Hello, is it possible to say how much data will HEM write into postgre database if you play few thousands hand in a day? I am thinking about getting SSD just for database, but large amount of data written (not read) might be a problem. Thank you.
I generally recommend at least a 256gb SSD for someone that plays a lot of hands now that they are cheaper. Bigger SSD's also perform better than their smaller counterparts so if you can afford it bigger is always better when it comes to SSD. If you use NoteCaddy it will require additional space as well so you have to double (or more) the numbers below. The HM2 data numbers given below are split between the \postgresql\data folder and the \roaming\holdemmanager\database folders about 60/40.
You can put your postgres data folder and the HM2 database parent folder on a different drive but the best performance comes from having all the data on the SSD - http://hm2faq.holdemmanager.com/ques...e+or+Partition
In HM1 there are a lot of tables that deal with a specific hand played by a specific player. In HM2 we don't have these tables, instead, for player related hand data, we use a flat file approach. Each player has his own folder and each file represents a single day. Each line within the file is a tokenized version of a single hand with incredible amounts of detail. Your overall HM2 spaced used (DB + Files) is about 2/3 of HM1 and we store probably 2-3 times as much info plus it can be accessed many, many times faster and allows us to do some things that wouldn't be possible otherwise. If you don't want the space in your Roaming folder due to limited C: file space, we do give you the option of storing this data anywhere when creating the database.
General rule is 1 million hands = 10 GB.
In reality it's: 1 million hands
HM1: 6.8 GB
HM2: 4.3 GB
PT4: 13.6 GB
So with a 10 million hand database you need a SSD of at least 60 GB (Windows) + 100 GB = 160 GB. We use 1 million = 10 GB (instead of 6.8) because when you want to perform a vacuum/analyse or backup/restore a database it will require a LOT of disk space to perform such a task.