HAL9000

HAL9000
"It just isn't conceivable that you can design a program strong enough to beat players like me."

December 31, 2015

Android XB engine update: ExChess 7.88 beta JA

I really like ExChess for its flawless compatibility with Chess for Android. It can use multi threads and delivers one of the best chess performance among the Xboard/Winboard engines. At present only Crafty is playing at relatively "close" ELO level.

The new version 7.88 beta is now ported to Android by Jim Ablett. I expect this one will break the 2900 ELO barrier when i'll have time to introduce it in Rapidroid.

The package below includes both arm5 and arm7 versions and a book as well, in case you want to use it. Don't forget to install search.par too before running the engine. As far as i remember the book is not crucial but the search.par is absolutely necesssary.

Download: HERE

Android UCI engine update: Stockfish 7 beta 2 JA

Today's weather report was: Cloudy with fish rains... It seems we'll have to see some more alphas, betas and gammas before the final release.

The second beta version is compiled by Jim Ablett and this is the most recent version you can run on your Android device.

Download Stockfish 7 beta 2 for arm5: HERE
Download Stockfish 7 beta 2 for arm7: HERE

December 30, 2015

Android UCI engine update: Stockfish 7 beta 1 JA

The seventh fish is getting closer. Jim Ablett has already compiled the source code with commit number a5c76d6 of the first beta. Thanks Jim!

It would be interesting to compare these JA builds to Peter Österlund's builds released with the development version of Droidfish 1.59. Maybe some people will find enough time to conduct a head-to-head match. 

I hope the official version will not delay any longer because the whole community is waiting for v7 since one year. And one year of wait, in today's scale, is like a decade of the 80's.

Download Stockfish 7 beta 1 for arm5: HERE
Download Stockfish 7 beta 1 for arm7: HERE

Droidfish 1.59 development version featuring Stockfish 7 beta 1

Another development version of Droidfish  just came from Peter Österlund. Even though, it's still entitled 1.59, the package is different and the good news is that Stockfish 7 beta 1 comes with it.

I will upload the engine binaries included later. However, if you are interested to use them under other GUI's like Chess for Android, it's possible. You can easily strip them out. Just open the apk with 7-Zip and go to the folder named assets and extract the one which meets your architecture.

If you have Lollipop, choose x86_64 for Intel and arm8 for Arm.

To run 32-bit versions under Lollipop, choose arm7 or armeabi (arm5) but with nopie option, though 32-bit is not recommended when 64-bit builds are available!

If you run Kitkat or older like me, take arm7 or armeabi (arm5). Arm7 is better. Nopie is not mandatory.

I didn't install this version of Droidfish yet. I've thought it Stockfish 7 beta 1 embedded in the apk because the string was present in the binaries. In case i'm wrong, don't hesitate to correct me.

Download the apk: HERE

December 28, 2015

Stockfish 121215 to overtake Komodo 9.3 in Rapidroid

The fish is back on duty...


This time, the roof above 3300 ELO is fully in flames, hotter than hell.

In contrast to my expectations, the development version of Stockfish looks unstoppable in Rapidroid after 9 rounds played out of 17. I guess we will soon celebrate a new owner of the crown.

The ELO gain over Stockfish 6 is visible and looks reflected, as seen in the snapshot. The biggest surprise here is Komodo 9.3. It plays more drawish than 9.2 with default settings and loses more often to other engines like Critter, Mamba and Firenzina, not only Stockfish. It's incredible that Stockfish 121215 lost only one game so far, not to Komodo but to Critter. Highly interesting.

I'm surprised with the results and i can't wait to see how this race will end. To follow up!

December 27, 2015

Droidfish 1.59 without hash limit

By default, the Google Play edition of Droidfish is designed to check the hash size selected by user and lower it whenever the remaining free memory is low and risky.

System safety takes priority here. However, the experienced user may prefer to use bigger hash sizes in exchange of some risk, especially on devices allocated to chess only, like tablets without cellphone functions or any Android device not yet populated with dozens of memory hungry social network apps etc.

I'm simply one of those out-of-the-standard users. Being limited to 16MB hash while 600+ MB RAM is free looks oversafe to me. The main problem in my case arises when i play manual games between Android engines and other chess progs. I just can't give 256MB hash to one Android engine under Droidfish the same way i used to do with two engines under Chess for Android.

Even with two engines running at the same time, Chess for Android runs with 2 x 256MB allocated when playing automated tournaments. However, playing manual time controlled games with CfA is impossible because it won't allow negative clocks and will simply flag. The game is terminated because of the extra time used by myself to enter the moves. Operating overheads are unavoidable.

That's why i had to use Droidfish as the only GUI for manual games, since it will not terminate the games when time runs out. 

Hopefully, Peter Österlund was very kind to compile a special version without hash limit. As Google Play installs the standard version only, the special version must be downloaded as apk and installed  manually. 

You can access related topics from Talkcess.com below for further details:
TOPIC-2 (with the link to the apk)

As an alternative download, i've placed the same apk in my repository at box.com: HERE


Arm5 to Arm7 migration tests and findings with various engines

Soon after Jim has released a nice bouquet of arm7 compiles, i've conducted some quick tests in order to verify whether these upgraded binaries really bring additional strength. As always, i need to decide whether they are eligible for replacing their elder ARM5 brothers in Rapidroid.

We all know that arm7 should, in theory, perform better than arm5 but does the practice meet the theory every time? Neuh!

I didn't write anything since years but i'm not far from programming principles. Let's just state many unexpected parameters can effect the outcome. The most remarkable one is the source code itself. An arm7 compile can significantly overcome an arm5 compile only if the source code includes algorithms which would benefit from the differences between arm5 and 7. Otherwise, there's no reason for the performance to be different.

