X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=test%2FCodeGen%2FX86%2Fmisched-matrix.ll;h=23b561f6e5db417430a93951458aea825b069095;hb=4a524934577d85e5095df8ea62ad6a3261076d0c;hp=d00d17392a7e842eccc14c3401e829a04e5246b6;hpb=178f7d08a41f2e9432b96cd27f0c8ea42fa0ac9e;p=oota-llvm.git diff --git a/test/CodeGen/X86/misched-matrix.ll b/test/CodeGen/X86/misched-matrix.ll index d00d17392a7..23b561f6e5d 100644 --- a/test/CodeGen/X86/misched-matrix.ll +++ b/test/CodeGen/X86/misched-matrix.ll @@ -8,9 +8,6 @@ ; RUN: -misched=ilpmax -verify-machineinstrs \ ; RUN: | FileCheck %s -check-prefix=ILPMAX ; -; Very temporary xfail during SchedDFSResult churn. -; XFAIL: * -; ; Verify that the MI scheduler minimizes register pressure for a ; uniform set of bottom-up subtrees (unrolled matrix multiply). ; @@ -18,19 +15,19 @@ ; been reordered with the stores. This tests the scheduler's cheap ; alias analysis ability (that doesn't require any AliasAnalysis pass). ; -; TOPDOWN: %for.body +; TOPDOWN-LABEL: %for.body ; TOPDOWN: movl %{{.*}}, ( ; TOPDOWN: imull {{[0-9]*}}( ; TOPDOWN: movl %{{.*}}, 4( ; TOPDOWN: imull {{[0-9]*}}( ; TOPDOWN: movl %{{.*}}, 8( ; TOPDOWN: movl %{{.*}}, 12( -; TOPDOWN: %for.end +; TOPDOWN-LABEL: %for.end ; ; For -misched=ilpmin, verify that each expression subtree is ; scheduled independently, and that the imull/adds are interleaved. ; -; ILPMIN: %for.body +; ILPMIN-LABEL: %for.body ; ILPMIN: movl %{{.*}}, ( ; ILPMIN: imull ; ILPMIN: imull @@ -56,12 +53,12 @@ ; ILPMIN: imull ; ILPMIN: addl ; ILPMIN: movl %{{.*}}, 12( -; ILPMIN: %for.end +; ILPMIN-LABEL: %for.end ; ; For -misched=ilpmax, verify that each expression subtree is ; scheduled independently, and that the imull/adds are clustered. ; -; ILPMAX: %for.body +; ILPMAX-LABEL: %for.body ; ILPMAX: movl %{{.*}}, ( ; ILPMAX: imull ; ILPMAX: imull @@ -87,7 +84,7 @@ ; ILPMAX: addl ; ILPMAX: addl ; ILPMAX: movl %{{.*}}, 12( -; ILPMAX: %for.end +; ILPMAX-LABEL: %for.end define void @mmult([4 x i32]* noalias nocapture %m1, [4 x i32]* noalias nocapture %m2, [4 x i32]* noalias nocapture %m3) nounwind uwtable ssp { @@ -97,57 +94,57 @@ entry: for.body: ; preds = %for.body, %entry %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ] %arrayidx8 = getelementptr inbounds [4 x i32]* %m1, i64 %indvars.iv, i64 0 - %tmp = load i32* %arrayidx8, align 4, !tbaa !0 + %tmp = load i32* %arrayidx8, align 4 %arrayidx12 = getelementptr inbounds [4 x i32]* %m2, i64 0, i64 0 - %tmp1 = load i32* %arrayidx12, align 4, !tbaa !0 + %tmp1 = load i32* %arrayidx12, align 4 %arrayidx8.1 = getelementptr inbounds [4 x i32]* %m1, i64 %indvars.iv, i64 1 - %tmp2 = load i32* %arrayidx8.1, align 4, !tbaa !0 + %tmp2 = load i32* %arrayidx8.1, align 