KPJ <kpj@sics.se> writes:
> It appears as if Anders Sandberg <asa@nada.kth.se> wrote:
> |
> |I have been thinking a bit about computation backup for distributed
> |uploads. If the left hemisphere computer goes down, then the right
> |hemisphere might wait until another left comes online, and then either
> |synchronise with it. This might of course mean that during crashes you
> |get "half-thoughts" that never get anywhere, but they would not be
> |consciously observable.
>
> I believe this would compare to the situation when a wetware processor goes
> into sleep mode while the other stays in normal mode. According to local
> emprical data, this does give some unusual indata to the processor which
> is not in sleep mode (very fast projected movies, as it were), but no
> "half-thoughts" could be detected.
Detecting a crash can likely be done within one timeslice of the simulation (processor A calculates the new state of the right hemisphere, waits for processor B, waits, waits, times out and tries to start up a new left hemisphere simulation based on stored data), so the right hemisphere will just experience things up to the end of the timeslice, which likely is extremely short. If it uses delayed evaluation and other tricks, things can get stranger.
-- ----------------------------------------------------------------------- Anders Sandberg Towards Ascension! asa@nada.kth.se http://www.nada.kth.se/~asa/ GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y