You want me to be clearer? Let's take the example of the most basic arithmetic formula: You would have a source code to multiply any user input number by 10. You would prepare the c++ code and compile it for arm5, then for arm7. Both binaries would probably do the job in exactly the same time, no matter what target architecture is choosen.

Coming back to recent arm7 releases by Jim Ablett, i wanted to compare the performance of both targets the quickest possible way. For this, i usually run the 3 minutes benchmark from initial position and compare the kNps together with the search depth reached by the engine.

I assume here that any binary with more depth and with higher kNpsthan the other should be superior.

In case the same depth applies to both builds but the kNps favors one of them, i decide according to the ratio between the kNps. If the ratio is significant enough like more than 10%, the higher kNps shows the better build.

If above criterias don't allow a selection, we must have two builds with very close depth and kNps scores. In this case a minimum of 100 head-to-head games at 180+2 may help decide. Or not...

If still no visible gap, the last thing to do is to conduct two gauntlets in parallel using same hardware and software conditions against another engine of similar strength.

Following this method, i've tested some JA compiles on Rockchip 1.4 Ghz quadcore processor and  i've come to below verdicts (All scores given as arm5 followed by arm7):

One thing that may have mattered is that i didn't use the newest arm5 compiles that JA releases with arm7, instead i kept using the older arm5 releases rated in Rapidroid. Unexpected results below may also be related to the efficiency of the compiles from different date.

* Bison 9.11: Old ARM5 is better than new ARM7.
The benchmark gave 89 kNps vs 85 at depth 17 vs 17. Head to head 100 games result is 55.5 vs 44.5 favors arm5 by 38 ELO.  Then i've done twin gauntlets vs Mini Rodent 1.0. Arm5 scored 52-48 (+14 ELO) while Arm7 scored 44.5-55.5 (-38 ELO). All 3 matches favored arm5 by ~45 ELO.

* Crafty 24.1: Old ARM5 is better than new ARM7.
362 kNps at depth 21 drops to 347 kNps at depth 21. No need to analyze more because v25.0 is already released.

* Crafty 25.0: ARM7 is clearly better.
332 kNps at depth 20 is way lower than 427 kNps at depth 20. Depths are the same but arm7 showed shorter time to depth performance thanks to higher nps.

* Discocheck 5.2.1: Old ARM5 is better than new ARM7.
Arm7 is disappointing. 144kNps @ D18 drops to 134 kNps @ D18. Arm-7 is slower considering the depths step up slightly faster with ARM5.

* ExChess 7.71b: ARM7 is clearly better.
387 kNps @ D24 rises to 555 kNps @ D25

* Fridolin 2.0.0: No verdict yet. I must do twin gauntlets vs other engines or keep ARM5.
1192 kNps @ D18 are close to 1142 kNps @ D18

* Hakkapeliitta 3.0: Old ARM5 is clearly better.
There's something terribly wrong with Arm7 compile because 439 kNps @ D23 vs 62 kNps @ D18 is more than a usual gap. Arm7 is very very slow.

* Phalanx XXIV: ARM7 is slightly better
154 kNps @ D15 rises to 170 kNps @ D15

* RedQueen 1.1.97: ARM7 is clearly better
174 kNps @ D21 rises to 198 kNps @ D22

* Scorpio 2.7.7: ARM7 is clearly better
The result is to blame on SMP function which still doesn't work for arm5. Hash setting of scorpio.ini works but arm5 build can't use 4 threads. Arm7 wins easily. 215 kNps @ D21 vs 786 kNps @ D24 !!! Parallel search rocks!

* Senpai 1.0: Old ARM5 is clearly better.
294 kNps @ D22 drops to 263 kNps @ D21. Arm7 compile may have suffered from something.

And finally, given the number of surpises noticed, i've finally wanted to check Stockfish 121215 which will probably take back its crown from Komodo 9.3 in Rapidroid. This time ARM7 overcomes ARM5 easily: 441 kNps @ D23 rises to 645 kNps @ D23. This result clearly confirms ARM7 is much better.

Now, i guess Jim will take a deeper look to discover what happened to Bison, Hakkapeliitta, DiscoCheck and Senpai. Meanwhile i will keep the relevant ARM5 builds running in Rapidroid.

Others? Migrating from ARM5 to ARM7 seems healthy. This means: Many gauntlets and reprises await for me.

Droidfish 1.59 featuring Stockfish 231015

If you are an Android user interested in computer chess and if you never ever installed Droidfish, then you must be really exceptional. That nice app is written by Peter Österlund and it's considered to be the best GUI by most chess engine fans.

So it's nice to see that Droidfish has been updated to v1.58 on December,25th.

UPDATE ON DECEMBER,26TH: 1.58 becomes 1.59 to heal the Stockfish startup problem on devices with Android 4.0 and older.

The update implements below improvements:
- Use a drawer layout instead of the options menu to fix problems on some new devices without a menu button. Swipe from left/right screen edges to open the drawer menus.
- Moved "number of search threads" from preferences to UCI engine options.
- Added configurable button action to change UCI options.
- Updated Stockfish from v6 to development version 231015.
- At least Android 3.0 required.

Droidfish is free and available at GOOGLE PLAY.

Android UCI engine update: Fridolin 2.0.0 JA for arm7

Fridolin is an SMP engine, meaning that it can do multi threaded parallel search. Although the function seems to work well, the engine still needs more progress to climb higher in rating lists.

