TalkChess.com

mclane wrote:http://schachcomputerwelt.foren-city.de ... chach.html

Interesting.

I have a cheap Android phone with a Qualcomm 600MHz single-core processor (not overclocked), with Android 2.3.7. I ran the test for two minutes (if I understood correctly) and got 25,750 NPS (17 plies FWIW)

I also tested this on my 7-inch Barnes and Noble Color, which I hacked to a normal Android tablet running a very smooth build. With it I got 62,500 NPS (18 plies). Also non-overclocked.

ok, i put them in the list


Droidfish with stockfish 2.2.2 engine,
beginning position, all cores (parallel computation automatical),
ANALYSIS-Mode after 2' computation:

Android phone Qualcomm 600MHz 1 core Android 2.3.7.
25.750 NPS

7-inch Barnes and Noble Color:
62.500 NPS

HTC HD2 1 Ghz 1 core:
94.000 NPS

Advent Vega Tablet, tegra2 (2 cores):
137.126

HTC one S qualcomm s4 (2 cores):
200.000

Ipad2 (2 cores):
274.000 NPS

Notebook core2duo T7600 (1 core/2 cores), vista64bit:
355.000 NPS / 641.000 NPS

HTC one X tegra3 (4 cores):
368.800

Droidfish mit stockfish 2.2.2 engine,
grundstellung, alle cores (parallelberechnung automatisch),
ANALYSE-Modus eingestellt nach ca. 2' rechnen:

Android phone Qualcomm 600MHz 1 core Android 2.3.7.
25.750 NPS

7-inch Barnes and Noble Color:
62.500 NPS

HTC HD2 1 Ghz 1 core:
94.000 NPS

Advent Vega Tablet, tegra2 (2 cores):
137.126

HTC Flyer, 1.5 Ghz:
140.000

HTC one S qualcomm s4 (2 cores):
200.000

Sony Tablet S:
218.000

samsung galaxy S2:
253.000

Ipad2 (2 cores):
274.000 NPS

Notebook core2duo T7600 (1 core/2 cores), vista64bit:
355.000 NPS / 641.000 NPS


HTC one X tegra3 (4 cores):
368.800

Anyone test on the new iPad 3rd Generation? It should be quite a bit faster then last years model.

Terry McCracken wrote:Anyone test on the new iPad 3rd Generation? It should be quite a bit faster then last years model.

http://talkchess.com/forum/viewtopic.ph ... =&start=20
The new iPad 3 only very slightly better than iPad 2, most of the extra processing power is dedicated graphics, so is still a Dualcore for cpu tasks like chess.

no - for chess it is not really faster. the amount of cpus or cores has not changed. apple will be overtaken by other companies ...

mclane wrote:no - for chess it is not really faster. the amount of cpus or cores has not changed. apple will be overtaken by other companies ...

If anyone has one to test, I would expect the Asus Transformer Prime to be better for chess because it has the Tegra-3 4 core processor.

I have also read that the latest Qualcomm Snapdragon 2-core processor is remarkably fast (but this may be related to graphics processing rather than program processing).

There is a free Linpack benchmark download for Android phones/tablets, and it would be very interesting to see whether there is a strong correlation between Linpack benchmark and chess NPS (nodes per second).

mclane wrote:HTC one X tegra3 (4 cores):
368.800

???
= 2 cores ???

towforce wrote:

mclane wrote:no - for chess it is not really faster. the amount of cpus or cores has not changed. apple will be overtaken by other companies ...

If anyone has one to test, I would expect the Asus Transformer Prime to be better for chess because it has the Tegra-3 4 core processor.

Paul, the tegra3 is IN the HTC one X that is benchmarked here.
the asus transformer prime has the tegra3 hardware too, but is a tablet pc while the new HTC one X is a mobile.

I have also read that the latest Qualcomm Snapdragon 2-core processor is remarkably fast (but this may be related to graphics processing rather than program processing).

Paul, the Qualcomm Snapdragon S4 CPU with 2 cores IS tested in this benchmark too. The mobile is called HTC one S. as you can see, the chess benchmark shows it is NOT faster than the HTC one X.
the benchmark those guys used was not using all cores !
therefore it showed the HTC one S ahead of the one X.

now our chess benchmark shows this was wrong.

There is a free Linpack benchmark download for Android phones/tablets, and it would be very interesting to see whether there is a strong correlation between Linpack benchmark and chess NPS (nodes per second).

i have no idea but what could be better to test stockfish and count the NPS ??