maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/polly/test/ScopInfo/invariant-loads-leave-read-...

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

118 lines
5.1 KiB
LLVM
Raw Normal View History

[tests] Force invariant load hoisting for test cases that need it This will make it easier to switch the default of Polly's invariant load hoisting strategy and also makes it very clear that these test cases indeed require invariant code hoisting to work. llvm-svn: 278667
2016-08-15 21:27:49 +08:00
; RUN: opt %loadPolly -polly-scops -polly-invariant-load-hoisting=true -analyze < %s | FileCheck %s
; RUN: opt %loadPolly -polly-codegen -polly-invariant-load-hoisting=true -analyze < %s
Prepare unit tests for update to ISL 0.16 ISL 0.16 will change how sets are printed which breaks 117 unit tests that text-compare printed sets. This patch re-formats most of these unit tests using a script and small manual editing on top of that. When actually updating ISL, most work is done by just re-running the script to adapt to the changed output. Some tests that compare IR and tests with single CHECK-lines that can be easily updated manually are not included here. The re-format script will also be committed afterwards. The per-test formatter invocation command lines options will not be added in the near future because it is ad hoc and would overwrite the manual edits. Ideally it also shouldn't be required anymore because ISL's set printing has become more stable in 0.16. Differential Revision: http://reviews.llvm.org/D16095 llvm-svn: 257851
2016-01-15 08:48:42 +08:00
; CHECK: Statements {
[Invariant Loads] Do not consider invariant loads to have dependences. We need to relax constraints on invariant loads so that they do not create fake RAW dependences. So, we do not consider invariant loads as scalar dependences in a region. During these changes, it turned out that we do not consider `llvm::Value` replacements correctly within `PPCGCodeGeneration` and `ISLNodeBuilder`. The replacements dictated by `ValueMap` were not being followed in all places. This was fixed in this commit. There is no clean way to decouple this change because this bug only seems to arise when the relaxed version of invariant load hoisting was enabled. Differential Revision: https://reviews.llvm.org/D35120 llvm-svn: 307907
2017-07-13 20:18:56 +08:00
; CHECK-NEXT: Stmt_L_4
Prepare unit tests for update to ISL 0.16 ISL 0.16 will change how sets are printed which breaks 117 unit tests that text-compare printed sets. This patch re-formats most of these unit tests using a script and small manual editing on top of that. When actually updating ISL, most work is done by just re-running the script to adapt to the changed output. Some tests that compare IR and tests with single CHECK-lines that can be easily updated manually are not included here. The re-format script will also be committed afterwards. The per-test formatter invocation command lines options will not be added in the near future because it is ad hoc and would overwrite the manual edits. Ideally it also shouldn't be required anymore because ISL's set printing has become more stable in 0.16. Differential Revision: http://reviews.llvm.org/D16095 llvm-svn: 257851
2016-01-15 08:48:42 +08:00
; CHECK-NEXT: Domain :=
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
; CHECK-NEXT: [tmp8, tmp22, tmp15] -> { Stmt_L_4[i0, i1, i2] : 0 <= i0 < tmp8 and 0 <= i1 < tmp8 and 0 <= i2 < tmp8 };
Prepare unit tests for update to ISL 0.16 ISL 0.16 will change how sets are printed which breaks 117 unit tests that text-compare printed sets. This patch re-formats most of these unit tests using a script and small manual editing on top of that. When actually updating ISL, most work is done by just re-running the script to adapt to the changed output. Some tests that compare IR and tests with single CHECK-lines that can be easily updated manually are not included here. The re-format script will also be committed afterwards. The per-test formatter invocation command lines options will not be added in the near future because it is ad hoc and would overwrite the manual edits. Ideally it also shouldn't be required anymore because ISL's set printing has become more stable in 0.16. Differential Revision: http://reviews.llvm.org/D16095 llvm-svn: 257851
