harware vs software advances

[Don]

ok.

from the position
r2qkbnr/1ppb1ppp/p1np4/1B2p3/3PP3/2N2N2/PPP2PPP/R1BQK2R w KQkq - 0 6

chess genius play Ab5xc6 after 2:44 with depth 7
line b5xc6 d7xc6 d4xe5 f7f5 c1g5 f8e7 g5e7

Well then something is not right. It takes about 5 minutes to complete the 5 ply search.

It must be the dosemu - it's supposed to be very close to native speed but I don't believe it is.

I'm also getting a different principal variation than you are, so are you sure you are using chess genius 3?

now modern programs.
let start with Komodo.
(Jim A. compile 32 bits)

commands
uci
position startpos
go movetime 60000
quit

Code: Select all

id name Komodo32 1.2 JA
id author Don Dailey, Larry Kaufman
option name Ponder type check default true
option name Hash type spin default 32 min 0 max 4096
option name OwnBook type check default false
option name MultiPV type spin default 1 min 1 max 99
option name Book File type string default komodo.bin
uciok
info depth 1
info depth 1 time 380 nodes 20 score cp 63 nps 52 pv e2e4 
info depth 2
info depth 2 time 770 nodes 160 score cp 9 nps 207 pv e2e4 e7e5 
info depth 3
info depth 3 time 1760 nodes 558 score cp 45 nps 317 pv e2e4 e7e5 d2d4 
info depth 4
info depth 4 time 4940 nodes 1943 score cp 12 nps 393 pv g1f3 e7e6 d2d4 b8c6 
info depth 5
info depth 5 time 7579 nodes 2865 score cp 33 nps 377 pv g1f3 d7d5 d2d3 b8c6 b1c3 
info depth 5 time 8899 nodes 3682 score cp 33 nps 413 pv g1f3 d7d5 d2d3 b8c6 b1c3 
info depth 6
info currmove g1f3 currmovenumber 1
info depth 6 time 11420 nodes 5131 score cp 9 nps 449 pv g1f3 d7d5 d2d4 g8f6 b1c3 b8c6 
info currmove e2e4 currmovenumber 2
info depth 6 time 21690 nodes 8865 score cp 14 nps 408 pv e2e4 d7d5 d1f3 g8f6 b1c3 d5e4 c3e4 
info currmove d2d4 currmovenumber 3
info currmove e2e3 currmovenumber 4
info currmove b1c3 currmovenumber 5
info currmove d2d3 currmovenumber 6
info currmove b1a3 currmovenumber 7
info currmove g1h3 currmovenumber 8
info currmove b2b4 currmovenumber 9
info currmove b2b3 currmovenumber 10
info currmove g2g3 currmovenumber 11
info currmove g2g4 currmovenumber 12
info currmove c2c4 currmovenumber 13
info currmove a2a4 currmovenumber 14
info currmove h2h4 currmovenumber 15
info currmove f2f4 currmovenumber 16
info currmove c2c3 currmovenumber 17
info currmove f2f3 currmovenumber 18
info currmove a2a3 currmovenumber 19
info currmove h2h3 currmovenumber 20
info depth 6 time 25649 nodes 10751 score cp 14 nps 419 pv e2e4 d7d5 d1f3 g8f6 b1c3 d5e4 c3e4 
info depth 7
info currmove e2e4 currmovenumber 1
info depth 7 time 52010 nodes 22528 score cp 17 nps 433 pv e2e4 d7d5 e4d5 d8d5 b1c3 d5e6 g1e2 g8f6 d2d3 
info currmove g1f3 currmovenumber 2
info currmove d2d4 currmovenumber 3
info currmove e2e3 currmovenumber 4
info currmove b1c3 currmovenumber 5
info currmove d2d3 currmovenumber 6
info currmove b1a3 currmovenumber 7
info currmove g1h3 currmovenumber 8
info currmove b2b4 currmovenumber 9
info currmove b2b3 currmovenumber 10
info currmove g2g3 currmovenumber 11
info currmove g2g4 currmovenumber 12
info currmove c2c4 currmovenumber 13
info currmove a2a4 currmovenumber 14
info currmove h2h4 currmovenumber 15
info currmove f2f4 currmovenumber 16
info currmove c2c3 currmovenumber 17
info currmove f2f3 currmovenumber 18
info currmove a2a3 currmovenumber 19
info currmove h2h3 currmovenumber 20
info depth 7 time 63550 nodes 28590 score cp 17 nps 449 pv e2e4 d7d5 e4d5 d8d5 b1c3 d5e6 g1e2 g8f6 d2d3 
bestmove e2e4 ponder d7d5

a lot of swap, do you have sttings for a more low memory friendly?.
(komodo loses on time )

Crafty:
new
level 60 60 0
go
end

results in:

Code: Select all

unable to open book file [./book.bin].
book is disabled
unable to open book file [./books.bin].

