On Sat, 2022-05-07 at 15:18 -0700, Amnon wrote: > The other interesting question is what algorithm we use to find the > pattern in each line. > Generally bytes.Contains uses Rabin-Karp. But as the pattern is the > word "test" which is only 4 bytes long, > a brute force search is used, using SSE type instructions where > available. > So the naive Go approach will give you a very fast execution. The > main thing is to set up your scanner with a large buffer, to minimize > the number > of file system reads, and to avoid the newbie error of working with > strings rather than []byte, and forcing the code to do vast numbers > of > unnecessary and expensive allocations.
There is an interesting post from the author of grep that goes over some of these details https://lists.freebsd.org/pipermail/freebsd-current/2010-August/019310.html . -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/27eef405ac38a8e32488f15ee7726c3478866ede.camel%40kortschak.io.
