | | |
+|`cublasZsyr_v2_64`|12.0| | | |`hipblasZsyr_v2_64`|6.2.0| | | | |`rocblas_zsyr_64`|6.2.0| | | | |
 |`cublasZtbmv`| | | | |`hipblasZtbmv_v2`|6.0.0| | | | |`rocblas_ztbmv`|3.5.0| | | | |
 |`cublasZtbmv_64`|12.0| | | |`hipblasZtbmv_v2_64`|6.2.0| | | | | | | | | | |
 |`cublasZtbmv_v2`| | | | |`hipblasZtbmv_v2`|6.0.0| | | | |`rocblas_ztbmv`|3.5.0| | | | |

diff --git a/docs/tables/CUBLAS_API_supported_by_ROC.md b/docs/tables/CUBLAS_API_supported_by_ROC.md
@@ -751,9 +751,9 @@
 |`cublasCher2_64`|12.0| | | | | | | | | |
 |`cublasCher2_v2`| | | | |`rocblas_cher2`|3.5.0| | | | |
 |`cublasCher2_v2_64`|12.0| | | | | | | | | |
-|`cublasCher_64`|12.0| | | | | | | | | |
+|`cublasCher_64`|12.0| | | |`rocblas_cher_64`|6.2.0| | | | |
 |`cublasCher_v2`| | | | |`rocblas_cher`|3.5.0| | | | |
-|`cublasCher_v2_64`|12.0| | | | | | | | | |
+|`cublasCher_v2_64`|12.0| | | |`rocblas_cher_64`|6.2.0| | | | |
 |`cublasChpmv`| | | | |`rocblas_chpmv`|3.5.0| | | | |
 |`cublasChpmv_64`|12.0| | | | | | | | | |
 |`cublasChpmv_v2`| | | | |`rocblas_chpmv`|3.5.0| | | | |
@@ -775,9 +775,9 @@
 |`cublasCsyr2_64`|12.0| | | | | | | | | |
 |`cublasCsyr2_v2`| | | | |`rocblas_csyr2`|3.5.0| | | | |
 |`cublasCsyr2_v2_64`|12.0| | | | | | | | | |
-|`cublasCsyr_64`|12.0| | | | | | | | | |
+|`cublasCsyr_64`|12.0| | | |`rocblas_csyr_64`|6.2.0| | | | |
 |`cublasCsyr_v2`| | | | |`rocblas_csyr`|1.7.1| | | | |
-|`cublasCsyr_v2_64`|12.0| | | | | | | | | |
+|`cublasCsyr_v2_64`|12.0| | | |`rocblas_csyr_64`|6.2.0| | | | |
 |`cublasCtbmv`| | | | |`rocblas_ctbmv`|3.5.0| | | | |
 |`cublasCtbmv_64`|12.0| | | | | | | | | |
 |`cublasCtbmv_v2`| | | | |`rocblas_ctbmv`|3.5.0| | | | |
@@ -839,9 +839,9 @@
 |`cublasDsyr2_64`|12.0| | | | | | | | | |
 |`cublasDsyr2_v2`| | | | |`rocblas_dsyr2`|3.5.0| | | | |
 |`cublasDsyr2_v2_64`|12.0| | | | | | | | | |
-|`cublasDsyr_64`|12.0| | | | | | | | | |
+|`cublasDsyr_64`|12.0| | | |`rocblas_dsyr_64`|6.2.0| | | | |
 |`cublasDsyr_v2`| | | | |`rocblas_dsyr`|1.7.1| | | | |
-|`cublasDsyr_v2_64`|12.0| | | | | | | | | |
+|`cublasDsyr_v2_64`|12.0| | | |`rocblas_dsyr_64`|6.2.0| | | | |
 |`cublasDtbmv`| | | | |`rocblas_dtbmv`|3.5.0| | | | |
 |`cublasDtbmv_64`|12.0| | | | | | | | | |
 |`cublasDtbmv_v2`| | | | |`rocblas_dtbmv`|3.5.0| | | | |
@@ -903,9 +903,9 @@
 |`cublasSsyr2_64`|12.0| | | | | | | | | |
 |`cublasSsyr2_v2`| | | | |`rocblas_ssyr2`|3.5.0| | | | |
 |`cublasSsyr2_v2_64`|12.0| | | | | | | | | |
-|`cublasSsyr_64`|12.0| | | | | | | | | |
+|`cublasSsyr_64`|12.0| | | |`rocblas_ssyr_64`|6.2.0| | | | |
 |`cublasSsyr_v2`| | | | |`rocblas_ssyr`|1.7.1| | | | |
-|`cublasSsyr_v2_64`|12.0| | | | | | | | | |
+|`cublasSsyr_v2_64`|12.0| | | |`rocblas_ssyr_64`|6.2.0| | | | |
 |`cublasStbmv`| | | | |`rocblas_stbmv`|3.5.0| | | | |
 |`cublasStbmv_64`|12.0| | | | | | | | | |
 |`cublasStbmv_v2`| | | | |`rocblas_stbmv`|3.5.0| | | | |
@@ -959,9 +959,9 @@
 |`cublasZher2_64`|12.0| | | | | | | | | |
 |`cublasZher2_v2`| | | | |`rocblas_zher2`|3.5.0| | | | |
 |`cublasZher2_v2_64`|12.0| | | | | | | | | |
-|`cublasZher_64`|12.0| | | | | | | | | |
+|`cublasZher_64`|12.0| | | |`rocblas_zher_64`|6.2.0| | | | |
 |`cublasZher_v2`| | | | |`rocblas_zher`|3.5.0| | | | |
-|`cublasZher_v2_64`|12.0| | | | | | | | | |
+|`cublasZher_v2_64`|12.0| | | |`rocblas_zher_64`|6.2.0| | | | |
 |`cublasZhpmv`| | | | |`rocblas_zhpmv`|3.5.0| | | | |
 |`cublasZhpmv_64`|12.0| | | | | | | | | |
 |`cublasZhpmv_v2`| | | | |`rocblas_zhpmv`|3.5.0| | | | |
@@ -983,9 +983,9 @@
 |`cublasZsyr2_64`|12.0| | | | | | | | | |
 |`cublasZsyr2_v2`| | | | |`rocblas_zsyr2`|3.5.0| | | | |
 |`cublasZsyr2_v2_64`|12.0| | | | | | | | | |
-|`cublasZsyr_64`|12.0| | | | | | | | | |
+|`cublasZsyr_64`|12.0| | | |`rocblas_zsyr_64`|6.2.0| | | | |
 |`cublasZsyr_v2`| | | | |`rocblas_zsyr`|1.7.1| | | | |
-|`cublasZsyr_v2_64`|12.0| | | | | | | | | |
+|`cublasZsyr_v2_64`|12.0| | | |`rocblas_zsyr_64`|6.2.0| | | | |
 |`cublasZtbmv`| | | | |`rocblas_ztbmv`|3.5.0| | | | |
 |`cublasZtbmv_64`|12.0| | | | | | | | | |
 |`cublasZtbmv_v2`| | | | |`rocblas_ztbmv`|3.5.0| | | | |