In the previous article on 3D point cloud for FMX, systematic memory copy when drawing primitives came out as #1 bottleneck in FMX performance when rendering millions of points.
But was it just a false positive ?
I hacked FMXU.Context.DX11 to eliminate the memory copies… in the most hacky way, don’t ask for code it’s not re-useable 🙂 (more…)
Recent commits to the DWScript repository doubled the compiler performance when compiling many small scripts, like happens in the unit tests suites.
This started from a first profiling run where the memory allocations around the UnicodeLowerCase function came out as top bottlenecks.
Thing is, Pascal being a case-insensitive language, there are lots of case-insensitive comparisons, lookups, searches and hashes, and turns out a key hash code was computed with code like
SQLite offers through its JSON1 extension a lot of capability. Generating JSON outputs from SQL, even complex structured, multi-level JSON is very straightforward.
But what about the opposite use case ? When you have store data as JSON in an SQLite database, and you want to search or filter it ? This is a short exploration of what to expect.
Hash, IndexOf, Lookup, Performance, Search, Sorted, String
As a followup to the previous String Lookup: Hash, Sorted or Unsorted here is a look at what happens for longer strings.
When the string you’re searching have more than a dozen characters, and you only have a few hundreds of them, a clear winner emerges.
When looking up a string, what is the fastest strategy?
A hash map, a sorted list or an unsorted list?
Of course it depends on how many strings you have, but where are the cutoff points?
Here is a quick test, and an interesting tidbit is uncovered…
(more…)
How fast are mobile CPUs these days? Well, they are pretty fast processors.
While some still consider that x86 is ten times faster than ARM, that is just not the case anymore these days, the difference is much smaller, and closer to just two times if you compare a regular smartphone to a regular laptop/desktop.