f513f44bc6
625038d5d5f406385f44983ac2a442b20ad7b241 added NewPM CGSCC coverage to all of the inliner tests. Now that the NewPM is the default for the middle end, both check lines cover the same thing (with -passes="inline" expanding to -passes="cgscc(inline)"). Given that, remove them to get rid of the redundancy and make the tests run slightly faster (not measured).
27 lines
1.1 KiB
LLVM
27 lines
1.1 KiB
LLVM
; RUN: opt < %s -passes=inline -inline-threshold=0 -debug-only=inline-cost -print-instruction-comments -S -mtriple=x86_64-unknown-linux-gnu 2>&1 | FileCheck %s
|
|
; REQUIRES: asserts
|
|
|
|
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
|
|
target triple = "x86_64-unknown-unknown"
|
|
|
|
; Check that insertvalue's aren't free.
|
|
|
|
; CHECK: Analyzing call of callee... (caller:caller_range)
|
|
; CHECK-NEXT: Initial cost: -40
|
|
; CHECK-NEXT: define { i32, i32 } @callee({ i32, i32 } %arg, i32 %arg1) {
|
|
; CHECK-NEXT: ; cost before = -40, cost after = -35, threshold before = 0, threshold after = 0, cost delta = 5
|
|
; CHECK-NEXT: %r = insertvalue { i32, i32 } %arg, i32 %arg1, 0
|
|
; CHECK-NEXT: ; cost before = -35, cost after = -35, threshold before = 0, threshold after = 0, cost delta = 0
|
|
; CHECK-NEXT: ret { i32, i32 } %r
|
|
; CHECK-NEXT: }
|
|
|
|
define {i32, i32} @callee({i32, i32} %arg, i32 %arg1) {
|
|
%r = insertvalue {i32, i32} %arg, i32 %arg1, 0
|
|
ret {i32, i32} %r
|
|
}
|
|
|
|
define {i32, i32} @caller_range({i32, i32} %arg, i32 %arg1) {
|
|
%r = call {i32, i32} @callee({i32, i32} %arg, i32 %arg1)
|
|
ret {i32, i32} %r
|
|
}
|