maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/merge-store-partially-alias...

53 lines
2.2 KiB
LLVM
Raw Normal View History

Use new TokenFactor chain when merging stores If the stores are storing values from loads which partially alias the stores, we could end up placing the merged loads and stores on the same chain which has the potential to break. Each store may have a different chain dependency on only some of the original loads. Create a new TokenFactor to capture all of the required dependencies of the stores rather than assuming all stores can use the same chain. The testcase is a situation where this happens, although it does not have an observable change from this. The DAG nodes just happened to not be reordered before despite this missing chain dependency. This is based on an off-list report for an out of tree target which regressed due to r246307 and I haven't managed to find a case where the nodes do end up reordered with an in tree target. llvm-svn: 248468
2015-09-24 15:22:38 +08:00
; REQUIRES: asserts
; RUN: llc -march=x86-64 -mtriple=x86_64-unknown-linux-gnu < %s | FileCheck -check-prefix=X86 %s
; RUN: llc -march=x86-64 -mtriple=x86_64-unknown-linux-gnu -debug-only=isel < %s 2>&1 | FileCheck -check-prefix=DBGDAG %s
; It's OK to merge the load / store of the first 2 components, but
; they must not be placed on the same chain after merging.
; X86-LABEL: {{^}}merge_store_partial_overlap_load:
[X86] Fix PR23155 by turning on X86FixupBWInsts by default. Differential Revision: http://reviews.llvm.org/D18866 llvm-svn: 265830
2016-04-09 02:58:29 +08:00
; X86-DAG: movzwl ([[BASEREG:%[a-z]+]]), %e[[LO2:[a-z]+]]
Use new TokenFactor chain when merging stores If the stores are storing values from loads which partially alias the stores, we could end up placing the merged loads and stores on the same chain which has the potential to break. Each store may have a different chain dependency on only some of the original loads. Create a new TokenFactor to capture all of the required dependencies of the stores rather than assuming all stores can use the same chain. The testcase is a situation where this happens, although it does not have an observable change from this. The DAG nodes just happened to not be reordered before despite this missing chain dependency. This is based on an off-list report for an out of tree target which regressed due to r246307 and I haven't managed to find a case where the nodes do end up reordered with an in tree target. llvm-svn: 248468
2015-09-24 15:22:38 +08:00
; X86-DAG: movb 2([[BASEREG]]), [[HI1:%[a-z]+]]
[X86] Fix PR23155 by turning on X86FixupBWInsts by default. Differential Revision: http://reviews.llvm.org/D18866 llvm-svn: 265830
2016-04-09 02:58:29 +08:00
; X86-NEXT: movw %[[LO2]], 1([[BASEREG]])
Use new TokenFactor chain when merging stores If the stores are storing values from loads which partially alias the stores, we could end up placing the merged loads and stores on the same chain which has the potential to break. Each store may have a different chain dependency on only some of the original loads. Create a new TokenFactor to capture all of the required dependencies of the stores rather than assuming all stores can use the same chain. The testcase is a situation where this happens, although it does not have an observable change from this. The DAG nodes just happened to not be reordered before despite this missing chain dependency. This is based on an off-list report for an out of tree target which regressed due to r246307 and I haven't managed to find a case where the nodes do end up reordered with an in tree target. llvm-svn: 248468
2015-09-24 15:22:38 +08:00
; X86-NEXT: movb [[HI1]], 3([[BASEREG]])
; X86-NEXT: retq
; DBGDAG-LABEL: Optimized lowered selection DAG: BB#0 'merge_store_partial_overlap_load:'
; DBGDAG: [[ENTRYTOKEN:t[0-9]+]]: ch = EntryToken
; DBGDAG-DAG: [[BASEPTR:t[0-9]+]]: i64,ch = CopyFromReg [[ENTRYTOKEN]],
SelectionDAGDumper: Print simple operands inline. Print simple operands inline instead of their pointer/value number. Simple operands are SDNodes without predecessors like Constant(FP), Register, UNDEF. This unifies the behaviour with dumpr() which was already doing this. Previously: t0: ch = EntryToken t1: i64 = Register %vreg0 t2: i64,ch = CopyFromReg t0, t1 t3: i64 = Constant<1> t4: i64 = add t2, t3 t5: i64 = Constant<2> t6: i64 = add t2, t5 t10: i64 = undef t11: i8,ch = load t0, t2, t10<LD1[%tmp81]> t12: i8,ch = load t0, t4, t10<LD1[%tmp10]> t13: i8,ch = load t0, t6, t10<LD1[%tmp12]> Now: t0: ch = EntryToken t2: i64,ch = CopyFromReg t0, Register:i64 %vreg0 t4: i64 = add t2, Constant:i64<1> t6: i64 = add t2, Constant:i64<2> t11: i8,ch = load<LD1[%tmp81]> t0, t2, undef:i64 t12: i8,ch = load<LD1[%tmp10]> t0, t4, undef:i64 t13: i8,ch = load<LD1[%tmp12]> t0, t6, undef:i64 Differential Revision: http://reviews.llvm.org/D12567 llvm-svn: 248628
2015-09-26 06:27:02 +08:00
; DBGDAG-DAG: [[ADDPTR:t[0-9]+]]: i64 = add [[BASEPTR]], Constant:i64<2>
