-
@ Denis Warburton
2025-02-26 21:06:27
( To be fair, the last time I had one of these clusters running, was back in the ‘00s.
The “killer problem” back then was - and probably still is - how the computers share the source file to the other computers on the network.
In the ‘00s, Compressor would copy *the entire file* to every machine in the cluster before starting processing. Which, as you might imagine, would … take … some … time.
It literally took longer to do that step, then process most of the file on a single machine. )