Quote:
Originally Posted by ticenl
I just bought and installed PT4 on a relatively old computer (but it does have an ssd) and I noticed that its performance with for example making a graph is way slower.
An old computer with an SSD is still an old slower computer, the SSD cannot compensate for that. Processor speed, RAM, OS Version, and PostgreSQL tuning are the three other major factors that determine speed of PT4, all four work in conjunction with one another. Additionally we have seen SSDs which are no faster than a standard hard drive due to improper configuration.. it is not a guarantee of speed.
Now lets tackle your issue using your existing computer without any hardware changes:
1) It depends on the graph, and what you are graphing.
2) Have you rebuilt your cache? Often this is all that needs to be done, the cache needs to be rebuilt anytime a stat is used in a database query after that stat has been updated by our developers. Custom stats are uncached, this will also affect graphing performance (again, this all depends on what you are graphing).
Quote:
Originally Posted by powder_8s
Is this a known fact? Pt4 is slower than HM? I did get some warning by the way that the fact that I'm using Postgresql 9.1.4 and not 9.0 could cause reliability and performance issues. Anyone know if this is really the case and will my pt4 be clearly faster if I downgrade my postgresql?
We don't support PostgreSQL 9.1 or 9.2, these versions have known issues that will not only hurt performance, they could result in other data or crashing issues (a polite way of saying PT4 will crash is by calling it a reliability issue, but it still means PT4 may crash). 9.1 has to be replaced by 9.0 or 8.4 ... we cannot even provide support if 9.1 is installed because we do not design PT4 to work with it.
As for the speed of PT4 vs Hm2, our two companies code our applications in very different ways. We use a pure PostgreSQL compliant solution while integrating enumerated all-in equity calculations, ICM, and tournament detection along with NoteTracker autonote processing as each hand is imported. HM2 on the other hand uses a flat file "bit bucket" approach to import, this is an intermediary stage before storing into the database, the positive of this approach is that import speeds are faster, the negative to this approach is increased risk of data damage and additional "moving parts" that could be the source of development problems; additionally they do post-processing of some data points. Both methods have their pluses and minuses, but we are confident we made the right choice because in general our solution has resulted in fewer crashes and very rarely has been the source of database problems.