On 21/11/2012, at 8:05 AM, Alan Silverstein wrote: >> Could you get a slight improvement by making a private version of >> JudySLGet() that returns the location of Index's branch for subsequent >> use in seeding the call to JudySLNext()? Or does all this happen in >> the cache quickly enough not to make a difference? > > That last point was the theory based on Doug's measurements... No point > in keeping seeds, once it's in the cache it's basically free to revisit.
However that only applies internally. The argument fails if the client is scanning sequentially because there is intervening code. It also doesn't apply if the CPU is shared my multiple threads which could pre-empt and spill the cache at anytime. > Today -- who knows what changed? Probably not much on x86 like devices. What's changed is that the ARM is now the most common device, and people care about performance not because of speed, but because of battery drain. -- john skaller [email protected] http://felix-lang.org ------------------------------------------------------------------------------ Monitor your physical, virtual and cloud infrastructure from a single web console. Get in-depth insight into apps, servers, databases, vmware, SAP, cloud infrastructure, etc. Download 30-day Free Trial. Pricing starts from $795 for 25 servers or applications! http://p.sf.net/sfu/zoho_dev2dev_nov _______________________________________________ Judy-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/judy-devel
