HM2 and your competitor performance

[holdemstats]

Hello there, the last update from your most direct competitor brought an incredible speed boost to the HUD and the program overall even on large databases.
I hope that you are aware of this and do something to improve HM2 performance.
I'm a HM user myself, and want to continue to be, but against facts there's no arguments. I suggest HM2 staff to try out their last update and try to learn something from what they did.

[DogNamedBluff]

Thank you for the feedback. I have forwarded this information to the proper personnel.

[morny]

Thanks, competition is always good for the consumer and keeps everyone on their toes. The last time we tested theirs on a large database it was unusable so we couldn't do any comparisons, giving the method we use to store a lot of the information in memory i expect were still way ahead of the competition (and also HM1) when it comes to large database performance and speeds but will be interesting to compare. I just ran the overall report on a hero with 1.5 million hands on a high spec PC and it took about 10 secs to complete the report which is pretty impressive.

As to plans to improve performance we have been undergoing several internal projects to improve on this. One is for RAM usage, although we think that people should make use of their RAM to increase speed as their is no point in having 8gb of Ram and only ever using 500mb, the downsides of this is obviously higher CPU usage, slower initial load times and people with lower performance PCs have more issues but the upside is because we store most of the information in memory you can access it substantially quicker and thus reports will be much much faster, we took peoples feedback about this into consideration and were implementing an option to use less ram which will be less performance intensive, quicker initial load time but the sacrifice is slower report speeds. The choice will be yours, both methods have big advantages and disadvantages so you will have the choice

The other areas were working on is with CPU usage while playing, there are a couple of areas where the CPU usage can spike for some people and were looking into solving this, its less evident on higher spec machines and not really noticeable when you only play a couple of tables but as you increase the amount of tables you play or if you use a lower spec machine it can become an issue, so those people should see a big benefit there.

Finally were changing our reports grid to a more lightweight version which should deal with a bunch of bugs and issues with reports, we were using the standard one but it was a bit heavy and cluncky and we decided to code our own, all these projects were pretty big projects but we expect to start rolling them out in beta testing over the coming weeks.

[holdemstats]

Thanks very much for the kind and compreehensive response, I had by no means intention to bash on HM staff, it was just a constructive alarm. I like HM2 overall, and beside some tweaks, if the performance issue gets through we definitily have a winner so keep up, and hope to see some improvements in the near future.

[rowhousepd]

First off, Morny (and the rest of the HM2 gang), thank you! Seriously, this means a lot to folks like me who loved HM1 and want to love HM2 and who believe it really can be a supurb piece of software that easily rivals a certain unnamed competitor. I personally really appreciate you guys not only addressing the issue behind the scenes but also coming out and saying (or at least implying here) that there really are some performance issues w/ this release and that tweaks here & there aren't going to cut it for people w/ non-super computers.

One is for RAM usage, although we think that people should make use of their RAM to increase speed as their is no point in having 8gb of Ram and only ever using 500mb, the downsides of this is obviously higher CPU usage, slower initial load times and people with lower performance PCs have more issues but the upside is because we store most of the information in memory you can access it substantially quicker and thus reports will be much much faster

I'm a bit confused. Do you mean this is something the new version will offer as an option for us to select in the software itself, or that we will have to adjust the way our machines run via the settings in windows?

Thanks very much for the kind and compreehensive response, I had by no means intention to bash on HM staff, it was just a constructive alarm. I like HM2 overall, and beside some tweaks, if the performance issue gets through we definitily have a winner so keep up, and hope to see some improvements in the near future.

+1 Thanks again HM folks!

[Patvs]

By editing the holdemmanager.config file, you'll be able to change a value:
how many hands are cached for each GB available RAM.


This gives you the option to:
-1 cache a LOT of hands, where HoldemManager will use a LOT of RAM. Startup time will take longer (caching those hands to RAM). Then running a report will be faster.
-2 NOT cashing a lot of hands. Startup time will be a LOT faster. Running a report can be slower.
So it will greatly depend of the size of your database (number of hands) / available RAM, if you can benefit from the hybrid build.

[rowhousepd]

By editing the holdemmanager.config file, you'll be able to change a value:
how many hands are cached for each GB available RAM.

Sorry if I'm being dense here, but are you saying we can do that now w/ the current version (in which case I guess there is something that in fact I didn't try yet), or is this an option that will be available only w/ the hybrid?

[Patvs]

This will be working in the upcoming hybrid build.
The line already exists in the current .config file, but changing the value has no influence with the current non-hybrid build.

Key Name="MaxHandsCachedPerGBofRAM" Value="500000" /

[rowhousepd]

Gotcha. Thanks for the clarification.

[Corey]

HEM2: I have 16 postgres.exes running

HEM1: I have 4 or 5 postgres.exes running

What is wrong with this picture?