On a iPhone 6 or Galaxy S6, Chess Genius would play on what ELO level at best setting?
Thanks in advance.
Chess Genius (Mobile)
Moderators: hgm, Rebel, chrisw
-
- Posts: 6340
- Joined: Mon Mar 13, 2006 2:34 pm
- Location: Acworth, GA
Chess Genius (Mobile)
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
__________________________________________________________________
Ted Summers
-
- Posts: 1796
- Joined: Thu Sep 18, 2008 10:24 pm
Re: Chess Genius (Mobile)
At blitz I'd guess it would give Magnus a run for his money.
At tournament time controls I'd guess 2500.
At tournament time controls I'd guess 2500.
-
- Posts: 4889
- Joined: Thu Mar 09, 2006 6:34 am
- Location: Pen Argyl, Pennsylvania
Re: Chess Genius (Mobile)
Chess Genius on Palm Zire 21/31 was considered by some to be rated near 2250 on the SSDF list. The iPhone Chess Genius is nearly the same , if not the exact same chess engine , but on 6 or 6+ runs nearly 40 times faster. If you use the conservative estimate that a doubling of speed is worth 50 ELO, you could argue that the iPhone Chess Genius would be rated at ~ 2500 range. As one who played both at one second per move, I could beat the Zire on occasion at 1 second per move. Have not yet accomplished that feat yet on the iphone 6. So I believe it is fair estimate and the hardware has made a notable change in strength.AdminX wrote:On a iPhone 6 or Galaxy S6, Chess Genius would play on what ELO level at best setting?
Thanks in advance.
Below is a position from the Chess genius site has used in the past to measure hardware speed.
[D] R5Nk/7p/4pbpP/n1p4n/2p1b3/K1p3pq/8/8 w - - 0 1
My iPhone 6 at the indicated setting of hash=64kb solves this in 3 seconds averaging over 1M nps. With higher hash settings, it's even faster.
The Zire 31 took 95 seconds to see mate.
A very worthy opponent for most of us that are carbon based.
-
- Posts: 6340
- Joined: Mon Mar 13, 2006 2:34 pm
- Location: Acworth, GA
Re: Chess Genius (Mobile)
Werewolf wrote:At blitz I'd guess it would give Magnus a run for his money.
At tournament time controls I'd guess 2500.
Thanks to both of you for the reply. I tested the position on my S6 Edge, it solved it in one second using the settings you gave. On my Samsung Galaxy Note 12.2 Tablet it took two seconds @623355/sec.MikeB wrote: Chess Genius on Palm Zire 21/31 was considered by some to be rated near 2250 on the SSDF list. The iPhone Chess Genius is nearly the same , if not the exact same chess engine , but on 6 or 6+ runs nearly 40 times faster. If you use the conservative estimate that a doubling of speed is worth 50 ELO, you could argue that the iPhone Chess Genius would be rated at ~ 2500 range. As one who played both at one second per move, I could beat the Zire on occasion at 1 second per move. Have not yet accomplished that feat yet on the iphone 6. So I believe it is fair estimate and the hardware has made a notable change in strength.
A very worthy opponent for most of us that are carbon based.
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
__________________________________________________________________
Ted Summers
-
- Posts: 61
- Joined: Wed Mar 14, 2012 9:43 pm
Re: Chess Genius (Mobile)
Probably a more complicated position is better suited for faster CPUs, because ChessGenius 3.0 solves the Mate in about 1/10 of a second
on an 1.7 GHz Pentuim M Dothan from 2004.
[D]rbb2rk1/1p2qppp/p4n2/2p5/P1B1p3/1P2P3/1BQPNPPP/R4RK1 b - -
Genius 3.0 with 32M hash needs 23 seconds on an i7 @ 3.3 GHz to find Bxh2 in depth 8. Some other results for comparison:
Pentium M @ 600 MHz 256k hash 00:02:36
Pentium M @ 600 MHz 32M hash 00:02:06
Pentium M @ 1.7 GHz 256k hash 00:00:53
Pentium M @ 1.7 GHz 32M hash 00:00:46
on an 1.7 GHz Pentuim M Dothan from 2004.
