Man vs Machine weak hardware?

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

bigo

Man vs Machine weak hardware?

Post by bigo »

Be honest how do you think Rybka 3 32bit would rate against the top 20 strongest humans running on 1.6 atom netbook? Larry K? anyone?
Milos
Posts: 4190
Joined: Wed Nov 25, 2009 1:47 am

Re: Man vs Machine weak hardware?

Post by Milos »

bigo wrote:Be honest how do you think Rybka 3 32bit would rate against the top 20 strongest humans running on 1.6 atom netbook? Larry K? anyone?
+150 ELO (or 70%) at least...

Btw. Atom 1.6 is not slow at all, it is as fast as (if not even faster than) P4 at 3.0 GHz.
User avatar
Mike S.
Posts: 1480
Joined: Thu Mar 09, 2006 5:33 am

Re: Man vs Machine weak hardware?

Post by Mike S. »

In 1994, Fritz 3 on a Pentium 90 MHz won a blitz tournament ex aequo with Garry Kasparov, who achieved the same score, by 12.5/17 each. The opponents included Anand, Short, Gelfand, Kramnik, Leko...

(Kasparov won the playoff match, including a game where he played 1.e3.)

It was blitz, but also it wasn't Rybka but Fritz 3, and not an Atom 1.6 but a P90, only. So, this is a question I stopped asking ~15 years ago, when the P90 was "outdated" again, as usual with any CPU.
Regards, Mike
User avatar
Thomas Mayer
Posts: 385
Joined: Thu Mar 09, 2006 6:45 pm
Location: Nellmersbach, Germany

Re: Man vs Machine weak hardware?

Post by Thomas Mayer »

Milos wrote:
bigo wrote:Be honest how do you think Rybka 3 32bit would rate against the top 20 strongest humans running on 1.6 atom netbook? Larry K? anyone?
+150 ELO (or 70%) at least...

Btw. Atom 1.6 is not slow at all, it is as fast as (if not even faster than) P4 at 3.0 GHz.
Benchmarks anyone ? Afaik it is compareable to a 1 GHz Centrino ?!

Greets, Thomas
Milos
Posts: 4190
Joined: Wed Nov 25, 2009 1:47 am

Re: Man vs Machine weak hardware?

Post by Milos »

Thomas Mayer wrote:
Milos wrote: Btw. Atom 1.6 is not slow at all, it is as fast as (if not even faster than) P4 at 3.0 GHz.
Benchmarks anyone ? Afaik it is compareable to a 1 GHz Centrino ?!
Maybe something like this:
I was amazed to see a benchmark of the 901 Atom. Running w/Prime 32M the Atom beat my 3.06GHz P4.
The Atom is a 1.6GHz cpu but according to the article I read it runs overclocked at 1.8GHz when set to performance mode.
http://hwbot.org/benchmark.do

Asus Eee PC 901 (Intel Atom @ 1.8GHz) 111 seconds
My P4 @ 3.06GHz w/915GAG chipset 138 seconds
My other P4 @ 2.4GHz w/SiS651 chipset 155 seconds
Asus Eee PC 900 (Intel Celeron M ULV @ 900MHz) 203 seconds

HP 2133 Mini-Note (Via CV7-M ULV @ 1.6GHz) 168 seconds
Asus Eee PC 4G (Intel Celeron M ULV @ 630MHz) 289 seconds
Asus Eee PC 4G (Intel Celeron M ULV @ 900MHz) 200 seconds
Everex CloudBook (VIA C7-M ULV @ 1.2GHz) 248 seconds
Fujitsu U810 Tablet PC (Intel A110 @ 800MHz) 209 seconds
Sony VAIO VGN-G11XN/B (Core Solo U1500 @ 1.33GHz) 124 seconds
Sony VAIO TZ (Core 2 Duo U7600 @ 1.2GHz 76 seconds
The record is 2x Core 2 QX9775 (3.2GHz CPU's @ 5708mhz) 3.45 seconds

This is one of the few benchmarks that show the C7 winning against the 900. But the Atom is the real surprise.
Uri Blass
Posts: 10815
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Man vs Machine weak hardware?

Post by Uri Blass »

bigo wrote:Be honest how do you think Rybka 3 32bit would rate against the top 20 strongest humans running on 1.6 atom netbook? Larry K? anyone?
This is a very strong hardware
I expected something at least 10 times slower to qualify for weak hardware if you want equal chances.

Uri
User avatar
Dr.Wael Deeb
Posts: 9773
Joined: Wed Mar 08, 2006 8:44 pm
Location: Amman,Jordan

Re: Man vs Machine weak hardware?

