2022-11-30 10:21:47 <snow_richard> hi
2022-11-30 10:22:02 <snow_richard> do i need to register a nick here
2022-11-30 10:36:45 <olle        > No?
2022-11-30 12:33:51 <Zarutian_iPad> interesting way of taking a few conditional branches and combine them into one inderect branch
2022-11-30 12:34:07 <Zarutian_iPad> indirect*
2022-11-30 12:35:10 <Zarutian_iPad> take the conditional and have their boolean outputs be mapped to spefic bits in a cell-value
2022-11-30 12:36:16 <Zarutian_iPad> requires that the conditionals are not data dependant on the effects of the consequent or alternative blocks of each other
2022-11-30 12:36:40 <Zarutian_iPad> use that cell-value as an index into a jump table
2022-11-30 12:37:20 <Zarutian_iPad> that is basically it
2022-11-30 17:14:39 <KipIngram   > That sounds like it ought to work fine.
2022-11-30 17:15:36 <KipIngram   > and yeah - it would let you compress a bunch of control transfers into one, which can have big payoff on complex pipelines.
2022-11-30 17:16:35 <KipIngram   > If your table was sparse there'd be a space efficiency to consider.  There might be numerous combinations of test results that would never happen together.
2022-11-30 17:27:16 <KipIngram   > Kind of strikes me as an implementation of the "Calculate, don't choose" principle.
2022-11-30 17:35:37 <KipIngram   > snow_richard: Not for this channel, I don't think.
2022-11-30 17:42:33 <thrig       > they gone done left
2022-11-30 17:57:55 <KipIngram   > Ooops.