[D]rbb2rk1/1p2qppp/p4n2/2p5/P1B1p3/1P2P3/1BQPNPPP/R4RK1 b - -
Genius 3.0 with 32M hash needs 23 seconds on an i7 @ 3.3 GHz to find Bxh2 in depth 8. Some other results for comparison:
Pentium M @ 600 MHz 256k hash 00:02:36
Pentium M @ 600 MHz 32M hash 00:02:06
Pentium M @ 1.7 GHz 256k hash 00:00:53
Pentium M @ 1.7 GHz 32M hash 00:00:46
-
- Posts: 18754
- Joined: Thu Mar 09, 2006 6:40 pm
- Location: US of Europe, germany
- Full name: Thorsten Czub
Re: Chess Genius (Mobile)
The lastest Genius Versions are 64 bit btw.
The old dedicated genius modules were maximum 32 bit and the PC programs were also 32 bit only, or even 16.
Of course genius is no multi CPU program. So on a 8 core mobile it uses only 1 core.
The old dedicated genius modules were maximum 32 bit and the PC programs were also 32 bit only, or even 16.
Of course genius is no multi CPU program. So on a 8 core mobile it uses only 1 core.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
Here we have a fairy tale of the day after tomorrow....
-
- Posts: 6340
- Joined: Mon Mar 13, 2006 2:34 pm
- Location: Acworth, GA
Re: Chess Genius (Mobile)
That I did not know. (The latest Genius Versions are 64 bit btw.) Thanks for the info.mclane wrote:The latest Genius Versions are 64 bit btw.
The old dedicated genius modules were maximum 32 bit and the PC programs were also 32 bit only, or even 16.
Of course genius is no multi CPU program. So on a 8 core mobile it uses only 1 core.
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers
__________________________________________________________________
Ted Summers
-
- Posts: 196
- Joined: Wed Oct 29, 2014 12:42 am
Re: Chess Genius (Mobile)
I will introduce Cg 2.6.4 android in next release of Rapidroid. So far 4 manual games and it started around 2350 on exynos 4412. I expect it reaches barely 2450 on iphone5. All above concerns 32 bit single core.
You should not expect 2600 even on extreme mobiles. Shredder and hiarcs look both stronger than CG.
You should not expect 2600 even on extreme mobiles. Shredder and hiarcs look both stronger than CG.
My blog for Android users: http://chesstroid.blogspot.com
-
- Posts: 61
- Joined: Wed Mar 14, 2012 9:43 pm
Re: Chess Genius (Mobile)
According to the manual CG3 DOS is a 32 bit program. Thats the reason why it needs at least a 80386.mclane wrote:The lastest Genius Versions are 64 bit btw.
The old dedicated genius modules were maximum 32 bit and the PC programs were also 32 bit only, or even 16.
-
- Posts: 73
- Joined: Mon Mar 24, 2014 9:48 am
- Location: New Zealand
Re: Chess Genius (Mobile)
The original problem had the black Qh3 and pawn g3 so might take longer. This mate test position looks like the one that Novag dedicated machines had in the manual showing the difference in the find mate in x compared to general game play. Mate in x was faster on this position than a general analysis search.AdminX wrote:Werewolf wrote:At blitz I'd guess it would give Magnus a run for his money.
At tournament time controls I'd guess 2500.Thanks to both of you for the reply. I tested the position on my S6 Edge, it solved it in one second using the settings you gave. On my Samsung Galaxy Note 12.2 Tablet it took two seconds @623355/sec.MikeB wrote: Chess Genius on Palm Zire 21/31 was considered by some to be rated near 2250 on the SSDF list. The iPhone Chess Genius is nearly the same , if not the exact same chess engine , but on 6 or 6+ runs nearly 40 times faster. If you use the conservative estimate that a doubling of speed is worth 50 ELO, you could argue that the iPhone Chess Genius would be rated at ~ 2500 range. As one who played both at one second per move, I could beat the Zire on occasion at 1 second per move. Have not yet accomplished that feat yet on the iphone 6. So I believe it is fair estimate and the hardware has made a notable change in strength.
A very worthy opponent for most of us that are carbon based.