File: loopvectorize-costmodel.test

package info (click to toggle)
llvm-toolchain-16 1%3A16.0.6-15~deb12u1
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 1,634,792 kB
  • sloc: cpp: 6,179,261; ansic: 1,216,205; asm: 741,319; python: 196,614; objc: 75,325; f90: 49,640; lisp: 32,396; pascal: 12,286; sh: 9,394; perl: 7,442; ml: 5,494; awk: 3,523; makefile: 2,723; javascript: 1,206; xml: 886; fortran: 581; cs: 573
file content (13 lines) | stat: -rw-r--r-- 867 bytes parent folder | download | duplicates (12) 
1
2
3
4
5
6
7
8
9
10
11
12
13
 
# REQUIRES: x86-registered-target, asserts

## Check that --filter works properly.
# RUN: cp -f %S/Inputs/x86-loopvectorize-costmodel.ll %t.ll && %update_analyze_test_checks --filter "LV: Found an estimated cost of [0-9]+ for VF [0-9]+ For instruction:\s*%v0 = load float, ptr %in0, align 4" %t.ll
# RUN: diff -u %t.ll %S/Inputs/x86-loopvectorize-costmodel.ll.expected

## Check that running the script again does not change the result:
# RUN: %update_analyze_test_checks --filter "LV: Found an estimated cost of [0-9]+ for VF [0-9]+ For instruction:\s*%v0 = load float, ptr %in0, align 4" %t.ll
# RUN: diff -u %t.ll %S/Inputs/x86-loopvectorize-costmodel.ll.expected

## Check that running the script again, without arguments, does not change the result:
# RUN: %update_analyze_test_checks %t.ll
# RUN: diff -u %t.ll %S/Inputs/x86-loopvectorize-costmodel.ll.expected