Initializing multiple threads.
System is SMP, not NUMA.

Crafty v23.3 JA (1 cpus)

White(1): White(1): 60 moves/3600 seconds primary time control
60 moves/3600 seconds secondary time control
White(1):               time surplus   0.00  time limit 1:00 (+0.00) (6:00)
              depth   time  score   variation (1)
               12     9.54  13/20*  1. c4      (21Knps)             
               12     9.60  14/20*  1. Nc3     (21Knps)             
               12     9.67  15/20*  1. d3      (21Knps)             
               12     9.71  16/20*  1. a3      (21Knps)             
               12     9.79  17/20*  1. Nh3     (21Knps)             
               12     9.84  18/20*  1. b4      (21Knps)             
               12     9.86  19/20*  1. c3      (21Knps)             
               12     9.88  20/20*  1. h3      (21Knps)             
               12->   9.95   0.13   1. Nf3 Nf6 2. Nc3 Nc6 3. e3 e6 4. Bd3
                                    Bd6 5. O-O O-O 6. Ng5 g6
               13     9.95   1/20*  1. Nf3     (21Knps)             
               13    14.55   0.13   1. Nf3 Nf6 2. Nc3 Nc6 3. e4 e6 4. Bb5
                                    Bb4 5. O-O O-O 6. d3 d5 7. Bd2 Bxc3
                                    8. Bxc3 dxe4 9. Bxc6 bxc6
               13    14.56   2/20*  1. d4      (21Knps)             
               13    16.53   2/20*  1. d4      (21Knps)             
               13    16.53   3/20*  1. e3      (21Knps)             
               13    24.63   0.20   1. e3 Nf6 2. Bd3 e5 3. Nc3 d5 4. Nh3
                                    Nc6 5. O-O Bxh3 6. gxh3 Bc5 7. Bf5
               13    24.64   4/20*  1. Na3     (23Knps)             
               13    25.29   5/20*  1. g4      (23Knps)             
               13    25.50   6/20*  1. f4      (22Knps)             
               13    25.54   7/20*  1. e4      (22Knps)             
               13    25.63   8/20*  1. h4      (22Knps)             
               13    25.67   9/20*  1. a4      (22Knps)             
               13    25.70  10/20*  1. g3      (22Knps)             
               13    25.73  11/20*  1. b3      (22Knps)             
               13    25.78  12/20*  1. Nc3     (22Knps)             
               13    25.83  13/20*  1. a3      (22Knps)             
               13    25.88  14/20*  1. h3      (22Knps)             
               13    25.91  15/20*  1. c4      (22Knps)             
               13    25.96  16/20*  1. Nh3     (22Knps)             
               13    25.98  17/20*  1. d3      (22Knps)             
               13    26.07  18/20*  1. f3      (22Knps)             
               13    26.08  19/20*  1. c3      (22Knps)             
               13    26.09  20/20*  1. b4      (22Knps)             
               13->  26.11   0.20   1. e3 Nf6 2. Bd3 e5 3. Nc3 d5 4. Nh3
                                    Nc6 5. O-O Bxh3 6. gxh3 Bc5 7. Bf5
               14    26.11   1/20*  1. e3      (22Knps)             
               14    37.05   0.05   1. e3 Nf6 2. Nc3 Nc6 3. Nf3 e6 4. Bd3
                                    Bd6 5. O-O O-O 6. Nb5 Nb4 7. Nfd4 Nfd5
               14    37.06   1/20*  1. e3      (22Knps)             
               14    37.06   2/20*  1. Nf3     (23Knps)             
               14    44.53   0.12   1. Nf3 Nf6 2. Nc3 Nc6 3. e4 d5 4. Bd3
                                    dxe4 5. Nxe4 Nb4 6. Bb5+ c6 7. Nxf6+
                                    exf6 8. Bc4 Qe7+ 9. Kf1
