thanks! i got another update on the way, actually. I didn't realize until now that mutations in this bench happen at the db level (rows) rather than cell level (tds). A simple optimization of oldDb === newDb yeilds 100% perf boost at 1% mutations, maybe 7% improvement at 50% and -5% regression at 100% (still investigating).
the resulting code is nearly identical by splitting off the row/db tpl into a sub-view and comparing the dbs (not at all esoteric). would this be considered an "optimized version" or should i just replace the main impl and maybe leave the old code in a large comment?
@mathieuancelin
thanks! i got another update on the way, actually. I didn't realize until now that mutations in this bench happen at the db level (rows) rather than cell level (tds). A simple optimization of
oldDb === newDb
yeilds 100% perf boost at 1% mutations, maybe 7% improvement at 50% and -5% regression at 100% (still investigating).the resulting code is nearly identical by splitting off the row/db tpl into a sub-view and comparing the dbs (not at all esoteric). would this be considered an "optimized version" or should i just replace the main impl and maybe leave the old code in a large comment?