https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107905
--- Comment #7 from Georgi <sanmayce at sanmayce dot com> --- (In reply to Alexander Monakov from comment #5) > Not sure what you don't like about the inputs, they appear quite reasonable. > Perhaps GCC's estimation of bb frequencies is off (with profile feedback we > achieve good performance). > > Georgi: you'll likely see better results with profile-guided optimization. > You can first compile the benchmark with -O2 -fprofile-generate, run the > output (it will generate *.gcda files), then compile again with -O2 > -fprofile-use. For Clang the options are spelled -fprofile-instr-generate > and -fprofile-instr-use, respectively. Thank you Alexander, your help is much appreciated. For the first time I see such problem, no idea what is the core-problem. Regarding inputs, I agree, they are not only okay, they form one DEFINITIVE benchmark, anyway, glad seeing the problem resolved when using your -fdisable-rtl-bbro: Rerunning, on Zen2 4800H: D:\Wildtest_2022-Nov-29>3 D:\Wildtest_2022-Nov-29>wildtest_CLANG_14.0.1.exe WILDTEST, wildcard benchmark written by Dogan Kurt (dogan.k...@dodobyte.com), modified by Kaze (twitter.com/Sanmayce), revision 2022-Nov-29. Dogan Kurt's 'Antimalware', 2016, Iterative (wild_iterative): 70.441000 s Dogan Kurt's 'Antimalware', 2016, Iterative Optimised (wild_iterative_opt): 61.099000 s Tcheburaschka_r3, 2022, (Tcheburaschka_Wildcard_Iterative_Kaze_CaseSensitive): 72.853000 s JackHandy_Iterative, 2005, (IterativeWildcards): 80.415000 s Kirk J. Krauss, 2014, DrDobbs (FastWildCompare): 42.279000 s Alessandro Cantatore, 2003, (szWildMatch7): 97.681000 s Nondeterministic Finite Automaton (wild_nfa): 162.023000 s D:\Wildtest_2022-Nov-29>wildtest_Intel_19.0.exe WILDTEST, wildcard benchmark written by Dogan Kurt (dogan.k...@dodobyte.com), modified by Kaze (twitter.com/Sanmayce), revision 2022-Nov-29. Dogan Kurt's 'Antimalware', 2016, Iterative (wild_iterative): 98.859000 s Dogan Kurt's 'Antimalware', 2016, Iterative Optimised (wild_iterative_opt): 72.175000 s Tcheburaschka_r3, 2022, (Tcheburaschka_Wildcard_Iterative_Kaze_CaseSensitive): 70.278000 s JackHandy_Iterative, 2005, (IterativeWildcards): 107.693000 s Kirk J. Krauss, 2014, DrDobbs (FastWildCompare): 45.988000 s Alessandro Cantatore, 2003, (szWildMatch7): 79.309000 s Nondeterministic Finite Automaton (wild_nfa): 170.198000 s D:\Wildtest_2022-Nov-29>wildtest_GCC_11.3.0.exe WILDTEST, wildcard benchmark written by Dogan Kurt (dogan.k...@dodobyte.com), modified by Kaze (twitter.com/Sanmayce), revision 2022-Nov-29. Dogan Kurt's 'Antimalware', 2016, Iterative (wild_iterative): 113.664000 s Dogan Kurt's 'Antimalware', 2016, Iterative Optimised (wild_iterative_opt): 72.522000 s Tcheburaschka_r3, 2022, (Tcheburaschka_Wildcard_Iterative_Kaze_CaseSensitive): 78.138000 s JackHandy_Iterative, 2005, (IterativeWildcards): 83.266000 s Kirk J. Krauss, 2014, DrDobbs (FastWildCompare): 56.878000 s Alessandro Cantatore, 2003, (szWildMatch7): 119.861000 s Nondeterministic Finite Automaton (wild_nfa): 176.290000 s Hope future GCCs to fix this, the Tcheburaschka function is good in my view.