               14    44.54   3/20*  1. d4      (23Knps)             
               14    46.80   3/20*  1. d4      (23Knps)             
               14    48.83   4/20*  1. Na3     (23Knps)             
               14    50.83   5/20*  1. g4      (23Knps)             
               14    51.93   6/20*  1. e4      (23Knps)             
               14    53.04   7/20*  1. d3      (23Knps)             
               14    53.55   8/20*  1. Nc3     (23Knps)             
               14    54.22   9/20*  1. b3      (23Knps)             
               14    54.44  10/20*  1. c4      (23Knps)             
               14    54.84  11/20*  1. f4      (23Knps)             
               14    55.16  12/20*  1. a3      (23Knps)             
               14    55.44  13/20*  1. a4      (23Knps)             
               14    55.74  14/20*  1. h4      (23Knps)             
               14    56.03  15/20*  1. Nh3     (23Knps)             
               14    56.34  16/20*  1. h3      (23Knps)             
               14    56.57  17/20*  1. g3      (23Knps)             
               14    56.90  18/20*  1. b4      (23Knps)             
               14    57.04  19/20*  1. f3      (23Knps)             
               14    57.10  20/20*  1. c3      (23Knps)             
               14->  57.13   0.12   1. Nf3 Nf6 2. Nc3 Nc6 3. e4 d5 4. Bd3
                                    dxe4 5. Nxe4 Nb4 6. Bb5+ c6 7. Nxf6+
                                    exf6 8. Bc4 Qe7+ 9. Kf1
               15    57.14   1/20*  1. Nf3     (23Knps)             
               15     1:01   1/20*  1. Nf3     (23Knps)             
              time=1:01  mat=0  n=1445421  fh=91%  nps=23K
              extensions=3K qchecks=17K reduced=211K pruned=192K
              predicted=0  evals=757K  50move=0  EGTBprobes=0  hits=0
              SMP->  splits=0  aborts=0  data=0/65536  elap=1:01
 White(1): Nf3
              time used:   1:01
Black(1): Nf6 [pondering]
              time surplus   0.00  time limit 59.97 (+0.00) (5:59)
              depth   time  score   variation (13)
               13     8.83   8/22?  2. Ng5     (24Knps)             
               13     8.90   9/22?  2. Nd4     (24Knps)             
               13     8.96  10/22?  2. a4      (24Knps)             
               13     9.08  11/22?  2. a3      (24Knps)             
               13     9.22  12/22?  2. h3      (24Knps)             
               13     9.40  13/22?  2. g3      (24Knps)             
               13    11.33  14/22?  2. b3      (24Knps)             
               13    11.60  15/22?  2. h4      (24Knps)             
               13    11.74  16/22?  2. c4      (24Knps)             
               13    11.84  17/22?  2. c3      (23Knps)             
               13    12.00  18/22?  2. b4      (23Knps)             
               13    12.10  19/22?  2. Nh4     (23Knps)             
               13    12.15  20/22?  2. Ng1     (23Knps)             
               13    12.17  21/22?  2. e4      (23Knps)             
               13    12.40  22/22?  2. g4      (23Knps)             
               13->  12.48   0.38   2. Nc3 Nc6 3. e4 e6 4. e5 Ng4 5. d4
                                    Be7 6. h3 Nh6 7. d5 exd5 8. Bxh6 <HT>
               14    12.48   1/22?  2. Nc3     &#40;23Knps&#41;             
               14    17.10   1/22?  2. Nc3     &#40;23Knps&#41;             
Black&#40;1&#41;&#58; quit