Use new TokenFactor chain when merging stores If the stores are storing values from loads which partially alias the stores, we could end up placing the merged loads and stores on the same chain which has the potential to break. Each store may have a different chain dependency on only some of the original loads. Create a new TokenFactor to capture all of the required dependencies of the stores rather than assuming all stores can use the same chain. The testcase is a situation where this happens, although it does not have an observable change from this. The DAG nodes just happened to not be reordered before despite this missing chain dependency. This is based on an off-list report for an out of tree target which regressed due to r246307 and I haven't managed to find a case where the nodes do end up reordered with an in tree target. llvm-svn: 248468
2015-09-24 15:22:38 +08:00
SelectionDAGDumper: Print simple operands inline. Print simple operands inline instead of their pointer/value number. Simple operands are SDNodes without predecessors like Constant(FP), Register, UNDEF. This unifies the behaviour with dumpr() which was already doing this. Previously: t0: ch = EntryToken t1: i64 = Register %vreg0 t2: i64,ch = CopyFromReg t0, t1 t3: i64 = Constant<1> t4: i64 = add t2, t3 t5: i64 = Constant<2> t6: i64 = add t2, t5 t10: i64 = undef t11: i8,ch = load t0, t2, t10<LD1[%tmp81]> t12: i8,ch = load t0, t4, t10<LD1[%tmp10]> t13: i8,ch = load t0, t6, t10<LD1[%tmp12]> Now: t0: ch = EntryToken t2: i64,ch = CopyFromReg t0, Register:i64 %vreg0 t4: i64 = add t2, Constant:i64<1> t6: i64 = add t2, Constant:i64<2> t11: i8,ch = load<LD1[%tmp81]> t0, t2, undef:i64 t12: i8,ch = load<LD1[%tmp10]> t0, t4, undef:i64 t13: i8,ch = load<LD1[%tmp12]> t0, t6, undef:i64 Differential Revision: http://reviews.llvm.org/D12567 llvm-svn: 248628
2015-09-26 06:27:02 +08:00
; DBGDAG-DAG: [[LD2:t[0-9]+]]: i16,ch = load<LD2[%tmp81](align=1)> [[ENTRYTOKEN]], [[BASEPTR]], undef:i64
; DBGDAG-DAG: [[LD1:t[0-9]+]]: i8,ch = load<LD1[%tmp12]> [[ENTRYTOKEN]], [[ADDPTR]], undef:i64
Use new TokenFactor chain when merging stores If the stores are storing values from loads which partially alias the stores, we could end up placing the merged loads and stores on the same chain which has the potential to break. Each store may have a different chain dependency on only some of the original loads. Create a new TokenFactor to capture all of the required dependencies of the stores rather than assuming all stores can use the same chain. The testcase is a situation where this happens, although it does not have an observable change from this. The DAG nodes just happened to not be reordered before despite this missing chain dependency. This is based on an off-list report for an out of tree target which regressed due to r246307 and I haven't managed to find a case where the nodes do end up reordered with an in tree target. llvm-svn: 248468
2015-09-24 15:22:38 +08:00
In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled. Recommiting with compiler time improvements Recommitting after fixup of 32-bit aliasing sign offset bug in DAGCombiner. * Simplify Consecutive Merge Store Candidate Search Now that address aliasing is much less conservative, push through simplified store merging search and chain alias analysis which only checks for parallel stores through the chain subgraph. This is cleaner as the separation of non-interfering loads/stores from the store-merging logic. When merging stores search up the chain through a single load, and finds all possible stores by looking down from through a load and a TokenFactor to all stores visited. This improves the quality of the output SelectionDAG and the output Codegen (save perhaps for some ARM cases where we correctly constructs wider loads, but then promotes them to float operations which appear but requires more expensive constant generation). Some minor peephole optimizations to deal with improved SubDAG shapes (listed below) Additional Minor Changes: 1. Finishes removing unused AliasLoad code 2. Unifies the chain aggregation in the merged stores across code paths 3. Re-add the Store node to the worklist after calling SimplifyDemandedBits. 4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is arbitrary, but seems sufficient to not cause regressions in tests. 5. Remove Chain dependencies of Memory operations on CopyfromReg nodes as these are captured by data dependence 6. Forward loads-store values through tokenfactors containing {CopyToReg,CopyFromReg} Values. 7. Peephole to convert buildvector of extract_vector_elt to extract_subvector if possible (see CodeGen/AArch64/store-merge.ll) 8. Store merging for the ARM target is restricted to 32-bit as some in some contexts invalid 64-bit operations are being generated. This can be removed once appropriate checks are added. This finishes the change Matt Arsenault started in r246307 and jyknight's original patch. Many tests required some changes as memory operations are now reorderable, improving load-store forwarding. One test in particular is worth noting: CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store forwarding converts a load-store pair into a parallel store and a memory-realized bitcast of the same value. However, because we lose the sharing of the explicit and implicit store values we must create another local store. A similar transformation happens before SelectionDAG as well. Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle llvm-svn: 297695
