FOLLOWUP:

Funny, i ran benches for elements up to 6000 and it seems /set becomes considerably slower than /write

Code:
Test results for hadd,setv,file tests:
 
size:  1000	 2000	 3000	
-----------------------------------------------------
hadd:  0.047	 0.078	 0.125	
setv:  0.094	 0.593	 1.984	
file:  0.453	 1.016	 1.39	
fopen:  0.047	 0.109	 0.172	
 
Test results for hadd,setv,file tests:
 
size:  1000	 2000	 3000	
-----------------------------------------------------
hadd:  0.031	 0.078	 0.187	
setv:  0.078	 0.609	 2.016	
file:  0.453	 0.969	 1.625	
fopen:  0.047	 0.094	 0.156	
 
Test results for hadd,setv,file tests:
 
size:  6000	
-----------------------------------------------------
hadd:  0.469	
setv:  9.766	
file:  3.079	
fopen:  0.375
 
Test results for hadd,setv,file tests:
 
size:  6000	
-----------------------------------------------------
hadd:  0.375	
setv:  9.672	
file:  3	
fopen:  0.282


Notice that i added in code for /fopen testing, tests seem to show that it's actually faster than hashes at high element counts... weird?

Code:
_fopen {
  if ($fopen(FILE)) .fclose FILE
  .fopen -no FILE test.txt
  var %x = 0
  var %ms = $ticks 
  while (%x < $$1) {
    .fwrite -n FILE 0
    inc %x
  }
  %x = $calc(($ticks - %ms) / 1000) 
  write -c test.txt
  .fclose FILE
  return %x
}

- argv[0] on EFnet #mIRC
- "Life is a pointer to an integer without a cast"