Optimizing Rust Binaries: Observation of Musl versus Glibc and Jemalloc versus System Alloc

community
7

Here are the results for my uncbv project:

Jemalloc + Glibc

time ./target/release/uncbv x tests/others/MegaDatabase2016.cbv -o ~/CBVTests/ --no-confirm
75.49user 3.26system 1:24.75elapsed 92%CPU (0avgtext+0avgdata 971064maxresident)k
0inputs+3993304outputs (0major+15854minor)pagefaults 0swaps

System Alloc + Glibc

time ./target/release/uncbv x tests/others/MegaDatabase2016.cbv -o ~/CBVTests/ --no-confirm
80.45user 2.90system 1:29.13elapsed 93%CPU (0avgtext+0avgdata 971580maxresident)k
0inputs+3993304outputs (0major+19533minor)pagefaults 0swaps

Jemalloc + Musl

time ./target/x86_64-unknown-linux-musl/release/uncbv x tests/others/MegaDatabase2016.cbv -o ~/CBVTests/ --no-confirm
77.63user 2.62system 1:25.63elapsed 93%CPU (0avgtext+0avgdata 967984maxresident)k
0inputs+3993304outputs (0major+15783minor)pagefaults 0swaps

System Alloc + Musl

time ./target/x86_64-unknown-linux-musl/release/uncbv x tests/others/MegaDatabase2016.cbv -o ~/CBVTests/ --no-confirm
84.84user 3.16system 1:32.07elapsed 95%CPU (0avgtext+0avgdata 970204maxresident)k
0inputs+3993304outputs (0major+311665minor)pagefaults 0swaps

It seems the default is the best in my case.

1 Like