Android is not different here. Fridolin is just above the average of the population even with four cores. I think Jim Ablett's new compile for arm7 may help Fridolin gain some more ELOs in Rapidroid. But let's wait and see whether the practice meets the theory!

Download Fridolin for arm7: HERE

Android XB engine update: ExChess 7.71b JA for arm7

ExChess is currently the strongest Winboard engine available for Android, ahead of Scorpio 2.7.7 and Crafty 24.1.

If you think it's not enough, HERE is the arm7 binary of ExChess, compiled by Jim Ablett and ready to increase the gap with others.

Android UCI engine update: Bison 9.11 JA for arm7

Bison is now upgraded from arm5 to arm7, following many other engines. By definition it should play stronger than the arm5 binary.

Meanwhile, don't be surprised by the variety of the engine ID's shown here and there. The original version number is 9.11 (probably for November,2009) while the GUI's will show 15.12, the year and the month the binary was compiled. Don't take it as the real version number.

Anyway, it's an old engine abandoned long ago and we know it's 9.11, regardless of what will be shown.

Don't forget to install Parameters.txt as well the same way you install the engine. The name MUST BE Parameters.txt and NOT parameters.txt!

Download Bison 9.11 for arm7: HERE

Android UCI engine update: Hakkapeliitta 3.0 JA for arm7

Another arm7 binary, this time for Hakkapeliitta, which will probably perform slightly better than arm5: HERE

Android UCI engine update: DiscoCheck 5.2.1 JA for arm7

Lucas Braesch's DiscoCheck is not maintained anymore since he participated to the Stockfish team long. It's a dead engine today but it's among the top engines in Rapidroid ranking.

The binary being shared HERE is nothing but a recompile made for arm7 devices and it is supposed to replace the old arm5 compile.

Android XB engine update: Crafty 25.0 JA

The new Crafty 25.0 is now ported to Android. Thanks to Jim Ablett for bringing it to our mobile devices very quickly.

Download for arm5: HERE
Download for arm7: HERE

Android XB engine update: Crafty 24.1 JA for arm7

Though the newest v25.0 is quickly ported to Android as well by Jim, the previous version had been made available for arm7 just one day ago.

I think very few people will still need 24.1 upgraded to arm7 but i share it anyway: HERE

Android XB engine update: Phalanx XXIV for arm7

Phalanx is recompiled by Jim Ablett for arm7 devices. The engine uses Winboard protocol and works only with Chess for Android. Enjoy: HERE

Android UCI engine update: RedQueen 1.1.97 JA for arm7

RedQueen upgrades from arm5 to arm7 thanks to Jim Ablett. What i've been telling about Senpai in the previous post, applies to RQ too. I want to see how stronger is the arm7 compile vs arm5 before switching to it in Rapidroid.

Those who want to go for arm7 directly may download the new binary: HERE

Please note that the engine will show up as RedQueen 1.1.98 once imported although Jim's release notes refer to 1.1.97 recompiled for arm7. 

Android UCI engine update: Senpai 1.0 JA for arm7

New compiles for arm7 by Jim Ablett will probably keep raining for a while. Another candidate to replace the older arm5 compile is Senpai for arm7.

Rapidroid used Senpai arm5 compiled by Jim instead of the original arm5 because Jim's binary delivers 30% more nodes than the original one, meaning more ELO.

I hope arm7 will add even more but i prefer to see this confirmed after a head to head match before switching to arm7.

Download Senpai 1.0 arm7: HERE

Android UCI engine update: Galjoen 0.3.1.1 JA

Galjoen just gets another update from 0.3.1 to 0.3.1.1. Nice to hear but although i'm glad to share a link to the binary, i must notify that the previous version is now disqualified from Rapidroid.  It will not be included in the next release because of permanently hogging one of the cores despite the ponder off setting of the GUI.

When i've tried the new version compiled by Jim Ablett, i've seen the same behaviour continues.

No matter what, if you're interested you can still use the engine but don't forget that it will not stop running when it's the opponent's turn.

Download Galjoen 0.3.1.1: HERE   .

December 25, 2015

Merry Christmas and Happy New Year to all chess friends


Chess has progressed in 2015. So did the computers and processors. So did the computer chess. But, unfortunately, it's hard to think that the same applied to worldwide peace and freedom and happiness.
Nevertheless, i will still dream for a peacefull world where the only battles would happen on 64 squares of the chessboards...
...and i wish a Merry Christmas and a happy new year to all.

December 23, 2015

Rapidroid to include Stockfish 121215 and Komodo 9.3 in next release

I've started to replace all games played by Stockfish 6 and Komodo 9.2 with their most recent versions available for Android.

After first 50 games, Stockfish 121215 seems to compete much better than Stockfish 6 could.

A longer time frame between two Stockfish versions compared to Komodo's state, there's nothing unusual and unexpected here.

Anyway, we will soon find out how close they will rank. It's really hard to guess which will perform better and i'm sure even 300+ games will not be enough to assure a concrete verdict as long as the ELO difference remains narrower than the error margins. 

Then, let's stay tuned for Rapidroid January-2016 while the two Exynos processors try to survive the quadcore torture running at 1.6 Ghz in my refrigerator!

December 16, 2015

New Android UCI engine: Stockfish Matefinder 1e8836d JA

Matefinder is a fork of Stockfish engine specially tuned for finding checkmating variations. I've read it's maintained by Joseph Ellis but i don't have a deep knowledge about its abilities.

I'm not interested in rating its chess strength and introduce it in Rapidroid. Meanwhile, those who wish to use it for chess puzzles may enjoy two compiles made by Jim Ablett.

