It looks as if raw 'nc' works not too badly, with a 40GByte file I get
oak$ time nc -l 172.26.128.2 7777 > /scratch/frogface real 1m59.020s user 0m1.813s sys 1m32.519s (obviously there is a bit of a pause here between pressing enter on the listening machine and on the sending machine)
---
pineapple@pineapple:/scratch/L5895B$ time cat msieve.dat | nc 172.26.128.2 7777 real 1m58.149s user 0m6.762s sys 1m52.296s
so that's forty gigabytes in two minutes, with some sort of speed-of-light set by pineapple@pineapple:/scratch/L5895B$ time md5sum msieve.dat e88bcd4df8064f250cac50be2d79adfc msieve.dat
real 1m13.308s user 1m3.819s sys 0m9.488s oak@oak:/home/nfsworld/fibonacci/F2157/e/s2$ time md5sum /scratch/frogface e88bcd4df8064f250cac50be2d79adfc /scratch/frogface
real 1m53.066s user 1m36.066s sys 0m9.336s
oak@oak:/home/nfsworld/fibonacci/F2157/e/s2$ time cat /scratch/frogface | wc -c note deliberate use of cat, otherwise wc -c reduces to a single stat call 41629616378
no subject
Date: 2019-01-12 01:55 pm (UTC)oak$ time nc -l 172.26.128.2 7777 > /scratch/frogface
real 1m59.020s
user 0m1.813s
sys 1m32.519s
(obviously there is a bit of a pause here between pressing enter on the listening machine and on the sending machine)
---
pineapple@pineapple:/scratch/L5895B$ time cat msieve.dat | nc 172.26.128.2 7777
real 1m58.149s
user 0m6.762s
sys 1m52.296s
so that's forty gigabytes in two minutes, with some sort of speed-of-light set by
pineapple@pineapple:/scratch/L5895B$ time md5sum msieve.dat
e88bcd4df8064f250cac50be2d79adfc msieve.dat
real 1m13.308s
user 1m3.819s
sys 0m9.488s
oak@oak:/home/nfsworld/fibonacci/F2157/e/s2$ time md5sum /scratch/frogface
e88bcd4df8064f250cac50be2d79adfc /scratch/frogface
real 1m53.066s
user 1m36.066s
sys 0m9.336s
oak@oak:/home/nfsworld/fibonacci/F2157/e/s2$ time cat /scratch/frogface | wc -c
note deliberate use of cat, otherwise wc -c reduces to a single stat call
41629616378
real 0m25.615s
user 0m0.571s
sys 0m24.695s