individual operations extremely slow compared to environments

[kforner]

(empty) hashmap creation: ~ 80x slower

microbenchmark(new.env(), hashmap(numeric(0), numeric(0)))
# Unit: nanoseconds
#                             expr   min      lq      mean  median      uq    max neval cld
#                        new.env()   640   803.5   1245.52  1182.5  1310.5  14147   100  a 
#  hashmap(numeric(0), numeric(0)) 93833 95027.0 101540.09 95987.5 98018.5 258528   100   b

inserting one element: ~ 10x slower

microbenchmark(assign("toto", TRUE, envir = env), hm$insert("toto", TRUE))
# Unit: nanoseconds
#                               expr  min     lq    mean median     uq   max neval cld
#  assign("toto", TRUE, envir = env)  514  622.5  743.01  713.0  800.5  2406   100  a 
#            hm$insert("toto", TRUE) 7512 7765.0 8881.30 7860.5 8039.5 93803   100   b

I understand (from the benchmarks) that you are focusing on bulk operations, but still considering the title "The Faster Hash Map" I was quite disappointed.