On December 23th Jim has released reworked versions based on commit 1e8836d expected to run without crash on Android devices. Below new links replace the old ones which crashed on load and will direct you to the corrected builds:

Download arm5 binary: HERE
Download arm7 binary: HERE

Arm7 vs Arm5: How many more ELO?

By default, it's expected that an arm7 compile plays better chess than the same source compiled for arm5. We know that arm5 is older, standard, trustable and compatible while arm7 intruduces new commands that increase the performance and arm-8 goes even further to cover the requirements of an 64-bit Android.

Shortly said: arm5, 7, 8 are something like Intel's MMX, SSE, AVX and so on. The progress does not and will not stop.

Regarding the engine compiles for Android, Jim Ablett is recently replacing some of his older arm5 compiles by arm7 but previous arm7 introductions i've done for Rapidroid ranking had not yield remarkable ELO improvements for Cheng and Deep Saros.

Since i don't like practices conflicting with theory, i wanted to insist on my humble assumption of +30 to +50 ELO between two target architectures. So, i've just conducted a self play match between two binaries of Scorpio 2.7.7, using the last 50 openings of TCEC-7.

Both versions played with default settings, without scorpio.ini installed under Chess for Android.

At 5 seconds/move fixed, the Arm7 compile won the match with 57,5 to 42,5. The ELO gap is 52 ELO in these conditions, not far from expected, taking into account the usual self play distortion, shorter time controls and the error margin of only 100 games.

Anyway, theory meets practice. Now, i can sleep happier.

December 13, 2015

Android UCI engine update: Komodo 9.3

At last... the latest reptile comes to Android fields. We had to wait longer than expected for this because the PC version had been released quickly after TCEC-8 victory but nothing for Android.

After all it's nice to see Komodo 9.3 available for Android. All procedures remain the same: If you have Komodo 9.2, it's nothing but a simple update which will make 9.2 disappear.

Let me remind that Komodo 8, if you have it installed, will stay in place as Komodo 9 is a different app. You can have both v8 and v9.3 working on the same device!

Komodo 9.3 is not free but not expensive either. It costs only $5. So don't expect to find an apk somewhere and install it to get it running. That won't work at all. Just buy it enjoy a 3300+ ELO engine available in your pocket: HERE.

Android UCI engine update: Stockfish Dev 2015-12-12 JA

TCEC-8 was a disappointment for Stockfish. It lost the superfinal to Komodo 9.3 with only 2 wins vs 8 after 100 games played. All the remaining games were drawn.

You may read more about the superfinal: HERE

The result is not satisfying for Stockfish team and it's a logical reason for them to postpone the release of Stockfish 7, long awaited by the community.

My guess it that we will only see development versions until February. What can be done during a "few" more time is to enjoy the development versions ported to Android.

Jim Ablett has just released two binaries of the latest code stamped 7904a7d of December 12th. Although i still keep myself patient for the version update in Rapidroid, others may start using the latest available version on their devices.

Download Stockfish Dev 2015-12-12 for arm5: HERE
Download Stockfish Dev 2015-12-12 for arm7: HERE

Android UCI engine update: Komodo Beekay-1028 JA for arm7

Here's another arm7 binary supposed to perform better than the arm5 version which was previously posted HERE.

Komodo Beekay is not rated in Rapidroid because of potential ELO distortion issues. However that does not apply to other users who may be interested in running the Android brother of Komodo 5x series for Windows.

See it for yourself: HERE

Android XB engine update: Scorpio 2.7.7 JA for arm7

Scorpio is the second strongest XB engine available for Android after ExChess. The current compile was prepared for arm5 which is by default, the most compatible.

Now we can run an arm7 compile of the same source thanks to Jim Ablett. This one should perform slightly better than arm5.

Normally 30 to 40 ELO more can be expected but the theory is not always fully working, depending on how efficiently the source benefits from the instructions specific to arm7. Don't be surprised to see that the difference is barely noticeable.

Once again, the package comes with scorpio.ini but i must state here that i could never ever obtain any effect with the settings. No matter how i rename the file, Scorpio.ini or scorpio.ini, the engine does not seem to access it and uses default values.

As usual, Scorpio needs Chess for Android, the only Winboard compatible GUI for Android.

Erratum on 19-Dec-15: No. It can use scorpio.ini very well but the file must be named scorpio.ini and not Scorpio.ini. I've checked today once again and after reinsralling the file, the engine can use all four cores as well as 256MB hash. This explains why during the last rounds of Rapidroid Scorpio performed ~200 elo lower. Now it should be okay because all rounds will be repeated with the arm7 compile.

One thing will remain unknown anyway: why did i decide that SMP and hash don't function while the old logs prove the search depths were high enough?

Finally, with single core and without hash, 3 minutes thinking result is around 235kNps while quad core and 256MB hash boost the performance to 830kNps on Exynos 4412. Indeed the SMP scaling of Scorpio is very good.

Download Scorpio 2.7.7 for arm7: HERE

Android UCI engine update: Arasan 18.2 JA

Latest version of Jon Dart's Arasan is now ported to Android by Jim Ablett. There are two binaries for arm5 and arm7 to choose from.

Arasan 18.1 for Android is currently rated 2881 ELO in Rapidroid. We will soon find out how many more comes with 18.2

Download Arasan 18.2 for arm5: HERE
Download Arasan 18.2 for arm7: HERE

December 8, 2015

GreKo 2015 for Android: Why engines don't play until checkmate?