I'm going to rest, more fun tomorrow.

[Don]

Try also DOSBox. It's very good.

But DosBox emulates the processor and is slow for this reason.

Dosemu is supposed to use hardware virtualization which means it runs at native speed. However it does not appear to run at native speed so far.

[Janzert]

Yeah, something is certainly messed up there. The page giving the processor details that each processor on that page links to shows the two releases as Nov. 1 1996 and Jan. 1 1997 with the second release being a die shrink and adding an additional, larger, L2 size.

Looking between the two pages the pentium 2 dates seem to agree better.

Janzert

[Dann Corbit]

Chess Genius 7 is included in this list:
http://www.husvankempen.de/nunn/40_40%2 ... liste.html
In particular:

Code: Select all

no Program Elo + - Games Score Av.Op. Draws 
1 Deep Rybka 4 x64 6CPU 3249 23 23 500 71.2% 3092 43.6% 
...
730 Genius 7 2409 25 25 528 40.5% 2475 31.8% 

and the SSDF has a number of entries for various Chess Genius versions:
http://ssdf.bosjo.net/rlwww101.txt

In particular:

Code: Select all

      THE SSDF RATING LIST 2010-03-21   %120316 games played by  311 computers
                                           Rating   +     -  Games   Won  Oppo
                                           ------  ---   --- -----   ---  ----
   1 Deep Rybka 3 x64  2GB Q6600 2,4 GHz     3227   27   -25  1005   83%  2962
...
 117 Genius 6.5  128MB  K6-2 450 MHz         2480   28   -28   605   48%  2493
 134 Genius 5.0 DOS  46MB P200 MMX           2395   20   -20  1197   49%  2400
 145 Genius 5.0 DOS Pentium 90 MHz           2333   18   -18  1558   47%  2353
 148 Genius 3.0 Pentium 90 MHz               2312   25   -24   824   60%  2241
 151 Genius 4.0 DOS Pentium 90 MHz           2306   23   -23   924   59%  2238
 161 Genius 4.0 DOS 486/50-66 MHz            2268   23   -23   900   49%  2272
 164 Genius 3.0 486/50-66 MHz                2256   24   -23   890   62%  2171
 184 Chess Genius 1.0  486/50-66 MHz         2183   23   -23   936   54%  2151
 191 Chess Genius 1.4 SX1  OMAP 310 120 MHz  2151   50   -48   210   60%  2081
 192 Chess Genius 1.0  486/33 MHz            2140   32   -32   483   50%  2143

[Mincho Georgiev]

This discussion remind me of how I played Sargon and that it was impossible for 6502 to reach ply 9 on this "platform":
http://www.pravetz.info/pravetz-8c.html
and earlier on this:
http://www.sofiaprint.com/pravetz8d/8DInt.JPG
I'm wondering, if this is the progress that hardware reaches for 20-25 years to the current stage, what it would be after another 20 years or so (that of course if not 2012's worst scenario is becoming real ) It is a very interesting subject, Don. If we think about what is the percentage of algorithms and refinements that we still use today in our programs, initially developed at that time (and even much earlier like alpha-beta) and compare that to the percentage old hardware components, used in today's platforms (I can only recall the fact that von Neumann architecture is still in use and probably some insignificant insides like 8253/8254 counter-timer chip and 8255 PPI chip, even if they are not in use anymore, just cannot be removed because some organizations defends that they are usable for people with illness (I agree with the last)), we can determine that hardware is developing fast and faster.
Sometimes I think that all the refinements and tiny optimizations we do (let's exclude the main algorithms that still works pretty well even 50 years later) is to compensate what current hardware state just can't accomplish undependable. Which suggests, that even today, maybe the hardware's room for improvement is bigger than Empire State Building. And that is a good thing, the way I see it, otherwise if we had the hardware for it (solving chess), probably our job/hobby is done.
By the way, there is a guy at FICS that plays with Toga and Robbo, installed on router. It's very interesting (the poor device probably have serious headache while calculating bitboards) and they are arround 2100-2200 ELO.
The conclusion? The imperfections in hardware and software makes both to develop and that is a good thing. One of them could grow faster, but they work in combination so people like you still are and will be usable for that process.
Just my 2 cents...

[Don]

ok.

from the position
r2qkbnr/1ppb1ppp/p1np4/1B2p3/3PP3/2N2N2/PPP2PPP/R1BQK2R w KQkq - 0 6

chess genius play Ab5xc6 after 2:44 with depth 7
line b5xc6 d7xc6 d4xe5 f7f5 c1g5 f8e7 g5e7

Ok, I got freedos working on a usb stick so I can run genius. I tried your position and it responds quickly. I really need you to run something longer because it only takes 3 seconds so the resolution is not going to be very good.

I assume you are going by the PV time which is on the far right - this is not the time it takes to complete the search but apparently it is a time that is updated when the PV is calculated.

If it took 2:44 to do the 7 ply PV on your machine and it's taking 3 seconds on mine, then the hardware is about 55 to 1. This is on my core 2 laptop but it would be about twice that fast on my i7. But the resolution is poor because you did not run a long enough timing.

Let me suggest this test. start the program and do things in this exact order:

1. set level to 9 ply fixed depth
2. turn OFF permanent brain (otherwise it starts thinking right away)
3. turn OFF book
4. start the search with computer (or altC)

I have found that I don't get predictable times or PV's. Even when I restart the program from scratch.

So I suggest you run this 2 or 3 times if you have the patience and we can go by the median value, or we can average them or something.

It should take you at least 15 minutes to run this as it takes me about 18 seconds which is the average of 3 tries.