4 %arrayidx12.1 = getelementptr inbounds [4 x i32]* %m2, i64 1, i64 0 - %tmp3 = load i32* %arrayidx12.1, align 4, !tbaa !0 + %tmp3 = load i32* %arrayidx12.1, align 4 %arrayidx8.2 = getelementptr inbounds [4 x i32]* %m1, i64 %indvars.iv, i64 2 - %tmp4 = load i32* %arrayidx8.2, align 4, !tbaa !0 + %tmp4 = load i32* %arrayidx8.2, align 4 %arrayidx12.2 = getelementptr inbounds [4 x i32]* %m2, i64 2, i64 0 - %tmp5 = load i32* %arrayidx12.2, align 4, !tbaa !0 + %tmp5 = load i32* %arrayidx12.2, align 4 %arrayidx8.3 = getelementptr inbounds [4 x i32]* %m1, i64 %indvars.iv, i64 3 - %tmp6 = load i32* %arrayidx8.3, align 4, !tbaa !0 + %tmp6 = load i32* %arrayidx8.3, align 4 %arrayidx12.3 = getelementptr inbounds [4 x i32]* %m2, i64 3, i64 0 - %tmp8 = load i32* %arrayidx8, align 4, !tbaa !0 + %tmp8 = load i32* %arrayidx8, align 4 %arrayidx12.137 = getelementptr inbounds [4 x i32]* %m2, i64 0, i64 1 - %tmp9 = load i32* %arrayidx12.137, align 4, !tbaa !0 - %tmp10 = load i32* %arrayidx8.1, align 4, !tbaa !0 + %tmp9 = load i32* %arrayidx12.137, align 4 + %tmp10 = load i32* %arrayidx8.1, align 4 %arrayidx12.1.1 = getelementptr inbounds [4 x i32]* %m2, i64 1, i64 1 - %tmp11 = load i32* %arrayidx12.1.1, align 4, !tbaa !0 - %tmp12 = load i32* %arrayidx8.2, align 4, !tbaa !0 + %tmp11 = load i32* %arrayidx12.1.1, align 4 + %tmp12 = load i32* %arrayidx8.2, align 4 %arrayidx12.2.1 = getelementptr inbounds [4 x i32]* %m2, i64 2, i64 1 - %tmp13 = load i32* %arrayidx12.2.1, align 4, !tbaa !0 - %tmp14 = load i32* %arrayidx8.3, align 4, !tbaa !0 + %tmp13 = load i32* %arrayidx12.2.1, align 4 + %tmp14 = load i32* %arrayidx8.3, align 4 %arrayidx12.3.1 = getelementptr inbounds [4 x i32]* %m2, i64 3, i64 1 - %tmp15 = load i32* %arrayidx12.3.1, align 4, !tbaa !0 - %tmp16 = load i32* %arrayidx8, align 4, !tbaa !0 + %tmp15 = load i32* %arrayidx12.3.1, align 4 + %tmp16 = load i32* %arrayidx8, align 4 %arrayidx12.239 = getelementptr inbounds [4 x i32]* %m2, i64 0, i64 2 - %tmp17 = load i32* %arrayidx12.239, align 4, !tbaa !0 - %tmp18 = load i32* %arrayidx8.1, align 4, !tbaa !0 + %tmp17 = load i32* %arrayidx12.239, align 4 + %tmp18 = load i32* %arrayidx8.1, align 4 %arrayidx12.1.2 = getelementptr inbounds [4 x i32]* %m2, i64 1, i64 2 - %tmp19 = load i32* %arrayidx12.1.2, align 4, !tbaa !0 - %tmp20 = load i32* %arrayidx8.2, align 4, !tbaa !0 + %tmp19 = load i32* %arrayidx12.1.2, align 4 + %tmp20 = load i32* %arrayidx8.2, align 4 %arrayidx12.2.2 = getelementptr inbounds [4 x i32]* %m2, i64 2, i64 2 - %tmp21 = load i32* %arrayidx12.2.2, align 4, !tbaa !0 - %tmp22 = load i32* %arrayidx8.3, align 4, !tbaa !0 + %tmp21 = load i32* %arrayidx12.2.2, align 4 + %tmp22 = load i32* %arrayidx8.3, align 4 %arrayidx12.3.2 = getelementptr inbounds [4 x i32]* %m2, i64 3, i64 2 - %tmp23 = load i32* %arrayidx12.3.2, align 4, !tbaa !0 - %tmp24 = load i32* %arrayidx8, align 4, !tbaa !0 + %tmp23 = load i32* %arrayidx12.3.2, align 4 + %tmp24 = load i32* %arrayidx8, align 4 %arrayidx12.341 = getelementptr inbounds [4 x i32]* %m2, i64 0, i64 3 - %tmp25 = load i32* %arrayidx12.341, align 4, !tbaa !0 - %tmp26 = load i32* %arrayidx8.1, align 4, !tbaa !0 + %tmp25 = load i32* %arrayidx12.341, align 4 + %tmp26 = load i32* %arrayidx8.1, align 4 %arrayidx12.1.3 = getelementptr inbounds [4 x i32]* %m2, i64 1, i64 3 - %tmp27 = load i32* %arrayidx12.1.3, align 4, !tbaa !0 - %tmp28 = load i32* %arrayidx8.2, align 4, !tbaa !0 + %tmp27 = load i32* %arrayidx12.1.3, align 4 + %tmp28 = load i32* %arrayidx8.2, align 4 %arrayidx12.2.3 = getelementptr inbounds [4 x i32]* %m2, i64 2, i64 3 - %tmp29 = load i32* %arrayidx12.2.3, align 4, !tbaa !0 - %tmp30 = load i32* %arrayidx8.3, align 4, !tbaa !0 + %tmp29 = load i32* %arrayidx12.2.3, align 4 + %tmp30 = load i32* %arrayidx8.3, align 4 %arrayidx12.3.3 = getelementptr inbounds [4 x i32]* %m2, i64 3, i64 3 - %tmp31 = load i32* %arrayidx12.3.3, align 4, !tbaa !0 - %tmp7 = load i32* %arrayidx12.3, align 4, !tbaa !0 + %tmp31 = load i32* %arrayidx12.3.3, align 4 + %tmp7 = load i32* %arrayidx12.3, align 4 %mul = mul nsw i32 %tmp1, %tmp %mul.1 = mul nsw i32 %tmp3, %tmp2 %mul.2 = mul nsw i32 %tmp5, %tmp4 @@ -177,13 +174,13 @@ for.body: ; preds = %for.body, %entry %add.2.3 = add nsw i32 %mul.2.3, %add.1.3 %add.3.3 = add nsw i32 %mul.3.3, %add.2.3 %arrayidx16 = getelementptr inbounds [4 x i32]* %m3, i64 %indvars.iv, i64 0 - store i32 %add.3, i32* %arrayidx16, align 4, !tbaa !0 + store i32 %add.3, i32* %arrayidx16, align 4 %arrayidx16.1 = getelementptr inbounds [4 x i32]* %m3, i64 %indvars.iv, i64 1 - store i32 %add.3.1, i32* %arrayidx16.1, align 4, !tbaa !0 + store i32 %add.3.1, i32* %arrayidx16.1, align 4 %arrayidx16.2 = getelementptr inbounds [4 x i32]* %m3, i64 %indvars.iv, i64 2 - store i32 %add.3.2, i32* %arrayidx16.2, align 4, !tbaa !0 + store i32 %add.3.2, i32* %arrayidx16.2, align 4 %arrayidx16.3 = getelementptr inbounds [4 x i32]* %m3, i64 %indvars.iv, i64 3 - store i32 %add.3.3, i32* %arrayidx16.3, align 4, !tbaa !0 + store i32 %add.3.3, i32* %arrayidx16.3, align 4 %indvars.iv.next = add i64 %indvars.iv, 1 %lftr.wideiv = trunc i64 %indvars.iv.next to i32 %exitcond = icmp eq i32 %lftr.wideiv, 4 @@ -192,7 +189,3 @@ for.body: ; preds = %for.body, %entry for.end: ; preds = %for.body ret void } - -!0 = metadata !{metadata !"int", metadata !1} -!1 = metadata !{metadata !"omnipotent char", metadata !2} -!2 = metadata !{metadata !"Simple C/C++ TBAA"}