e3cf80c5c1
BlockFrequencyInfo calculates block frequencies as Scaled64 numbers but as a last step converts them to unsigned 64bit integers (`BlockFrequency`). This improves the factors picked for this conversion so that: * Avoid big numbers close to UINT64_MAX to avoid users overflowing/saturating when adding multiply frequencies together or when multiplying with integers. This leaves the topmost 10 bits unused to allow for some room. * Spread the difference between hottest/coldest block as much as possible to increase precision. * If the hot/cold spread cannot be represented loose precision at the lower end, but keep the frequencies at the upper end for hot blocks differentiable.
46 lines
1.7 KiB
LLVM
46 lines
1.7 KiB
LLVM
; RUN: opt -thinlto-bc %s -write-relbf-to-summary -thin-link-bitcode-file=%t1.thinlink.bc -o %t1.bc
|
|
; RUN: opt -thinlto-bc %p/Inputs/function_entry_count.ll -write-relbf-to-summary -thin-link-bitcode-file=%t2.thinlink.bc -o %t2.bc
|
|
|
|
; First perform the thin link on the normal bitcode file.
|
|
; RUN: llvm-lto2 run %t1.bc %t2.bc -o %t.o -save-temps -disable-thinlto-funcattrs=0 -thinlto-synthesize-entry-counts \
|
|
; RUN: -r=%t1.bc,g, \
|
|
; RUN: -r=%t1.bc,f,px \
|
|
; RUN: -r=%t1.bc,h,px \
|
|
; RUN: -r=%t2.bc,h, \
|
|
; RUN: -r=%t2.bc,g,px
|
|
; RUN: llvm-dis -o - %t.o.1.3.import.bc | FileCheck %s
|
|
|
|
; RUN: llvm-lto -thinlto-action=run -disable-thinlto-funcattrs=0 -thinlto-synthesize-entry-counts -exported-symbol=f \
|
|
; RUN: -exported-symbol=g -exported-symbol=h -thinlto-save-temps=%t3. %t1.bc %t2.bc
|
|
; RUN: llvm-dis %t3.0.3.imported.bc -o - | FileCheck %s
|
|
|
|
; CHECK: define void @h() [[ATTR:#[0-9]+]] !prof ![[PROF2:[0-9]+]]
|
|
; CHECK: define void @f(i32{{.*}}) [[ATTR:#[0-9]+]] !prof ![[PROF1:[0-9]+]]
|
|
; CHECK: define available_externally void @g() !prof ![[PROF2]]
|
|
; CHECK-DAG: ![[PROF1]] = !{!"synthetic_function_entry_count", i64 10}
|
|
; CHECK-DAG: ![[PROF2]] = !{!"synthetic_function_entry_count", i64 200}
|
|
; CHECK-DAG: attributes [[ATTR]] = { norecurse nounwind }
|
|
|
|
target triple = "x86_64-unknown-linux-gnu"
|
|
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
|
|
|
|
declare void @g();
|
|
|
|
define void @h() {
|
|
ret void
|
|
}
|
|
|
|
define void @f(i32 %n) {
|
|
entry:
|
|
%cmp = icmp slt i32 %n, 1
|
|
br i1 %cmp, label %exit, label %loop
|
|
loop:
|
|
%n1 = phi i32 [%n, %entry], [%n2, %loop]
|
|
call void @g()
|
|
%n2 = sub i32 %n1, 1
|
|
%cmp2 = icmp slt i32 %n, 1
|
|
br i1 %cmp2, label %exit, label %loop
|
|
exit:
|
|
ret void
|
|
}
|