I tried to run rebel 10 but I cannot make it run. It runs in the emulator but not in a real OS.

now modern programs.
let start with Komodo.
(Jim A. compile 32 bits)

commands
uci
position startpos
go movetime 60000
quit

Code: Select all

id name Komodo32 1.2 JA
id author Don Dailey, Larry Kaufman
option name Ponder type check default true
option name Hash type spin default 32 min 0 max 4096
option name OwnBook type check default false
option name MultiPV type spin default 1 min 1 max 99
option name Book File type string default komodo.bin
uciok
info depth 1
info depth 1 time 380 nodes 20 score cp 63 nps 52 pv e2e4 
info depth 2
info depth 2 time 770 nodes 160 score cp 9 nps 207 pv e2e4 e7e5 
info depth 3
info depth 3 time 1760 nodes 558 score cp 45 nps 317 pv e2e4 e7e5 d2d4 
info depth 4
info depth 4 time 4940 nodes 1943 score cp 12 nps 393 pv g1f3 e7e6 d2d4 b8c6 
info depth 5
info depth 5 time 7579 nodes 2865 score cp 33 nps 377 pv g1f3 d7d5 d2d3 b8c6 b1c3 
info depth 5 time 8899 nodes 3682 score cp 33 nps 413 pv g1f3 d7d5 d2d3 b8c6 b1c3 
info depth 6
info currmove g1f3 currmovenumber 1
info depth 6 time 11420 nodes 5131 score cp 9 nps 449 pv g1f3 d7d5 d2d4 g8f6 b1c3 b8c6 
info currmove e2e4 currmovenumber 2
info depth 6 time 21690 nodes 8865 score cp 14 nps 408 pv e2e4 d7d5 d1f3 g8f6 b1c3 d5e4 c3e4 
info currmove d2d4 currmovenumber 3
info currmove e2e3 currmovenumber 4
info currmove b1c3 currmovenumber 5
info currmove d2d3 currmovenumber 6
info currmove b1a3 currmovenumber 7
info currmove g1h3 currmovenumber 8
info currmove b2b4 currmovenumber 9
info currmove b2b3 currmovenumber 10
info currmove g2g3 currmovenumber 11
info currmove g2g4 currmovenumber 12
info currmove c2c4 currmovenumber 13
info currmove a2a4 currmovenumber 14
info currmove h2h4 currmovenumber 15
info currmove f2f4 currmovenumber 16
info currmove c2c3 currmovenumber 17
info currmove f2f3 currmovenumber 18
info currmove a2a3 currmovenumber 19
info currmove h2h3 currmovenumber 20
info depth 6 time 25649 nodes 10751 score cp 14 nps 419 pv e2e4 d7d5 d1f3 g8f6 b1c3 d5e4 c3e4 
info depth 7
info currmove e2e4 currmovenumber 1
info depth 7 time 52010 nodes 22528 score cp 17 nps 433 pv e2e4 d7d5 e4d5 d8d5 b1c3 d5e6 g1e2 g8f6 d2d3 
info currmove g1f3 currmovenumber 2
info currmove d2d4 currmovenumber 3
info currmove e2e3 currmovenumber 4
info currmove b1c3 currmovenumber 5
info currmove d2d3 currmovenumber 6
info currmove b1a3 currmovenumber 7
info currmove g1h3 currmovenumber 8
info currmove b2b4 currmovenumber 9
info currmove b2b3 currmovenumber 10
info currmove g2g3 currmovenumber 11
info currmove g2g4 currmovenumber 12
info currmove c2c4 currmovenumber 13
info currmove a2a4 currmovenumber 14
info currmove h2h4 currmovenumber 15
info currmove f2f4 currmovenumber 16
info currmove c2c3 currmovenumber 17
info currmove f2f3 currmovenumber 18
info currmove a2a3 currmovenumber 19
info currmove h2h3 currmovenumber 20
info depth 7 time 63550 nodes 28590 score cp 17 nps 449 pv e2e4 d7d5 e4d5 d8d5 b1c3 d5e6 g1e2 g8f6 d2d3 
bestmove e2e4 ponder d7d5

a lot of swap, do you have sttings for a more low memory friendly?.
(komodo loses on time )

Crafty:
new
level 60 60 0
go
end

results in:

Code: Select all

unable to open book file [./book.bin].
book is disabled
unable to open book file [./books.bin].

Initializing multiple threads.
System is SMP, not NUMA.

