They are in your folder (or at least should be). If you use the updater with "force update" they should be there. If they are not drop us an email to support@piosolver.com and I will send them to you.

It assumes check-down postflop.

No, it calculates it on all possible flops.

Yes.

Full preflop contains postflop play and doesn't just assume check-down postflop.

I did small saves, but when I load them im able to see river ranges. what does the small save mean then? they are 20 kb files

Rivers are recalculated on the go; they just take a very short time to calculate

As tobakudan said, they are recalculated on the fly and the process is usually barely noticeable. I talk about small/micro saves and how they work here:

https://www.youtube.com/watch?v=VhG_YsEvfwU (question number 5)

Sorry this will seem a really simple question:

in the product description for pro it says

"Uses up to 12 CPU cores (12 hardware threads)"

When CPU's are advertised they usually use the format 8c/16t or 12c/24t

Does this mean that on the 8c/16t chip pio will be able to use all 8 cores and 12 of the 16 threads? and for the 12c/24t it will use all the cores and 12 of the threads?

When doing multiple files Aggregation analysis:

Can I an do river aggregation analysis on small saves? (i.e. Will it recalculate the rivers?)

hi punter, i have rented an Intel Xeon E5 2xE5-2630v3 16c/32t 2,4/3,2GHz server.
So it has 2 cpus with 16 cores each. I have Pio Pro, how do i get maximum out of this server?

I was thinking that, given the pio pro is capped at 12 cores per instance, I should use 3 instances of pio at once. If yes, should i change the number of threads that its using - from the default "0" to "10" in order to have a fair distribution? - or just leave it like it is and let the system split the resources in 3?

The 2nd number means that the cores can do hyperthreading which usually provides from 0% to 20% speed-up (depending on application). On 6 core CPU (like 5820k) Pio pro is going to use 6 cores + hyperthreading. On 12 core CPu it will use all 12 cores but without hyperthreading. The difference for hyperthreading for Pio is around 15%.

Yes but it will take a very long time (because it will need to recalculate all rivers). It's good idea to set recalc accuracy for rivers to something bigger than default if you want to follow up with that.

It will split them more or less. I don't the answer to your question as it's likely depends on the machine. Running three instances with something like 12 threads each should be good so you don't need to change anything.

Hey,

I'm putting together a pc for Pio use and I have a few questions. It was mentioned in this thread several times that the 4790k and 5820k are the best value for average Pio use.

The price difference where I'm buying is $50-$60 for those two. But getting the 5820k also means I need to get a x99 mobo, DDR4 RAM, more expensive PSU etc. since I want to build a pc that will be durable for the workload. So with all the different components that are needed for the 5820k I'm looking at around $600-$1000 difference between the builds depending on which components I get and if I'll use my old GPU for example.

Is the difference worth it if I'm only going to use Pio for postflop?

If not, is the difference worth it if we take into account future proofing? As new things are added to Pio and it is optimized etc. is having the ability to have DDR4 much better? Also, how much do different memory frequencies affect Pio?

If I decide to solve preflop spots as well does it then make more sense to get the 4790k build and use the rest of money to get the Edge version (I'm thinking of getting the Pro version first) and rent a server for pre flop like it was mentioned here?

Or is the 5820k still a better option even if I do decide to rent a server? And I could solve some pre flop situations on that build as well without renting the server but that would make that build even more expensive because of more DD4 RAM. I assume solving pre flop is not an option on 4790k because of the RAM being capped at 32gb.

They are likely the best performance/cost at this point although 5820k is going to be faster, especially if you overclock.

That's a question only you can answer. Bare 5820k is going to be around 1.25x faster but you can easily overclock it for additional ~20% which would result in about x1.5x performance. The difference between DDR3 and DDR4 should be negligent (at least it was I benchmarked it some time ago).

I don't know sadly. My best guess is that RAM frequency won't make much difference in the future (but get at least 1600Mhz).

I am a fan of renting instead of investing in own hardware but again it's a very individual decision. With preflop it's a bit more clearcut at you would need to get at least 64GB of RAM and even then you may sometime want to run bigger trees. You may rent something for a few days or a month to do it but building such a computer at home is quite an investment.

