Could Titan use GPUs to speed up encoding/decoding messages?
Submitted by Kristof Szabados
It might be worth a try to see if Titan could use GPUs to speed up the encoding/decoding of messages (at least for some encodings and some protocols).
Most of the time, the actual real-life encoding/decoding needed to be done is not very complicated. And for many messages or message parts, a GPU based implementation might be able to do the work much faster. For example encode all elements of a record of integer at the same time, so that they just need to be chained together to finis the coding work.