Hi Glenn,
Computers today are very fast and they keep getting faster.
However there are certain bottlenecks that can cause significant delays. One such bottleneck is accessing data over the Internet. Until these data are returned, the software can't really do anything. Compared to local data access, remote data access is EXTREMELY slow (the interesting thing is that "slow" is relative. 15 years ago if you told someone you could analyze 8800 stocks in a few hours, they'd have thought that was fast).
By implementing some algorithms to minimize the number of network data requests and using caching techniques, more of the processing can be done locally and that's where the vast majority of the performance improvement comes from. A small percentage comes from optimizing the code.
As it turns out there are other techniques that can be used to further increase performance. Currently the software uses a synchronous pattern (i.e. it has to wait for a response before proceeding). Changing to an asynchronous pattern and taking advantage of today's multi-core processors, using threads, would further increase performance. However processing 8800+ stocks in 1 minute 20 seconds is fast enough so that this change is not currently necessary. Once foreign support is added (e.g. Canadian and European stocks) and the number of tickers increase dramatically, then I'll most likely implement the asynchronous pattern.
PI 3.0 will be subscription-based but it is a superset of VSS. Basically it includes everything VSS includes (i.e. the fundamental analyzer and the asset allocator) plus a portfolio manager (similar to AI's), a classic rebalancing engine, News reader, historical analyzer (that supports walk-forward testing) and a new short-term rebalancing function that I've been working on for the past year.
AI