2017-03-14 08:34:14 +08:00
; DBGDAG-DAG: [[ST1:t[0-9]+]]: ch = store<ST1[%tmp14]> [[ENTRYTOKEN]], [[LD1]], t{{[0-9]+}}, undef:i64
; DBGDAG-DAG: [[LOADTOKEN:t[0-9]+]]: ch = TokenFactor [[LD2]]:1, [[LD1]]:1
[SDAG] Revert r296476 (and r296486, r296668, r296690). This patch causes compile times for some patterns to explode. I have a (large, unreduced) test case that slows down by more than 20x and several test cases slow down by 2x. I'm sending some of the test cases directly to Nirav and following up with more details in the review log, but this should unblock anyone else hitting this. llvm-svn: 296862
2017-03-03 18:02:25 +08:00
; DBGDAG-DAG: [[ST2:t[0-9]+]]: ch = store<ST2[%tmp10](align=1)> [[LOADTOKEN]], [[LD2]], t{{[0-9]+}}, undef:i64
In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled. Recommiting with compiler time improvements Recommitting after fixup of 32-bit aliasing sign offset bug in DAGCombiner. * Simplify Consecutive Merge Store Candidate Search Now that address aliasing is much less conservative, push through simplified store merging search and chain alias analysis which only checks for parallel stores through the chain subgraph. This is cleaner as the separation of non-interfering loads/stores from the store-merging logic. When merging stores search up the chain through a single load, and finds all possible stores by looking down from through a load and a TokenFactor to all stores visited. This improves the quality of the output SelectionDAG and the output Codegen (save perhaps for some ARM cases where we correctly constructs wider loads, but then promotes them to float operations which appear but requires more expensive constant generation). Some minor peephole optimizations to deal with improved SubDAG shapes (listed below) Additional Minor Changes: 1. Finishes removing unused AliasLoad code 2. Unifies the chain aggregation in the merged stores across code paths 3. Re-add the Store node to the worklist after calling SimplifyDemandedBits. 4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is arbitrary, but seems sufficient to not cause regressions in tests. 5. Remove Chain dependencies of Memory operations on CopyfromReg nodes as these are captured by data dependence 6. Forward loads-store values through tokenfactors containing {CopyToReg,CopyFromReg} Values. 7. Peephole to convert buildvector of extract_vector_elt to extract_subvector if possible (see CodeGen/AArch64/store-merge.ll) 8. Store merging for the ARM target is restricted to 32-bit as some in some contexts invalid 64-bit operations are being generated. This can be removed once appropriate checks are added. This finishes the change Matt Arsenault started in r246307 and jyknight's original patch. Many tests required some changes as memory operations are now reorderable, improving load-store forwarding. One test in particular is worth noting: CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store forwarding converts a load-store pair into a parallel store and a memory-realized bitcast of the same value. However, because we lose the sharing of the explicit and implicit store values we must create another local store. A similar transformation happens before SelectionDAG as well. Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle llvm-svn: 297695
2017-03-14 08:34:14 +08:00
; DBGDAG: X86ISD::RET_FLAG t{{[0-9]+}},
Use new TokenFactor chain when merging stores If the stores are storing values from loads which partially alias the stores, we could end up placing the merged loads and stores on the same chain which has the potential to break. Each store may have a different chain dependency on only some of the original loads. Create a new TokenFactor to capture all of the required dependencies of the stores rather than assuming all stores can use the same chain. The testcase is a situation where this happens, although it does not have an observable change from this. The DAG nodes just happened to not be reordered before despite this missing chain dependency. This is based on an off-list report for an out of tree target which regressed due to r246307 and I haven't managed to find a case where the nodes do end up reordered with an in tree target. llvm-svn: 248468
2015-09-24 15:22:38 +08:00
; DBGDAG: Type-legalized selection DAG: BB#0 'merge_store_partial_overlap_load:'
define void @merge_store_partial_overlap_load([4 x i8]* %tmp) {
%tmp8 = getelementptr inbounds [4 x i8], [4 x i8]* %tmp, i32 0, i8 0
%tmp10 = getelementptr inbounds [4 x i8], [4 x i8]* %tmp, i32 0, i8 1
%tmp12 = getelementptr inbounds [4 x i8], [4 x i8]* %tmp, i32 0, i8 2
%tmp14 = getelementptr [4 x i8], [4 x i8]* %tmp, i32 0, i8 3
%tmp9 = load i8, i8* %tmp8, align 1 ; base + 0
%tmp11 = load i8, i8* %tmp10, align 1 ; base + 1
%tmp13 = load i8, i8* %tmp12, align 1 ; base + 2
store i8 %tmp9, i8* %tmp10, align 1 ; base + 1
store i8 %tmp11, i8* %tmp12, align 1 ; base + 2
store i8 %tmp13, i8* %tmp14, align 1 ; base + 3
; Should emit
; load base + 0, base + 1
; store base + 1, base + 2
; load base + 2
; store base + 3
ret void
}