I wish i could do better with GreKo. Although it would not be able to compete with big bros, i'm sure a lot of ELO improvement is there and it would be nice to reflect the most recent status. Unfortunately all my attempts to get it running without interruptions fail since 12.6. Followers didn't help either.
Latest compile for Android labeled GreKo 2015 had given a new hope. Then, i've started two gauntlets on two of my Exynos 4412 devices to check whether it plays until the end. Ooops, mission fails in the morning, after a few hours.
GreKo is supposed to be compatible with UCI/XB  but only XB mode seems to work. Fischer clock mode is not working as always and ends up with 2 seconds per move systematically. That's why i went on trying fixed time per move mode directly but both gauntlets got stuck somewhere without termination.
I think GreKo sends something that Chess for Android can not interpret. I could adjudicate the games later if the engine simply exits and steps further to the next games but that's not the case and i have no clue to overcome suspended tournaments waiting for a move forever.
It most be noted that most of the hanged games of GreKo are wins or draws, adjudicatable according to a post analysis. A similar case to compare is Black Mamba, a very bad boy which terminates games on its own but i didn't see yet any game on which i can't decide about the result.
I don't understand what is different for GreKo...

RAPIDROID RELOADED: December-2015

While the wait for Komodo 9.3 and Stockfish 7 becomes unbearable after TCEC-8, be this release the latest one featuring the "outdated" versions of both supertoppers.

I've spent the last month almost with updates only, especially with Jim Ablett's first arm7 compiles. Therefore there's no remarkable changes this time, just narrower error margins and noticeable progress of Maverick from 1.0 to 1.5...

The highlights are:
* UPDATED Arasan 18.0 to 18.1
* UPDATED Deep Saros 2.3f arm5 to arm7 (no remarkable boost!)
* UPDATED Cheng 4.39 arm5 to arm7 (no remarkable boost!)
* UPDATED Maverick 1.0 to 1.5
* UPDATED Zurichess Fribourg to Geneva
* ADDED Floyd 0.6 as new Android engine
* ADDED Mini Rodent 1.0 as new Android engine

The list is calibrated using 39 engines rated in CCRL 40/4, assuming Exynos performs 65 and 110 ELO lower than 32 and 64 bit engines of CCRL respectively.

Some targets for the next release:
* Releasing on time :-)
* Update GreKo 12.5 to 2015 if it doesn't hang anymore during games
* Update Floyd from 0.6 to 0.7
* Add some more manual games (a working man's impossible dream)