Crafty v23.3 JA (1 cpus)

White(1): White(1): 60 moves/3600 seconds primary time control
60 moves/3600 seconds secondary time control
White(1):               time surplus   0.00  time limit 1:00 (+0.00) (6:00)
              depth   time  score   variation (1)
               12     9.54  13/20*  1. c4      (21Knps)             
               12     9.60  14/20*  1. Nc3     (21Knps)             
               12     9.67  15/20*  1. d3      (21Knps)             
               12     9.71  16/20*  1. a3      (21Knps)             
               12     9.79  17/20*  1. Nh3     (21Knps)             
               12     9.84  18/20*  1. b4      (21Knps)             
               12     9.86  19/20*  1. c3      (21Knps)             
               12     9.88  20/20*  1. h3      (21Knps)             
               12->   9.95   0.13   1. Nf3 Nf6 2. Nc3 Nc6 3. e3 e6 4. Bd3
                                    Bd6 5. O-O O-O 6. Ng5 g6
               13     9.95   1/20*  1. Nf3     (21Knps)             
               13    14.55   0.13   1. Nf3 Nf6 2. Nc3 Nc6 3. e4 e6 4. Bb5
                                    Bb4 5. O-O O-O 6. d3 d5 7. Bd2 Bxc3
                                    8. Bxc3 dxe4 9. Bxc6 bxc6
               13    14.56   2/20*  1. d4      (21Knps)             
               13    16.53   2/20*  1. d4      (21Knps)             
               13    16.53   3/20*  1. e3      (21Knps)             
               13    24.63   0.20   1. e3 Nf6 2. Bd3 e5 3. Nc3 d5 4. Nh3
                                    Nc6 5. O-O Bxh3 6. gxh3 Bc5 7. Bf5
               13    24.64   4/20*  1. Na3     (23Knps)             
               13    25.29   5/20*  1. g4      (23Knps)             
               13    25.50   6/20*  1. f4      (22Knps)             
               13    25.54   7/20*  1. e4      (22Knps)             
               13    25.63   8/20*  1. h4      (22Knps)             
               13    25.67   9/20*  1. a4      (22Knps)             
               13    25.70  10/20*  1. g3      (22Knps)             
               13    25.73  11/20*  1. b3      (22Knps)             
               13    25.78  12/20*  1. Nc3     (22Knps)             
               13    25.83  13/20*  1. a3      (22Knps)             
               13    25.88  14/20*  1. h3      (22Knps)             
               13    25.91  15/20*  1. c4      (22Knps)             
               13    25.96  16/20*  1. Nh3     (22Knps)             
               13    25.98  17/20*  1. d3      (22Knps)             
               13    26.07  18/20*  1. f3      (22Knps)             
               13    26.08  19/20*  1. c3      (22Knps)             
               13    26.09  20/20*  1. b4      (22Knps)             
               13->  26.11   0.20   1. e3 Nf6 2. Bd3 e5 3. Nc3 d5 4. Nh3
                                    Nc6 5. O-O Bxh3 6. gxh3 Bc5 7. Bf5
               14    26.11   1/20*  1. e3      (22Knps)             
               14    37.05   0.05   1. e3 Nf6 2. Nc3 Nc6 3. Nf3 e6 4. Bd3
                                    Bd6 5. O-O O-O 6. Nb5 Nb4 7. Nfd4 Nfd5
               14    37.06   1/20*  1. e3      (22Knps)             
               14    37.06   2/20*  1. Nf3     (23Knps)             
               14    44.53   0.12   1. Nf3 Nf6 2. Nc3 Nc6 3. e4 d5 4. Bd3
                                    dxe4 5. Nxe4 Nb4 6. Bb5+ c6 7. Nxf6+
                                    exf6 8. Bc4 Qe7+ 9. Kf1
               14    44.54   3/20*  1. d4      (23Knps)             
               14    46.80   3/20*  1. d4      (23Knps)             
               14    48.83   4/20*  1. Na3     (23Knps)             
               14    50.83   5/20*  1. g4      (23Knps)             
               14    51.93   6/20*  1. e4      (23Knps)             
               14    53.04   7/20*  1. d3      (23Knps)             
               14    53.55   8/20*  1. Nc3     (23Knps)             
               14    54.22   9/20*  1. b3      (23Knps)             
               14    54.44  10/20*  1. c4      (23Knps)             
               14    54.84  11/20*  1. f4      (23Knps)             
               14    55.16  12/20*  1. a3      (23Knps)             
               14    55.44  13/20*  1. a4      (23Knps)             
               14    55.74  14/20*  1. h4      (23Knps)             
               14    56.03  15/20*  1. Nh3     (23Knps)             
               14    56.34  16/20*  1. h3      (23Knps)             
               14    56.57  17/20*  1. g3      (23Knps)             
               14    56.90  18/20*  1. b4      (23Knps)             
               14    57.04  19/20*  1. f3      (23Knps)             
               14    57.10  20/20*  1. c3      (23Knps)             
               14->  57.13   0.12   1. Nf3 Nf6 2. Nc3 Nc6 3. e4 d5 4. Bd3
                                    dxe4 5. Nxe4 Nb4 6. Bb5+ c6 7. Nxf6+
                                    exf6 8. Bc4 Qe7+ 9. Kf1
               15    57.14   1/20*  1. Nf3     (23Knps)             
               15     1:01   1/20*  1. Nf3     (23Knps)             
              time=1:01  mat=0  n=1445421  fh=91%  nps=23K
              extensions=3K qchecks=17K reduced=211K pruned=192K
              predicted=0  evals=757K  50move=0  EGTBprobes=0  hits=0
              SMP->  splits=0  aborts=0  data=0/65536  elap=1:01
 White(1): Nf3
              time used:   1:01
Black(1): Nf6 [pondering]
              time surplus   0.00  time limit 59.97 (+0.00) (5:59)
              depth   time  score   variation (13)
               13     8.83   8/22?  2. Ng5     (24Knps)             
               13     8.90   9/22?  2. Nd4     (24Knps)             
               13     8.96  10/22?  2. a4      (24Knps)             
               13     9.08  11/22?  2. a3      (24Knps)             
               13     9.22  12/22?  2. h3      (24Knps)             
               13     9.40  13/22?  2. g3      (24Knps)             
               13    11.33  14/22?  2. b3      (24Knps)             
               13    11.60  15/22?  2. h4      (24Knps)             
               13    11.74  16/22?  2. c4      (24Knps)             
               13    11.84  17/22?  2. c3      (23Knps)             
               13    12.00  18/22?  2. b4      (23Knps)             
               13    12.10  19/22?  2. Nh4     (23Knps)             
               13    12.15  20/22?  2. Ng1     (23Knps)             
               13    12.17  21/22?  2. e4      (23Knps)             
               13    12.40  22/22?  2. g4      (23Knps)             
               13->  12.48   0.38   2. Nc3 Nc6 3. e4 e6 4. e5 Ng4 5. d4
                                    Be7 6. h3 Nh6 7. d5 exd5 8. Bxh6 <HT>
               14    12.48   1/22?  2. Nc3     &#40;23Knps&#41;             
               14    17.10   1/22?  2. Nc3     &#40;23Knps&#41;             
Black&#40;1&#41;&#58; quit

