blob: cd9e6c3826faa59e85edc48bf79f0b49909d6ee0 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
|
prelude: |
# Measure uncontended GVL performance via read/write with 1:1 threading
# If we switch to M:N threading, this will benchmark something else...
benchmark:
io_pipe_rw: |
r, w = IO.pipe
src = '0'.freeze
dst = String.new
i = 0
while i < 1_000_000
i += 1
w.write(src)
r.read(1, dst)
end
w.close
r.close
loop_count: 1
|