BAYES RATINGS AFTER 15785 GAMES PLAYED BY 126 PROGRAMS
Rnk Name                     O/S T  Elo   +   - gam sco oppo dra
001 Komodo 9.2               A32 4 3317  42  40 274 83% 3056 27%
002 Stockfish 6              A32 4 3286  40  38 274 80% 3060 31%
003 Critter 1.6a             A32 4 3137  34  34 278 59% 3075 44%
004 Sting SF 4.8.4 JA        A32 4 3098  36  36 278 58% 3034 31%
005 Firenzina 2.4.1 xTreme   A32 4 3097  34  34 276 54% 3067 45%
006 BlackMamba 2.0           A32 4 3070  34  34 276 50% 3069 45%
007 Rybka 2.3.2a mp          W64 4 3041  98 101  34 43% 3096 32%
008 Texel 1.05               A32 4 3032  36  36 280 51% 3023 32%
009 DeepSaros ver.2.3f       A32 4 2992  36  36 274 49% 2997 31%
010 Senpai 1.0               A32 4 2986  36  36 288 56% 2947 27%
011 Hiarcs 13.71             IOS 2 2960 121 114  26 65% 2850 23%
012 RobboLito 0.085e4l       A32 1 2957  35  35 286 51% 2954 33%
013 Cheng 4.39               A32 4 2930  35  35 286 48% 2950 36%
014 Shredder 1.7.0           IOS 2 2897 115 113  24 54% 2880 42%
015 Gaviota v1.0-d           A32 4 2876  35  35 290 47% 2903 28%
016 Hakkapeliitta 3.0        A32 1 2866  36  36 284 48% 2892 29%
017 Arasan 18.1              A32 4 2860  35  35 286 48% 2882 32%
018 Exchess_7.71.beta.JA_xb  A32 4 2860  35  35 280 51% 2868 32%
019 Grapefruit 1.0           A32 4 2825  33  33 302 47% 2847 36%
020 Cyclone 3.4              A32 4 2820  33  33 292 50% 2818 41%
021 DiscoCheck 5.2.1         A32 1 2816  34  34 302 51% 2808 33%
022 Deep Saros 0.9           A32 4 2796  33  33 300 50% 2794 35%
023 Toga II 3.0              A32 1 2789  34  34 288 49% 2797 35%
024 Deuterium v14.3.34.130   A32 1 2758  33  32 312 53% 2735 38%
025 Bobcat 6.4b              A32 1 2746  34  34 304 50% 2743 26%
026 Fruit Reloaded 2.1       A32 1 2740  34  33 278 53% 2718 41%
027 Doch32 1.3.4 JA          A32 1 2736  33  34 298 49% 2744 36%
028 GNU Chess 5.60           A32 1 2735  35  34 296 52% 2719 27%
029 Chess Pro 2016.02        IOS 2 2730 112 114  22 45% 2759 55%
030 Murka 3                  A32 1 2723  32  32 318 52% 2708 34%
031 The King 3.50 x64        W64 1 2696  46  47 152 43% 2743 33%
032 IvanHoe 9.46b            A32 4 2690  34  34 298 46% 2720 30%
033 Strelka 5                A32 1 2680  34  33 302 53% 2658 31%
034 Scorpio_2.7.7.JA_xb      A32 1 2677  38  38 248 45% 2721 28%
035 TheMadPrune 1.7.04       A32 4 2668  35  35 272 46% 2691 32%
036 Crafty_24.1.JA_xb        A32 1 2666  34  33 314 54% 2638 25%
037 Rodent 1.7 build 1       A32 1 2628  33  33 304 49% 2633 31%
038 Tucano_5.00.JA_xb        A32 1 2626  37  37 242 49% 2640 35%
039 CNVCS 1.2.0              IOS 2 2613  94  95  36 46% 2636 36%
040 Mini Rodent 1.0          A32 1 2594  35  35 276 46% 2623 32%
041 RedQueen 1.1.97          A32 4 2586  33  34 308 46% 2612 27%
042 Bison 15.1               A32 1 2586  33  33 306 51% 2582 30%
043 Rhetoric 1.4.1           A32 1 2579  33  33 308 50% 2581 29%
044 Alfil 12.10              A32 1 2551  33  33 306 48% 2561 30%
045 Chess Genius 4.0.00      IOS 2 2544 197 257   8 13% 2800 25%
046 Gull 1.2 JA              A32 1 2535  34  34 304 50% 2531 29%
047 Cheese 1.7               A32 1 2522  33  33 304 45% 2562 32%
048 Daydreamer 1.75 JA       A32 1 2518  33  33 302 50% 2520 33%
049 Rotor 0.8                A32 1 2517  34  34 306 46% 2552 28%
050 Fridolin 2.00            A32 4 2492  33  33 294 52% 2476 34%
051 GarboChess 3             A32 1 2488  34  34 302 48% 2504 24%
052 Chess Genius 2.6.4       A32 1 2486 224 239   4 38% 2534 75%
053 Glaurung Mainz           A32 1 2481  41  41 200 45% 2518 29%
054 Danasah_5.07.JA_xb       A32 1 2469  36  36 272 54% 2434 27%
055 Sloppy_0.23.JA_xb        A32 1 2450  34  34 278 52% 2432 33%
056 BBChess 1.3b JA          A32 4 2449  34  34 296 50% 2449 28%
057 Maverick 1.5 arm         A32 1 2445  34  33 302 58% 2383 31%
058 Dirty_030411.JA_xb       A32 1 2442  36  36 274 52% 2428 26%
059 Phalanx_XXIV.JA_xb       A32 1 2411  36  36 280 45% 2448 21%
060 GreKo_12.5.JA_xb         A32 1 2407  36  36 276 55% 2371 23%
061 Pawny_1.0.JA_uci2xb      A32 1 2404  35  35 280 53% 2381 30%
062 Pepito v1.59             A32 1 2393  33  34 292 48% 2403 35%
063 BetsabeII_1.47.JA_xb     A32 1 2375  37  37 276 49% 2379 20%
064 Ifrit_m1.8.JA_uci2xb     A32 1 2346  36  35 276 55% 2308 28%
065 Typhoon_1.0.r358.JA_xb   A32 1 2326  36  35 286 56% 2281 24%
066 zurichess geneva         A32 1 2326  53  53 128 52% 2310 27%
067 Diablo 0.5.1b JA         A32 1 2313  35  35 294 49% 2315 23%
068 Olithink_5.3.2.JA_xb     A32 1 2291  36  36 284 51% 2276 20%
069 Amy_0.8.JA_xb            A32 1 2277  35  35 302 47% 2301 22%
070 Myrddin_0.86.JA_xb       A32 1 2264  37  37 282 48% 2272 23%
071 TJchess 1.1U             A32 1 2248  34  35 310 48% 2259 24%
072 Natwarlal_0.14.JA_xb     A32 1 2237  36  36 292 53% 2210 23%
073 Bitfoot 150922.JA        A32 1 2236  36  36 310 57% 2177 16%
074 MangoPaola_1.1.JA_xb     A32 1 2233  36  36 292 53% 2207 20%
075 Sungorus 1.4 JA          A32 1 2205  35  35 288 45% 2237 24%
076 KmtChess_1.21.JA_xb      A32 1 2162  36  36 298 51% 2154 21%
077 NGplay_9.86.JA_xb        A32 1 2155  36  36 294 52% 2134 21%
078 Rattate_Nosferatu.JA_xb  A32 1 2147  36  36 298 48% 2159 17%
079 Scidlet_2.61b2.JA_xb     A32 1 2125  37  37 298 52% 2109 13%
080 Resp_0.19.JA_xb          A32 1 2120  35  35 314 53% 2099 19%
081 Clubfoot 150907.JA       A32 1 2092  37  36 312 63% 1980 14%
082 DanChess_1.04.JA_xb      A32 1 2051  36  36 298 49% 2059 18%
083 Kurt 0.9.2.2 JA          A32 1 2014  35  35 314 49% 2027 19%
084 Robocide 28.12.14.JA     A32 1 2002  33  33 342 51% 1991 18%
085 Witz_Alpha21.JA_xb       A32 1 1987  36  36 294 48% 2002 20%
086 Woodpecker_2.11.JA_xb    A32 1 1967  37  37 288 52% 1949 15%
087 Knightcap_3.7F.JA_xb     A32 1 1960  37  37 280 53% 1937 19%
088 AdroitChess0.4 JA        A32 1 1950  36  36 310 49% 1952 17%
089 BikJump v1.8  ğ          A32 1 1941  33  33 336 48% 1960 22%
090 Sjeng_1.12.JA_xb         A32 1 1920  37  37 280 51% 1910 16%
091 Gunborg_1.39.JA_uci2xb   A32 1 1915  39  39 264 61% 1824 20%
092 Galjoen_v0.30.2          A32 1 1913  47  47 172 56% 1865 21%
093 Floyd 0.6 JA             A32 1 1913  37  38 314 40% 2000 13%
094 Leonidas_r83.JA_xb       A32 1 1909  37  37 282 46% 1943 18%
095 ZCT-0.3.2500             A32 1 1887  36  36 314 41% 1959 12%
096 Faile_1.44.JA_xb         A32 1 1886  36  36 276 48% 1901 28%
097 Samchess_JA_xb           A32 1 1878  37  38 276 46% 1914 17%
098 Cilian_4.14.JA_xb        A32 1 1874  35  35 306 50% 1867 25%
099 Mephisto Roma Turbo      W64 1 1870  79  83  56 37% 1972 16%
100 Ecce rev. 508            A32 1 1835  35  36 306 44% 1884 17%
101 Sayuri 2015.10.01        A32 4 1828  37  37 302 54% 1781 14%
102 Surprise_4.3.b13.JA_xb   A32 1 1825  50  50 162 52% 1805 15%
103 Claudia v. 0.5           A32 1 1813  52  52 156 53% 1789 14%
104 Colchess_8.0.JA_xb       A32 1 1804  40  40 240 51% 1782 23%
105 Smash 1.03 JA            A32 1 1798  36  36 316 49% 1792 14%
106 Zzzzzz_3.5.1.JA_xb       A32 1 1714  39  39 230 51% 1680 30%
107 Hoichess_0.12.1.JA_xb    A32 1 1701  37  38 286 48% 1701 19%
108 Chenard_2015.08.15.JA_xb A32 1 1684  44  44 236 46% 1690  9%
109 Tscp_1.8.1.AB_xb         A32 1 1679  42  42 232 47% 1685 17%
110 Colossus 4.0 100X        C64 1 1672 236 207  10 80% 1296 20%
111 Kitteneitor_060513.JA_xb A32 1 1670  40  41 220 46% 1668 34%
112 Jester_0.84.JA_xb        A32 1 1665  45  45 214 50% 1612 13%
113 Rocinante 2.0 JA         A32 1 1658  38  38 290 47% 1646 19%
114 Pulse 1.5-cpp            A32 1 1623  38  38 288 52% 1546 28%
115 Mephisto Roma 68020 UCI  W64 1 1609 124 133  20 35% 1710 30%
116 VIRUTOR CHESS 1.1.1      A32 1 1484  43  44 284 54% 1403 11%
117 Superpawn b108 JA        A32 1 1437  51  50 184 52% 1418 18%
118 Chess for Android        A32 1 1429  44  44 279 51% 1375 13%
119 K2 v.075                 A32 1 1413  56  55 178 55% 1369  5%
120 Chess Titans             W64 1 1332 212 204   7 57% 1302 29%
121 Trappy_Beowulf_2.0.JA_xb A32 1 1191  47  48 272 37% 1308  8%
122 Colossus 4.0             C64 1 1167 170 190  14 36% 1252 14%
123 Byak 8.10.14.JA          A32 1 1124  50  52 198 26% 1360 18%
124 Xadreco_5.7.JA_xb        A32 1 1017  57  63 192 15% 1381  8%
125 Novag Secondo            TTC 1  944 270  50   6 42%  965 17%
126 OliveChess 0.2.7         A32 1  579 355-308 160  0% 1495  0%


