[llvm] [LV]Initial support for safe distance in predicated DataWithEVL vectorization mode. (PR #102897)

[Alexey Bataev via llvm-commits]

================
@@ -422,28 +422,37 @@ define void @no_high_lmul_or_interleave(ptr %p) {
 ; IF-EVL-NEXT:  entry:
 ; IF-EVL-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
 ; IF-EVL:       vector.ph:
+; IF-EVL-NEXT:    [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; IF-EVL-NEXT:    [[TMP1:%.*]] = sub i64 [[TMP0]], 1
+; IF-EVL-NEXT:    [[N_RND_UP:%.*]] = add i64 3002, [[TMP1]]
+; IF-EVL-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], [[TMP0]]
+; IF-EVL-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; IF-EVL-NEXT:    [[TMP2:%.*]] = call i64 @llvm.vscale.i64()
 ; IF-EVL-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; IF-EVL:       vector.body:
 ; IF-EVL-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; IF-EVL-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
-; IF-EVL-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i64> poison, i64 [[INDEX]], i64 0
-; IF-EVL-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i64> [[BROADCAST_SPLATINSERT]], <4 x i64> poison, <4 x i32> zeroinitializer
-; IF-EVL-NEXT:    [[VEC_IV:%.*]] = add <4 x i64> [[BROADCAST_SPLAT]], <i64 0, i64 1, i64 2, i64 3>
-; IF-EVL-NEXT:    [[TMP1:%.*]] = icmp ule <4 x i64> [[VEC_IV]], <i64 3001, i64 3001, i64 3001, i64 3001>
-; IF-EVL-NEXT:    [[TMP2:%.*]] = getelementptr i64, ptr [[P:%.*]], i64 [[TMP0]]
-; IF-EVL-NEXT:    [[TMP3:%.*]] = getelementptr i64, ptr [[TMP2]], i32 0
-; IF-EVL-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <4 x i64> @llvm.masked.load.v4i64.p0(ptr [[TMP3]], i32 32, <4 x i1> [[TMP1]], <4 x i64> poison)
-; IF-EVL-NEXT:    [[TMP4:%.*]] = add i64 [[TMP0]], 1024
-; IF-EVL-NEXT:    [[TMP5:%.*]] = getelementptr i64, ptr [[P]], i64 [[TMP4]]
-; IF-EVL-NEXT:    [[TMP6:%.*]] = getelementptr i64, ptr [[TMP5]], i32 0
-; IF-EVL-NEXT:    call void @llvm.masked.store.v4i64.p0(<4 x i64> [[WIDE_MASKED_LOAD]], ptr [[TMP6]], i32 32, <4 x i1> [[TMP1]])
-; IF-EVL-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 4
-; IF-EVL-NEXT:    [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], 3004
-; IF-EVL-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
+; IF-EVL-NEXT:    [[EVL_BASED_IV:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_EVL_NEXT:%.*]], [[VECTOR_BODY]] ]
+; IF-EVL-NEXT:    [[TMP3:%.*]] = icmp ult i64 [[EVL_BASED_IV]], 1024
+; IF-EVL-NEXT:    [[SAFE_AVL:%.*]] = select i1 [[TMP3]], i64 [[EVL_BASED_IV]], i64 1024
----------------
alexey-bataev wrote:

Yes, need to check if vectorization is safe for any VF before setting MaxSafeElements

https://github.com/llvm/llvm-project/pull/102897