mirror of
https://github.com/llvm/llvm-project.git
synced 2025-04-26 21:06:06 +00:00
d23c5c2d65
This implements a global function merging pass. Unlike traditional function merging passes that use IR comparators, this pass employs a structurally stable hash to identify similar functions while ignoring certain constant operands. These ignored constants are tracked and encoded into a stable function summary. When merging, instead of explicitly folding similar functions and their call sites, we form a merging instance by supplying different parameters via thunks. The actual size reduction occurs when identically created merging instances are folded by the linker. Currently, this pass is wired to a pre-codegen pass, enabled by the `-enable-global-merge-func` flag. In a local merging mode, the analysis and merging steps occur sequentially within a module: - `analyze`: Collects stable function hashes and tracks locations of ignored constant operands. - `finalize`: Identifies merge candidates with matching hashes and computes the set of parameters that point to different constants. - `merge`: Uses the stable function map to optimistically create a merged function. We can enable a global merging mode similar to the global function outliner (https://discourse.llvm.org/t/rfc-enhanced-machine-outliner-part-2-thinlto-nolto/78753/), which will perform the above steps separately. - `-codegen-data-generate`: During the first round of code generation, we analyze local merging instances and publish their summaries. - Offline using `llvm-cgdata` or at link-time, we can finalize all these merging summaries that are combined to determine parameters. - `-codegen-data-use`: During the second round of code generation, we optimistically create merging instances within each module, and finally, the linker folds identically created merging instances. Depends on #112664 This is a patch for https://discourse.llvm.org/t/rfc-global-function-merging/82608.
67 lines
2.6 KiB
Plaintext
67 lines
2.6 KiB
Plaintext
# REQUIRES: shell, aarch64-registered-target
|
|
# UNSUPPORTED: system-windows
|
|
|
|
# Test merge a single object file having both __llvm_outline and __llvm_merge into a cgdata.
|
|
# Effectively, this test combines merge-hashtree.test and merge-funcmap.test.
|
|
|
|
RUN: split-file %s %t
|
|
|
|
# Synthesize raw hashtree bytes without the header (32 byte) from the indexed cgdata.
|
|
RUN: llvm-cgdata --convert --format binary %t/raw-hashtree.cgtext -o %t/raw-hashtree.cgdata
|
|
RUN: od -t x1 -j 32 -An %t/raw-hashtree.cgdata | tr -d '\n\r\t' | sed 's/[ ]*$//' | sed 's/[ ][ ]*/\\\\/g' > %t/raw-hashtree-bytes.txt
|
|
|
|
# Synthesize raw funcmap bytes without the header (32 byte) from the indexed cgdata.
|
|
RUN: llvm-cgdata --convert --format binary %t/raw-funcmap.cgtext -o %t/raw-funcmap.cgdata
|
|
RUN: od -t x1 -j 32 -An %t/raw-funcmap.cgdata | tr -d '\n\r\t' | sed 's/[ ]*$//' | sed 's/[ ][ ]*/\\\\/g' > %t/raw-funcmap-bytes.txt
|
|
|
|
# Synthesize a bitcode file by creating two sections for the hash tree and the function map, respectively.
|
|
RUN: sed "s/<RAW_1_BYTES>/$(cat %t/raw-hashtree-bytes.txt)/g" %t/merge-both-template.ll > %t/merge-both-hashtree-template.ll
|
|
RUN: sed "s/<RAW_2_BYTES>/$(cat %t/raw-funcmap-bytes.txt)/g" %t/merge-both-hashtree-template.ll > %t/merge-both-hashtree-funcmap.ll
|
|
|
|
RUN: llc -filetype=obj -mtriple arm64-apple-darwin %t/merge-both-hashtree-funcmap.ll -o %t/merge-both-hashtree-funcmap.o
|
|
|
|
# Merge an object file having cgdata (__llvm_outline and __llvm_merge)
|
|
RUN: llvm-cgdata -m --skip-trim %t/merge-both-hashtree-funcmap.o -o %t/merge-both-hashtree-funcmap.cgdata
|
|
RUN: llvm-cgdata -s %t/merge-both-hashtree-funcmap.cgdata | FileCheck %s
|
|
|
|
CHECK: Outlined hash tree:
|
|
CHECK-NEXT: Total Node Count: 3
|
|
CHECK-NEXT: Terminal Node Count: 1
|
|
CHECK-NEXT: Depth: 2
|
|
CHECK-NEXT: Stable function map:
|
|
CHECK-NEXT: Unique hash Count: 1
|
|
CHECK-NEXT: Total function Count: 1
|
|
CHECK-NEXT: Mergeable function Count: 0
|
|
|
|
;--- raw-hashtree.cgtext
|
|
:outlined_hash_tree
|
|
0:
|
|
Hash: 0x0
|
|
Terminals: 0
|
|
SuccessorIds: [ 1 ]
|
|
1:
|
|
Hash: 0x1
|
|
Terminals: 0
|
|
SuccessorIds: [ 2 ]
|
|
2:
|
|
Hash: 0x2
|
|
Terminals: 4
|
|
SuccessorIds: [ ]
|
|
...
|
|
|
|
;--- raw-funcmap.cgtext
|
|
:stable_function_map
|
|
- Hash: 1
|
|
FunctionName: Func1
|
|
ModuleName: Mod1
|
|
InstCount: 2
|
|
IndexOperandHashes:
|
|
- InstIndex: 0
|
|
OpndIndex: 1
|
|
OpndHash: 3
|
|
...
|
|
|
|
;--- merge-both-template.ll
|
|
@.data1 = private unnamed_addr constant [72 x i8] c"<RAW_1_BYTES>", section "__DATA,__llvm_outline"
|
|
@.data2 = private unnamed_addr constant [60 x i8] c"<RAW_2_BYTES>", section "__DATA,__llvm_merge"
|