Post by Dr.Wael Deeb »

Uri Blass wrote:
bigo wrote:Be honest how do you think Rybka 3 32bit would rate against the top 20 strongest humans running on 1.6 atom netbook? Larry K? anyone?
This is a very strong hardware
I expected something at least 10 times slower to qualify for weak hardware if you want equal chances.

Uri
In less than 2 years a chess engine running on iPhone will wipe the floor with the strongest human chess player at any time control....
Dr.D
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….
alpha123
Posts: 660
Joined: Sat Dec 05, 2009 5:13 am
Location: Colorado, USA

Re: Man vs Machine weak hardware?

Post by alpha123 »

Dr.Wael Deeb wrote:
Uri Blass wrote:
bigo wrote:Be honest how do you think Rybka 3 32bit would rate against the top 20 strongest humans running on 1.6 atom netbook? Larry K? anyone?
This is a very strong hardware
I expected something at least 10 times slower to qualify for weak hardware if you want equal chances.

Uri
In less than 2 years a chess engine running on iPhone will wipe the floor with the strongest human chess player at any time control....
Dr.D
http://hiarcs.com/Games/Mercosur2009/mercosur09.htm

I expect it wouldn't be much different with an iPhone. Probably even better with an iPad.

Peter
CRoberson
Posts: 2093
Joined: Mon Mar 13, 2006 2:31 am
Location: North Carolina, USA

Re: Man vs Machine weak hardware?

Post by CRoberson »

Milos wrote:
Thomas Mayer wrote:
Milos wrote: Btw. Atom 1.6 is not slow at all, it is as fast as (if not even faster than) P4 at 3.0 GHz.
Benchmarks anyone ? Afaik it is compareable to a 1 GHz Centrino ?!
Maybe something like this:
I was amazed to see a benchmark of the 901 Atom. Running w/Prime 32M the Atom beat my 3.06GHz P4.
The Atom is a 1.6GHz cpu but according to the article I read it runs overclocked at 1.8GHz when set to performance mode.
http://hwbot.org/benchmark.do

Asus Eee PC 901 (Intel Atom @ 1.8GHz) 111 seconds
My P4 @ 3.06GHz w/915GAG chipset 138 seconds
My other P4 @ 2.4GHz w/SiS651 chipset 155 seconds
Asus Eee PC 900 (Intel Celeron M ULV @ 900MHz) 203 seconds

HP 2133 Mini-Note (Via CV7-M ULV @ 1.6GHz) 168 seconds
Asus Eee PC 4G (Intel Celeron M ULV @ 630MHz) 289 seconds
Asus Eee PC 4G (Intel Celeron M ULV @ 900MHz) 200 seconds
Everex CloudBook (VIA C7-M ULV @ 1.2GHz) 248 seconds
Fujitsu U810 Tablet PC (Intel A110 @ 800MHz) 209 seconds
Sony VAIO VGN-G11XN/B (Core Solo U1500 @ 1.33GHz) 124 seconds
Sony VAIO TZ (Core 2 Duo U7600 @ 1.2GHz 76 seconds
The record is 2x Core 2 QX9775 (3.2GHz CPU's @ 5708mhz) 3.45 seconds

This is one of the few benchmarks that show the C7 winning against the 900. But the Atom is the real surprise.
Prove to me that there exists a correlation between Prime performance and Chess program performance. There is a reason why there are so many different benchmarks. Different programs put different stresses on hardware. Thus, one benchmark may not correlate at all to another. Prime is a poor predictor of general performance.

Here are some real chess benchmarks. The CCRL standard is an AMD 4600+

Code: Select all

 Intel Atom N270    2 x 1600    767 nps  Fritz benchmark 2 procs
 P4 3.2 GHz         1 x 3200   1300 nps
 Amd 4600+          2 x 2400   2685 nps
A single cpu P4 3.2 GHz performs nearly 2 x faster than a 2 proc Atom.

Code: Select all

             Crafty Benchmarks
   CPU                    GHz          Bits                     NPS       Time
Intel Atom N270   1.60GHz   32 bit   2CPU     648171    200.29 
Pentium 4 524 HT  3.06GHz   32 bit   1CPU    1015201    110.09 
AMD 64 X2 4600+   2.40GHz   64 bit   2CPU    4167659    35.94
Here you can see the single CPU P4 at almost 2x the speed of a dual Atom. Also, you see the AMD 4600+ at almost 7x the speed of the Atom.

Another important note is that Crafty gained more speed going from Atom to the AMD 4600+ than Fritz. Even within chess programs there exists differences.

It is clear from the chess program benchmarks that Prime does not correlate with computer chess performance and thus is a poor predictor of chess performance.