2016-01-15 08:48:42 +08:00
; CHECK-NEXT: Schedule :=
[Invariant Loads] Do not consider invariant loads to have dependences. We need to relax constraints on invariant loads so that they do not create fake RAW dependences. So, we do not consider invariant loads as scalar dependences in a region. During these changes, it turned out that we do not consider `llvm::Value` replacements correctly within `PPCGCodeGeneration` and `ISLNodeBuilder`. The replacements dictated by `ValueMap` were not being followed in all places. This was fixed in this commit. There is no clean way to decouple this change because this bug only seems to arise when the relaxed version of invariant load hoisting was enabled. Differential Revision: https://reviews.llvm.org/D35120 llvm-svn: 307907
2017-07-13 20:18:56 +08:00
; CHECK-NEXT: [tmp8, tmp22, tmp15] -> { Stmt_L_4[i0, i1, i2] -> [i0, i1, i2] };
; CHECK-NEXT: ReadAccess := [Reduction Type: NONE] [Scalar: 0]
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
; CHECK-NEXT: [tmp8, tmp22, tmp15] -> { Stmt_L_4[i0, i1, i2] -> MemRef_tmp19[i1, i0] };
[Invariant Loads] Do not consider invariant loads to have dependences. We need to relax constraints on invariant loads so that they do not create fake RAW dependences. So, we do not consider invariant loads as scalar dependences in a region. During these changes, it turned out that we do not consider `llvm::Value` replacements correctly within `PPCGCodeGeneration` and `ISLNodeBuilder`. The replacements dictated by `ValueMap` were not being followed in all places. This was fixed in this commit. There is no clean way to decouple this change because this bug only seems to arise when the relaxed version of invariant load hoisting was enabled. Differential Revision: https://reviews.llvm.org/D35120 llvm-svn: 307907
2017-07-13 20:18:56 +08:00
; CHECK-NEXT: ReadAccess := [Reduction Type: NONE] [Scalar: 0]
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
; CHECK-NEXT: [tmp8, tmp22, tmp15] -> { Stmt_L_4[i0, i1, i2] -> MemRef_tmp5[i2, i0] };
[Invariant Loads] Do not consider invariant loads to have dependences. We need to relax constraints on invariant loads so that they do not create fake RAW dependences. So, we do not consider invariant loads as scalar dependences in a region. During these changes, it turned out that we do not consider `llvm::Value` replacements correctly within `PPCGCodeGeneration` and `ISLNodeBuilder`. The replacements dictated by `ValueMap` were not being followed in all places. This was fixed in this commit. There is no clean way to decouple this change because this bug only seems to arise when the relaxed version of invariant load hoisting was enabled. Differential Revision: https://reviews.llvm.org/D35120 llvm-svn: 307907
2017-07-13 20:18:56 +08:00
; CHECK-NEXT: ReadAccess := [Reduction Type: NONE] [Scalar: 0]
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
; CHECK-NEXT: [tmp8, tmp22, tmp15] -> { Stmt_L_4[i0, i1, i2] -> MemRef_tmp12[i2, i1] };
[Invariant Loads] Do not consider invariant loads to have dependences. We need to relax constraints on invariant loads so that they do not create fake RAW dependences. So, we do not consider invariant loads as scalar dependences in a region. During these changes, it turned out that we do not consider `llvm::Value` replacements correctly within `PPCGCodeGeneration` and `ISLNodeBuilder`. The replacements dictated by `ValueMap` were not being followed in all places. This was fixed in this commit. There is no clean way to decouple this change because this bug only seems to arise when the relaxed version of invariant load hoisting was enabled. Differential Revision: https://reviews.llvm.org/D35120 llvm-svn: 307907
2017-07-13 20:18:56 +08:00
; CHECK-NEXT: MustWriteAccess := [Reduction Type: NONE] [Scalar: 0]
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