I'm going to rest, more fun tomorrow.

[Don]

Finally I got a version of DOS working that boot via a usb memory stick that can run the Rebel 10 program, so I have accurate numbers.

Here is the scoop:

The Rebel built in benchmark returns 1,5 and 9 seconds for the 3 benchmarks and gives an estimated ELO of 2596.

ON the P90 results page they are solved in:

1:20 5:11 and 8:07

The speedup factor therefore is:

80, 62 and 54

This was roughly in agreement with crude chess genius timing test we attempted to do.

I'm using a core 2 duo laptop to run this test, so this does not quite represent the state of the art. I will check but I think one i7 core is about 2x the speed of my notebook at 1 core.

So I do not believe the P90 is nearly 1000 to 1 slower than even my fast i7-980x which is fastest clocked i7 and is a real screamer.

Nevertheless, this is still very far from clear when you consider a bunch of other factors:

1. Rebel was DESIGNED to run well on the P90, not the i7.

2. Rebel has 4 meg hash table by default when running this test.

3. We are not considering the 6 cores in the i7.

4. Complier technology was inferior.


Don

I'm pretty sure the Pentium 90 was the fastest processor at that time. Intel introduced the Pentium Pro family (to which the P200 belongs) in late 1995.

Yes, I'll assume that circa 1994 CG3 was running on the P90.

Any idea of the speed? I don't suppose anybody has a Pentium 90 laying around do they?

Don

This could be fun... I got REBEL 10 from this page and I'm now trying to get it to run in VMWare in order to compare with these benchmarks.

When did Rebel 10 come out?

I'm running CG3 in dosemu which is said to be native speed, but I'm going to check that out for sure.

1998 it seems. I only paid any attention to it because there's a Pentium 90 MHz benchmark for it.

VMWare is a pain in the ass to use with DOS, so I may try out Dosemu as well.

[rbarreira]

Interesting!

I just got MS-DOS 6.22 running from a CD, and I can access a USB pen drive as C:\ (thanks to an option in the BIOS). But when I tried running Rebel 10 it gave me an "insufficient memory" error. (4 GB not enough for you, eh? )

Your speedup factors make sense considering it's only using one core and that it's running in 32-bit mode. If you multiply your speedup of 80 by 6 and then by 2 for those two factors, that gives 960 which is pretty much spot-on with Moore's law!

[Don]

Interesting!