Rapidroid test platform:
* GT-N7100 1.6 * 4 + 256MB hash: All Android progs
* GT-N5105 1.6 * 4 + 256MB hash: All Android progs
* Codegen Novatab 1.4 * 4 + 256MB: Single thread Android progs,
* Polypad 1010IPS tablet 1.61 * 2 + 128MB: Single thread Android progs
* HTC Diam 528Mhz, 16MB hash: Windows Mobile
* i7 M620 2.67 Ghz dual + Arena 3.5 + 2GB hash: Windows 64
* iPhone5S A7 1.3 Ghz * 2: iOS progs
* DosBox 1.74: DOS progs,
* WinVICE 2.24: Commodore-64 progs
* Messtiny UCI adapters or CB-Emu2014: Mephisto progs
* Openings: 20 ply from Adam Hair or 16 ply from TCEC, no Q exchange, +0.15 to +0.40 evaluated by Stockfish and Komodo at depth 20 minimum, played twice both sides
* Repeating openings and twin games not allowed between two programs
* Tablebases and pondering off
* Time control: 10 to 30 sec/move or 600+0 to 1800+5 or closest known by both programs.

December 4, 2015

Android XB engine update: GreKo 2015

All Android compiles of GreKo used to cause problems to me since v12.5 on which Rapidroid is still stuck. GreKo works fine in xboard mode if you choose fixed time per move setting but Fischer clock will not work as expected because the engine uses always 2 seconds/move.

Starting from 12.6, all successors of 12.5 used to hang on move in the middle of the games several times, causing my tournaments to be interrupted. That's why 12.6, 12.8 and 13.1 could not be introduced.

I hope 2015 version will be more compatible so that i can rate the improvement.

You may give a try with Greko 2015 for Android: HERE

Android UCI engine update: Floyd 0.7.JA

Shortly after i posted the Android port of v0.6, the new version 0.7 of Floyd has just followed the same path to Android, thanks to Jim Ablett.

Floyd 0.6 is about to appear for the first and the last time in Rapidroid with ~1900 ELO. I expect that this new version will not be too different though. It remains to be ranked in December list.

The engine plays both Fischer and fixed time controls without issues.

Download Floyd 0.7: HERE



November 29, 2015

Android UCI engine update: Zurichess Geneva

Zurichess keeps travelling through Switzerland. Latest station is Geneva and we've just received an Android version compiled by Jim Ablett. I guess it's an arm5 build but i'm not sure as the binary did not disclose any trace like previous ones.

