Myracle GUI

[SSTATHIS]

Hi Martin!
There is a problem with new version.
I always have check the "Simple movelist" in "appearance". The moves tab at first show me 7 lines.
After moove 9 it shows 6 lines. After move 21-22 it shows 5 lines. This continues and finaly shows only one line!

[mar]

build 152 is up:
- fix broken autoscroll in movelist

[SSTATHIS]

Really fast!!!!
Thanks Martin!

[mar]

build 153 is up:
- fix: move list sometimes offset incorrectly when switching tabs, showing nothing
- fix: file list now shows file time converted to local time as it should

[mar]

there has been some fuss about some AVs false positives for the recent build,
namely AVG, Avira and Avast on VirusTotal (note they are all owned by Gen so likely they use the same detection engine),
so I've rebuilt the Windows binaries with control flow guard enabled (binary slighly bigger now) and now instead of x false positives
I only get 1 for some dev version of Defender, shrug.

I've uploaded new zip files for Windows, but this feels like fighting windmills... because tomorrow some xyz AV may flag it again.
I'm not going to pay for a signing certificate for a freeware program (that likely wouldn't help anyway)...

[jshriver]

Thank you for Myracle, I've been using it for CCRL gambits. May I make a request? A button or option to flip the board. I enjoy watching the games and would be nice to flip view.

[mar]

Thank you for Myracle, I've been using it for CCRL gambits. May I make a request? A button or option to flip the board. I enjoy watching the games and would be nice to flip view.

I'm glad you like it. there's already a shortcut (ctrl-shift-F) to flip the board. a button would occupy precious UI space,
so I don't like the idea very much. there's also view/flip board accessible from menu, only one extra click.

if you mean gauntlets then perhaps an option to auto-flip board for gauntlet seed(s) would be a good idea,
but in general it's quite a minor thing and feels a bit ad hoc.

I consider the GUI finished at this point and only plan to fix bugs (if any) from now on.
If another serious bug is found that would require a new build, I might add the auto-flip option for gauntlets
if that'd would be satisfactory (and solve the view problem automatically, assuming you want to watch from gauntlet seed's POV).

[mar]

build 154 is up:
- fix win version info so that it's listed as "Myracle GUI" in win10 taskmanager instead of "Chess GUI"
- auto-flip board from gauntlet seed's pov can be enabled in settings/tournament
- new details button in elo graph view (in tournament tab) can be used to toggle detailed opponent list along with points, score, rel. elo, WLD and individual results up to 200 games

people, for the n-th time already, please stop asking for new features, be respectful...
I'll start ignoring any feature request posts/PMs/emails from now on, since this is getting annoying

to sum up:
bugreports: yes, feature requests: no

[RubiChess]

Hi Martin.

First of all: Thank you for this GUI. I started testing it yesterday and it seems to be able to replace buggy Arena and low-featured Cutechess.
But... I have to report a bug.

I'm using build 153 of the Linux version in Ubuntu 26.04. Looking at the changes of build 154 I'm quite sure it won't fix the bug but if you insist, I will test 154.

Here's the bug description:
I started a round robin tournament with 22 versions of my engines with concurrency enabled. My system is a Ryzen 9 4900H with 8 cores / 16 hyper-threads
The tournament starts with 8 parallel games but after some games finished the myracle process takes more and more cpu time, and number of parallel games decreases.

Here's what 'top' shows after some time:

Code: Select all

    PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     ZEIT+ BEFEHL                                                                     
 113793 andreas   20   0 5454632 655016   8664 S  1005   4,3  57:18.80 myracle                                                                    
 118030 andreas   20   0  483632 204028  28060 S  70,5   1,3   2:35.36 RubiChess-1.9                                                              
 119816 andreas   20   0  317308 229520   4904 S  63,6   1,5   1:36.28 RubiChess-20230                                                            
 118139 andreas   20   0  295848 125844   5272 S  55,0   0,8   2:20.53 RubiChess-1.0                                                              
 118144 andreas   20   0  425288 238120  13016 S  45,0   1,6   2:19.54 RubiChess-20240                                                            
 119809 andreas   20   0  341428 230896   5904 S  36,4   1,5   1:21.38 RubiChess-20240                                                            
 118037 andreas   20   0  427120 219996  23704 S  29,1   1,4   2:40.07 RubiChess-2.0.1

So only three games are running and the myracle process uses 1000% cpu which is 10 logical cores.
Even after pausing the tournament and closing the Myracle GUI the process remained at 1200% cpu until I killed it.

I can reproduce this problem by restarting Myracle and resuming the tournament.

Yesterday I played another tournament with several Stockfish versions and I didn't notice any problems with high cpu usage.
So my guess is that either one of my engines does some strange thing regarding UCI communication with Myracle or one of the few differences in the tournament settings causes the problem which are
- using an epd opening book UHO_Lichess_4852_v1.epd (in the well running SF tournament I used one of your build-in books)
- override engine settings: 1 thread (which shouldn't change anything cause this is the default of all the engines)
- use global syzygy path (I'm not sure if I had enabled this in the SF tournament).

Any idea how to get behind that bug? Do you need any files to debug?

Regards, Andreas

[Modern Times]

That is very odd, I'm running a tournament with Build 154 under Windows, 18 engines, concurrency 8 and it is all running normally, 14 hours in.