; CHECK-NEXT: [tmp8, tmp22, tmp15] -> { Stmt_L_4[i0, i1, i2] -> MemRef_tmp19[i1, i0] };
Prepare unit tests for update to ISL 0.16 ISL 0.16 will change how sets are printed which breaks 117 unit tests that text-compare printed sets. This patch re-formats most of these unit tests using a script and small manual editing on top of that. When actually updating ISL, most work is done by just re-running the script to adapt to the changed output. Some tests that compare IR and tests with single CHECK-lines that can be easily updated manually are not included here. The re-format script will also be committed afterwards. The per-test formatter invocation command lines options will not be added in the near future because it is ad hoc and would overwrite the manual edits. Ideally it also shouldn't be required anymore because ISL's set printing has become more stable in 0.16. Differential Revision: http://reviews.llvm.org/D16095 llvm-svn: 257851
2016-01-15 08:48:42 +08:00
; CHECK-NEXT: }
Remove read-only statements from the SCoP We do not need to model read-only statements in the SCoP as they will not cause any side effects that are visible to the outside anyway. Removing them should safe us time and might even simplify the ASTs we generate. Differential Revision: http://reviews.llvm.org/D14272 llvm-svn: 251948
2015-11-04 00:54:49 +08:00
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
%jl_value_t = type { %jl_value_t* }
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
define %jl_value_t* @julia_gemm_22583(%jl_value_t*, %jl_value_t** %tmp1, i32) {
Remove read-only statements from the SCoP We do not need to model read-only statements in the SCoP as they will not cause any side effects that are visible to the outside anyway. Removing them should safe us time and might even simplify the ASTs we generate. Differential Revision: http://reviews.llvm.org/D14272 llvm-svn: 251948
2015-11-04 00:54:49 +08:00
top:
br label %top.split
top.split: ; preds = %top
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
%tmp3 = load %jl_value_t*, %jl_value_t** %tmp1, align 8
%tmp4 = bitcast %jl_value_t* %tmp3 to double**
%tmp5 = load double*, double** %tmp4, align 8
%tmp6 = getelementptr inbounds %jl_value_t, %jl_value_t* %tmp3, i64 3, i32 0
%tmp7 = bitcast %jl_value_t** %tmp6 to i64*
%tmp8 = load i64, i64* %tmp7, align 8
%tmp9 = getelementptr %jl_value_t*, %jl_value_t** %tmp1, i64 1
%tmp10 = load %jl_value_t*, %jl_value_t** %tmp9, align 8
%tmp11 = bitcast %jl_value_t* %tmp10 to double**
%tmp12 = load double*, double** %tmp11, align 8
%tmp13 = getelementptr inbounds %jl_value_t, %jl_value_t* %tmp10, i64 3, i32 0
%tmp14 = bitcast %jl_value_t** %tmp13 to i64*
%tmp15 = load i64, i64* %tmp14, align 8
%tmp16 = getelementptr %jl_value_t*, %jl_value_t** %tmp1, i64 2
%tmp17 = load %jl_value_t*, %jl_value_t** %tmp16, align 8
%tmp18 = bitcast %jl_value_t* %tmp17 to double**
%tmp19 = load double*, double** %tmp18, align 8
%tmp20 = getelementptr inbounds %jl_value_t, %jl_value_t* %tmp17, i64 3, i32 0
%tmp21 = bitcast %jl_value_t** %tmp20 to i64*
%tmp22 = load i64, i64* %tmp21, align 8
%tmp23 = icmp sgt i64 %tmp8, 0
%tmp24 = select i1 %tmp23, i64 %tmp8, i64 0
%tmp25 = add i64 %tmp24, 1
%tmp26 = icmp eq i64 %tmp24, 0
br i1 %tmp26, label %L.11, label %L.preheader
Remove read-only statements from the SCoP We do not need to model read-only statements in the SCoP as they will not cause any side effects that are visible to the outside anyway. Removing them should safe us time and might even simplify the ASTs we generate. Differential Revision: http://reviews.llvm.org/D14272 llvm-svn: 251948
2015-11-04 00:54:49 +08:00
L.preheader: ; preds = %top.split
br label %L
L: ; preds = %L.preheader, %L.9
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
%"#s5.0" = phi i64 [ %tmp27, %L.9 ], [ 1, %L.preheader ]
%tmp27 = add i64 %"#s5.0", 1
br i1 %tmp26, label %L.9, label %L.2.preheader
Remove read-only statements from the SCoP We do not need to model read-only statements in the SCoP as they will not cause any side effects that are visible to the outside anyway. Removing them should safe us time and might even simplify the ASTs we generate. Differential Revision: http://reviews.llvm.org/D14272 llvm-svn: 251948