Regarding the strength, some 30-50 ELO more may be expected according to the improvements mentioned by the author. I would be glad to test and verify this quickly but i'm too close to a new Rapidroid release including Fribourg version. That's why Geneva will be replacing Fribourg later on.

In the meantime, you may download and try Geneva: HERE

November 17, 2015

Fritz 15 powered by Vasik Rajlich: Coming on November, 25th

Most of the computer chess freaks were not yet born when Fritz 1.0 was released. After decades, the oldest trademarks in computer chess is about to come to 15th version.

In contrast to Arasan which is not only as ancient as Fritz but always maintained by the same author, Fritz has seen different authors behind the curtain.

The last version is powered by Vasik Rajlich who dominated an era with his famous Rybka until the so-long criticized ICGA act, after which Rajlich lost motivation and stayed away from the chess world for a while!

Now he's back on track and i hope the new probably "Rybkatized" Fritz will bring fresh air to the field.

Regarding the strength, if you were to force me to speculate, i would say it ahould be slightly better than Houdini but not as high as Komodo or Stockfish.

The bad news is that no Android port is in the schedule.

You may read more about Fritz 15 on related ChessBase page

November 15, 2015

News from the Rapidroid labs

Unless there are new engines or updates, Rapidroid is a non-stop and unlimited tournament running by divisions, in up runners and down runners basis. There are basically 11 divisions of 9-10 or 11 engines in each and after each round 3 winners go up, 3 losers go down and the roulette turns again.

When there are newcomers to introduce or updates, things slow down and get complicated because the strict rule (Why the heck did i invent it?) tells "Only one version of each engine must be included".
That means any updated engine requires that all previously played games are removed and replayed with the new version. All related PGNs, as well as all previous rankings, graphs, stats etc. must be also updated.

This is exactly an erase and rewind scenario. I know it's hard to do when there are dozens of rounds already played but i'm sure the final result is the most trustable one without rating distortion.

In case of a new engine to introduce, it's still complicated because i have to find the correct division to introduce it without losing the balance between the number of players in each division. It may happen that no place is available in the most suitable division. In such case, the first or the last engine of that round must be ported to the neighbor divison. More and more games to replay!

The number of ECO lines played by each engine must be also monitored and kept well balanced to avoid the effects of sensibility to a given opening category. An engine should not always play ECO A for example.

After October-2014 release, i'm currently working on November rankings which will integrate below changes:

Arasan 18.1 will replace Arasan 18.0 (+20 ELO expected)
Cheng 4.39 arm5 will be replaced by arm7 (+30 ELO expected)

Deep Saros 2.3f arm7 will replace arm5 (+10 ELO expected)

Galjoen 0.31 will replace 0.30.2 (+30 ELO expected)

Maverick 1.5 Leiden will replace Maverick 1.0 (+60 ELO expected)

Floyd 0.6 will be added

Mini Rodent 1.0 will be added (Yes, it may distort for a while but that Mini and Rodent 1.7 will be both gone when Rodent 2 will step in)

Due to above updates, i delayed the introduction of the second dedicated computer Excalibur LCD Express. Hopefully noone would cry out for that :-)

Now, back to digging it deep!

Must have tools for chess with Android

Referring to an expression i like very much which says "Better be late than sorry", my status here is a little bit paradoxal because i must say: "I'm late and i'm sorry too".

In short, this post should come months before!

This is about three system utilities i always keep continously running in the background on all my Android devices during Rapidroid games.

Without these three muskeeters in place, i would feel blind and unsafe.



1) CpuTemp in Statusbar

This tool shows the cpu temperature in the status bar and not in the notifications panel like many others do. The temp info is crucial and it must be visible on the status bar because pulling down the panel to read the temp needs a touch on the screen which is risky for me when two engines play chess, especially if case the GUI is Chess for Android. You simpy don't want to interrupt a tournament accidentally!

Another nice option is three temperatıre thresholds in Fahrenheit or Celcius and relevant colors fully customizable.

The app is free HERE as a tiny module which requires the installation of Xposed Framework to work. You need luck too because not all processors are equipped with temp sensor (Rockchip for example) or the app may simply not be compatible with yours.

2) CpuStats

Monitoring the running state of the cpu cores is as vital as checking the temperature. Nothing else can tell you whether a multi core engine is really using the expected number of cores. Another issue it comes to help is when an engine refuses to stop and exit when requested by the GUI. This is dangerous for any testing environment as the faulty engine remains running in the background and hogs one core or more depending the config. Once this occurs, in the next automated game, a multicore engine may start running without one of the cores vs a single core engine. This is simply unfair.

CpuStats freely available at Google Play, will help you figure out something goes wrong by showing the load of each core on-the-fly, every x seconds. Once again, the icon is pinned on the status bar and NOT in the notifications panel.

If monitoring the state of the cores is not enough (it's not enough of course), the app will show the current frequency as well, in readable numbers, next to the cpu load icon. This feature is extremely useful in case too much temp forces your cpu to throttle and your engines to play weaker.

3) Memory Bar

If you think the temperature, the load and the frequency are enough, think again. What about the RAM usage when you try to allocate as much hash memory as possible to the engines? What about the memory eaten by tablebases? Monitoring current free RAM becomes even more crucial. That's where the third free tool, Memory Bar Simple comes to help. It has a nice icon to reside in the status bar showing the amount of free RAM in MB. Although it will show 999MB beyond 1GB free, that should not be a problem as we know enough free RAM is available for chess.