How Misaligning Data Can Increase Performance 12x by Reducing Cache Misses via @FransBouma http://danluu.com/3c-conflict/
-
-
Replying to @kellabyte
@kellabyte@FransBouma does you know why the cache probing function isn't something less trivial? Seems like it should be possible…1 reply 0 retweets 0 likes -
Replying to @loganb
@loganb@kellabyte@FransBouma Cache lookups are often on the critical path; extra complexity may reduce clock speed or add a pipe stage.1 reply 0 retweets 0 likes -
Replying to @danluu
@danluu@kellabyte@FransBouma I guess, I'm just surprised there's not enough wiggle room to stuff in a XOR'ing w/the higher bits…1 reply 0 retweets 0 likes -
Replying to @loganb
@loganb@kellabyte@FransBouma My guess is that the effect is negligible for 8-way. See badly reproduced diagram http://web.cs.dal.ca/~mheywood/CSCI3121/Memory/02-MissRateReduction.pdf …1 reply 0 retweets 0 likes
@loganb @kellabyte @FransBouma When I have some time I'll generate a better graph with http://www.simplescalar.com/ .
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.