On Wednesday, May 28, 2014 10:17:20 AM UTC+10, Avik Sengupta wrote: > > While this is an old thread, I'm not sure if a K/Q benchmark is all that > relevant to Julia. I've only seen K/Q code written in conjunction with > KDB (which is an in memory columnar database). So its not just > benchmarking a language. Also given the specialised nature of KDB, I > imagine the answer will very much depend on the workload. >
As with any benchmark yes, but you could benchmark the language without any columnar database elements. In fact the latter in kdb not that prescriptive, and you are left to assemble the pieces you really want. It's not like writing a benchmark for an SQL database or anything where you're forced to use a large portion of the infrastructure. Tangentially, In regard to some of the earlier comments on this thread > about this being a properitary solution, there is a free version available > for download. However, its license disallows benchmarking. > Yes, that does limit the possibilities.. disappointing license provision :)
