mohammad1722 has quit [Quit: Ping timeout (120 seconds)]
cart_ has quit [Ping timeout: 260 seconds]
cbmuser has quit [Ping timeout: 246 seconds]
Guest55 has joined #openscad
Guest55 has quit [Client Quit]
cbmuser has joined #openscad
mmu_man has quit [Ping timeout: 256 seconds]
<peepsalot>
pca006132: iirc, it just throws out the largest items first.
<peepsalot>
i was thinking it would be neat to have a cache metric of: (memory size) / (time to compute). so the first things to remove would be the ones that are easiest to replace by re-computing. but would require that all entries have their the time tracked before putting into cache
Guest39 has joined #openscad
Guest39 has quit [Client Quit]
mohammad1722 has joined #openscad
R2robot has quit [Ping timeout: 240 seconds]
<peepsalot>
and its not immediately clear how to measure time to compute. e.g. for geometries that are result of nested operations on other cached objects, would you include accumulate the compute time of those objects in the output?
mohammad1722 has quit [Client Quit]
R2robot has joined #openscad
<peepsalot>
on the other hand, the speedups from manifold are so great, that the whole reason for the cache's existence is nearly moot. it was useful for glacially slow CGAL Nef ops, but seems like it may be just getting in the way for manifold ops
<Scopeuk>
I guess that gets even more interesting for compound objects of is the compounding more expensive or the raw generation, it may be more cost effective to throw away the result or the final op and keep precursors
<Scopeuk>
would be an interesting thing to play with and generate metrics, but equally modern machines do have so much disk space a user could just hand over 1GB and never think about it again
<peepsalot>
\
* peepsalot
stops trying to clean around his enter key
J24k25 has quit [Quit: Client closed]
J24k25 has joined #openscad
omegatron has quit [Quit: Power is a curious thing. It can be contained, hidden, locked away, and yet it always breaks free.]
mmu_man has joined #openscad
R2robot has quit [Read error: Connection reset by peer]