No, not for sure, you would know?
I just inferred from no speedup.
With a multi-threaded Julia program (Fasta), this time around, which actually gives multi-threaded speedup, I do not see snmalloc help (actually a bit slower) getting same speed as without LD_PRELOAD, unlike for a C program (binary-tree) where snmalloc gives 2.8x speedup.
$ time LD_PRELOAD=/home/pharaldsson_sym/snmalloc-master/build/libsnmallocshim.so ./binarytrees.gcc-5.gcc_run 20
stretch tree of depth 21 check: 4194303
[this time a round, no stop here, when not compiled with debug mode.]
real 0m1,036s # without LD_PRELOAD: 0m2,941s
user 0m5,568s
sys 0m0,108s