I just got MS-DOS 6.22 running from a CD, and I can access a USB pen drive as C:\ (thanks to an option in the BIOS). But when I tried running Rebel 10 it gave me an "insufficient memory" error. (4 GB not enough for you, eh? )

Your speedup factors make sense considering it's only using one core and that it's running in 32-bit mode. If you multiply your speedup of 80 by 6 and then by 2 for those two factors, that gives 960 which is pretty much spot-on with Moore's law!

I'm not going to adjust for cores as it makes the match almost impossible to conduct - I cannot give super heavy time handicap and hope to play lots of games. So we can basically interpret the results later and anyone who wants to is free to interpolate.

I don't see how 32 bit mode is a factor. Some of the very best programs today still run in 32 bits and even 64 bits programs do not deserve a double for this.

If this test is actually a contest where each side negotiates every possible advantage, I want no part of it. I want to see how well modern programs would run on older hardware and visa versa. I'm not trying to prove a point and even though I have been going round and round with Bob about this I don't care if I end up being wrong. Doubling the numbers for 64 bit and multiplying by 6 for a 6 core machine is not hardly going to give a realistic picture unless you just want to make up numbers to make the test come out a certain way and that's not what I am doing this for (if I actually do it that is

I plan on using a 100 to 1 factor when testing genius because the benchmark for a single core shows right around 100 to 1. We could argue over all kinds of ways to adjust up and down to compensate for this and that, but we probably would not agree on it anyway. This will be test under known conditions that people can fee free to interpret any way they choose, but at least they will know the exact conditions of the test.

This test is going to be kind of like the computer language benchmarks - it won't prove much - it will just show how things worked out give a very specific test setup and will serve as a jumping off point for other tests if anyone has the patience.

I though of another interesting test setup. Suppose you actually had a Pentium 90 and could run Rybka or Robboltto on it (in 32 bit mode no doubt.) You probably COULD if you installed Linux on one of these old machines and used robbolitto, stockfish or something really good.

Then what you do is play Rybka on the old hardware in the best configuration you can find for it and Chess Genius on the new hardware using the best configuration you can find. At even time controls you have something that is (at least in principle) a fair match. Can Rybka's superior software overcome Genius on superior hardware?

[rbarreira]

Interesting!

I just got MS-DOS 6.22 running from a CD, and I can access a USB pen drive as C:\ (thanks to an option in the BIOS). But when I tried running Rebel 10 it gave me an "insufficient memory" error. (4 GB not enough for you, eh? )

Your speedup factors make sense considering it's only using one core and that it's running in 32-bit mode. If you multiply your speedup of 80 by 6 and then by 2 for those two factors, that gives 960 which is pretty much spot-on with Moore's law!

I'm not going to adjust for cores as it makes the match almost impossible to conduct - I cannot give super heavy time handicap and hope to play lots of games. So we can basically interpret the results later and anyone who wants to is free to interpolate.

I don't see how 32 bit mode is a factor. Some of the very best programs today still run in 32 bits and even 64 bits programs do not deserve a double for this.

If this test is actually a contest where each side negotiates every possible advantage, I want no part of it. I want to see how well modern programs would run on older hardware and visa versa. I'm not trying to prove a point and even though I have been going round and round with Bob about this I don't care if I end up being wrong. Doubling the numbers for 64 bit and multiplying by 6 for a 6 core machine is not hardly going to give a realistic picture unless you just want to make up numbers to make the test come out a certain way and that's not what I am doing this for (if I actually do it that is

I plan on using a 100 to 1 factor when testing genius because the benchmark for a single core shows right around 100 to 1. We could argue over all kinds of ways to adjust up and down to compensate for this and that, but we probably would not agree on it anyway. This will be test under known conditions that people can fee free to interpret any way they choose, but at least they will know the exact conditions of the test.

This test is going to be kind of like the computer language benchmarks - it won't prove much - it will just show how things worked out give a very specific test setup and will serve as a jumping off point for other tests if anyone has the patience.

I though of another interesting test setup. Suppose you actually had a Pentium 90 and could run Rybka or Robboltto on it (in 32 bit mode no doubt.) You probably COULD if you installed Linux on one of these old machines and used robbolitto, stockfish or something really good.

Then what you do is play Rybka on the old hardware in the best configuration you can find for it and Chess Genius on the new hardware using the best configuration you can find. At even time controls you have something that is (at least in principle) a fair match. Can Rybka's superior software overcome Genius on superior hardware?

My point was not that you need to adjust your testing for these factors. I was just saying that the speedup makes sense according to Moore's law. If it was 1000x faster on a single core that would be unbelievable (in the other thread I didn't realize you were talking about single core comparisons).