On Tue, Apr 6, 2021 at 7:54 PM Matthew Flatt <mfl...@cs.utah.edu> wrote:
> In the cross-compilation path that the release builds use, there was an > unintended layer of compression. (I noticed the size difference before, > but had not yet investigated.) Checking that again, if the extra layer > is used, the 10% or so savings in file size causes a 5% decompression > slowdown for loading. Compressing just the code, meanwhile, makes load > times faster on machines that I tried. So, I think the current > (non-cross) default is still the best choice for most situations, and > I'll adjust cross-compilation to match. > … > Compressing embedded boot files would save 30 MB, which would make a > big difference in your case. Compressed boot files take 20-40ms longer > to load on a typical machine, which is significant relative to the > minimal load time, but it's very little time in many situations. > I don't have a problem with the current setup, but I've noticed that other projects (e.g. OpenZFS) have been adding support for Zstd compression, which has an especially wide range of options for trading off compression ratio and (de)compression speed. I don't know if it would be an improvement here—apparently it is both faster and more efficient than gzip but doesn't quite displace LZ4 from its core niche—but it's another direction that could be explored. -Philip -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/racket-users/01000178ae65a584-6a7d1479-38e4-45c3-88f3-1e8b9700b6c1-000000%40email.amazonses.com.