It isn't at the moment (with the exception of very simple spots in 6max like steal vs BB with one sizing). You can get even 128GB of RAM with 5820k but again, RAM is still very expensive and renting is likely to get cheaper.

Thanks a bunch for the answer. Got all the info I needed.

Sorry to come back to this so late...

If I don't want it to ever time out, do I just leave out the "go Y seconds" line?

If I want a full save, do I just leave out the "no_rivers" clause?

Also, when you say "at the top" and "at the bottom", I don't think you mean exactly that. So just to be clear, I include

set_threads 0
set_accuracy X

right after #EndFlop, correct? And I'm really not sure where to input

go Y seconds
wait_for_solver
dump_tree "C:\SAMPLE_PATH\yourfilename.cfr" no_rivers

Does it really go after the last line, "add_all_flops"?

Yes, just use "go"

Yes, or use "no_turns" to get preflop + flop only. Not using any 2nd argument will result in a full save.

All the lines which start with "#" are comments from the solver's perspective (they are ignored). They are there so the viewer can read the config from the script (from Viewer's perspective all lines not starting with "#" are comments). It doesn't matter where lines starting with "#" are.

Yes. Again if you are learning the scripting language of the solver it's a good idea to just type it and run it to see if it works (preferably on very simple trees and big accuracy first so you don't waste time). You can ask questions about it here but it will be faster if you just test it

Just tried scripting for preflop. When I loaded 2 scripts individually they worked fine, but when I put them in a nested script and ran it, both got the error:

ERROR: Scripts/Preflop/PreflopScript1 couldn't open file
ERROR: Scripts/Preflop/PreflopScript2 couldn't open file

Since they both work when opened individually, it doesn't seem like the problem is with the script itself...any idea what's going on?

Well, those things are very hard to debug remotely without seeing exactly what's in the script (it might be a typo, it might be incorrect path, it might be something else) and what exactly you are doing to start them.
The solver tries to find a file when it encounters load_script command so it's likely it can't find it. A good start would be to put them all in the same folder (the one the solver is located in) and seeing if it works then. If it does then you know the problem is in the file path. Maybe it's the best idea to provide full absolute path (like "c:\mysolver\myscripts\test1.txt") instead of relative ones. Remember to put paths/filenames in quotes ("C:\mypath") as well as it's needed if there are spaces there.

Hi,

How do I get this grid to show the equities of Raise/Bet/Check like in the quick start video?
Clicking Strategy + EV or Strategy won't display it either.

Edit: It shows the equities on grid for ROOT but no decisions after like in the quick start video, following it exactly. Ty

If you mean the averages displayed on 13x13 grid then the reason they are not displayed is that there is not enough space for minimum size font. You can try resizing the window (make it bigger) as well as moving vertical bar between 13x13 area and the buttons to the right.
I show it on the video, here:

https://www.youtube.com/watch?v=5Ozjfnatwms (question 11, see the video description)

Hi, in the image you linked me too, I don't have anything listed under 'Preflop Treebuilding'. Its blank. I am up to date with the current version and connected to 'PIOSOLVER - pro19.exe. PIOSOLVER-pro 1.9.2.' Can you advise me what to do next.

Thanks!

Does the rebuild_forgotten_streets command not work for preflop saves? I tried it on a very small preflop save and got: "ERROR: rebuild_forgotten_streets missing/incorrect tree"

It was my mistake not to ship sample preflop configs for basic/pro version. You can download PreflopTreeBuilding folder here:
https://www.dropbox.com/sh/4hk63ccnt...89_aB2CVa?dl=0

Put it in your Pio folder and the sample trees are going to appear.

Yes, it doesn't work for preflop at this point. There would be no benefits over building and solving the tree from scratch.

Well, it would save a lot of time and/or space

It wouldn't save space (rebuilding means you get the whole tree in memory again + overhead). It wouldn't save time either as more than 99% of the solution is on turn+. Solving from scratch is not slower.

What I said was assuming there would be times where you would want to do preflop calculations to a certain accuracy and then later continue it to a higher accuracy. Right now full saves are required for this (save space), or if you save smaller you have to start all over again (save time).

What I meant is that solving from scratch is not slower.