2015-11-04 00:54:49 +08:00
L.2.preheader: ; preds = %L
br label %L.2
L.2: ; preds = %L.2.preheader, %L.7
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
%"#s4.0" = phi i64 [ %tmp28, %L.7 ], [ 1, %L.2.preheader ]
%tmp28 = add i64 %"#s4.0", 1
br i1 %tmp26, label %L.7, label %L.4.preheader
Remove read-only statements from the SCoP We do not need to model read-only statements in the SCoP as they will not cause any side effects that are visible to the outside anyway. Removing them should safe us time and might even simplify the ASTs we generate. Differential Revision: http://reviews.llvm.org/D14272 llvm-svn: 251948
2015-11-04 00:54:49 +08:00
L.4.preheader: ; preds = %L.2
br label %L.4
L.4: ; preds = %L.4.preheader, %L.4
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
%"#s3.0" = phi i64 [ %tmp29, %L.4 ], [ 1, %L.4.preheader ]
%tmp29 = add i64 %"#s3.0", 1
%tmp30 = add i64 %"#s5.0", -1
%tmp31 = add i64 %"#s4.0", -1
%tmp32 = mul i64 %tmp31, %tmp22
%tmp33 = add i64 %tmp32, %tmp30
%tmp34 = getelementptr double, double* %tmp19, i64 %tmp33
%tmp35 = load double, double* %tmp34, align 8
%tmp36 = add i64 %"#s3.0", -1
%tmp37 = mul i64 %tmp36, %tmp8
%tmp38 = add i64 %tmp37, %tmp30
%tmp39 = getelementptr double, double* %tmp5, i64 %tmp38
%tmp40 = load double, double* %tmp39, align 8
%tmp41 = mul i64 %tmp36, %tmp15
%tmp42 = add i64 %tmp41, %tmp31
%tmp43 = getelementptr double, double* %tmp12, i64 %tmp42
%tmp44 = load double, double* %tmp43, align 8
%tmp45 = fmul double %tmp40, %tmp44
%tmp46 = fadd double %tmp35, %tmp45
store double %tmp46, double* %tmp34, align 8
%tmp47 = icmp eq i64 %tmp29, %tmp25
br i1 %tmp47, label %L.7.loopexit, label %L.4
Remove read-only statements from the SCoP We do not need to model read-only statements in the SCoP as they will not cause any side effects that are visible to the outside anyway. Removing them should safe us time and might even simplify the ASTs we generate. Differential Revision: http://reviews.llvm.org/D14272 llvm-svn: 251948
2015-11-04 00:54:49 +08:00
L.7.loopexit: ; preds = %L.4
br label %L.7
L.7: ; preds = %L.7.loopexit, %L.2
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
%tmp48 = icmp eq i64 %tmp28, %tmp25
br i1 %tmp48, label %L.9.loopexit, label %L.2
Remove read-only statements from the SCoP We do not need to model read-only statements in the SCoP as they will not cause any side effects that are visible to the outside anyway. Removing them should safe us time and might even simplify the ASTs we generate. Differential Revision: http://reviews.llvm.org/D14272 llvm-svn: 251948
2015-11-04 00:54:49 +08:00
L.9.loopexit: ; preds = %L.7
br label %L.9
L.9: ; preds = %L.9.loopexit, %L
[tests] Ensure all test cases use named variables This makes it easier to read and possibly even modify the test cases, as there is no need to keep the variable increment in steps of one. More importantly, by using explicit variable names we do not need to rely on the implicit numbering of statements when dumping the scop information. This makes it easier to read and possibly even modify the test cases. Furthermore, by using explicit variables we do not need to rely on the implicit numbering of statements when dumping the scop information. In a future commit, this implicit numbering will likely not be used any more to refer to LLVM-IR values as it is very expensive to construct. llvm-svn: 301689
2017-04-29 05:16:29 +08:00
%tmp49 = icmp eq i64 %tmp27, %tmp25
br i1 %tmp49, label %L.11.loopexit, label %L
Remove read-only statements from the SCoP We do not need to model read-only statements in the SCoP as they will not cause any side effects that are visible to the outside anyway. Removing them should safe us time and might even simplify the ASTs we generate. Differential Revision: http://reviews.llvm.org/D14272 llvm-svn: 251948
2015-11-04 00:54:49 +08:00
L.11.loopexit: ; preds = %L.9
br label %L.11
L.11: ; preds = %L.11.loopexit, %top.split
ret %jl_value_t* inttoptr (i64 140220477440016 to %jl_value_t*)
}