Phantom: everything about it finally known

mclane · Post by **mclane** » Tue Jul 10, 2007 5:06 pm

yes. but it could e.g. be damaged and as a result of that the RAM could be less.

of course ALL modules come at least with 1 MB . but if those modules are broken anything can be possible. i find the 11' in relation to the 15' not very much faster.

Steve B · Post by **Steve B** » Tue Jul 10, 2007 5:15 pm

mclane wrote:yes. but it could e.g. be damaged and as a result of that the RAM could be less.

of course ALL modules come at least with 1 MB . but if those modules are broken anything can be possible. i find the 11' in relation to the 15' not very much faster.

OK Thor
i test both Sparcs this weekend and post here

Lot of Work Regards
Steve

Steve B · Post by **Steve B** » Fri Jul 13, 2007 3:25 am

Steve B wrote:

mclane wrote:i see.

btw: can anybody with a sparc 4MB make a run for the following test position:

http://www.schachcomputer.info/forum/sh ... php?t=1352

[d] 2rr2k1/1b3ppp/pb2p3/1p2P3/1P2BPnq/P1N3P1/1B2Q2P/R4R1K b - - 0 22

Code: Select all


Super Forte B 5mhz   sel.7                           3'40"
                     sel.3                           7'54"
Saitek Sparc                                        11'54"
Fidelity MachIII Master                             15'56"
Analyst B 8Mhz                                      23'
ColossusChessX (atari ST) sel.                      32'32"
Mephisto Mirage                                     35'
Sphinx Dominator                                    38'
Simultano                                           44'
Supermondial                                        45'
Mephisto II                                         65'
Turbostar432                                        90'
Mephisto IIIS (16bit, 12mhz)                   ca.2h30'
ParExcellence                                  ca.6h
Mondial                                          18h30'
Constellation                                    22h
MMIV                                             23h30'
Mephisto III (8bit)                             nicht in 42h
Ameria 16Bit                                    nicht in 1h
Academy                                         nicht in 1h
Excell 68000                                    nicht in 1h
Super Forte                                     nicht in 1h

i can test the 4mb this coming weekend..not before..sorry

what is the winning move?
BxB??

Steve

Thor

you will be happy to learn there is nothing wrong with your Sparc module..
as promised..

Sparc 1-MB Finds Rxc3 in 11'54" .. 8 ply
Sparc 4-MB Finds Rxc3 in 07'44".. 8 ply
Testing Regards
Steve

mclane · Post by **mclane** » Fri Jul 13, 2007 8:56 am

thank you. so the ranking is now

Code: Select all

Novag Star Diamond ext.ON                               9"
Mephisto Magellan                                    1'33"
Novag Star Diamond                                   2'15"
Novag Super Forte B 5mhz   sel.7                     3'40"
Saitek Sparc 4MB                                     7'44"
Novag Super Forte B 5mhz   sel.3                     7'54"
Mephisto London 68020                               10'15" 
Saitek Sparc 1MB                                    11'54"
Fidelity MachIII Master                             15'56"
Vancouver 68000                                     17'12" 
Analyst B 8Mhz                                      23'
MM5 18mhz                                           30'15"
ColossusChessX (atari ST) sel.                      32'32"
Mephisto Mirage                                     35'
Sphinx Dominator                                    38'
Simultano                                           44'
Supermondial                                        45'
RISC I                                              46'45" 
Mephisto II                                         65'
Turbostar432                                        90'
Mephisto IIIS (16bit, 12mhz)                   ca.2h30'
ParExcellence                                  ca.6h
Mondial                                          18h30'
Constellation                                    22h
MMIV                                             23h30'
Mephisto III (8bit)                             nicht in 42h
Almeria 16Bit                                   nicht in 1h
Academy                                         nicht in 1h
Excell 68000                                    nicht in 1h
Super Forte                                     nicht in 1h

Terry McCracken · Post by **Terry McCracken** » Fri Jul 13, 2007 9:24 am

Steve B wrote:
Steve B wrote:
mclane wrote:i see.

btw: can anybody with a sparc 4MB make a run for the following test position:

http://www.schachcomputer.info/forum/sh ... php?t=1352

[d] 2rr2k1/1b3ppp/pb2p3/1p2P3/1P2BPnq/P1N3P1/1B2Q2P/R4R1K b - - 0 22
Code: Select all
Super Forte B 5mhz   sel.7                           3'40"
                     sel.3                           7'54"
Saitek Sparc                                        11'54"
Fidelity MachIII Master                             15'56"
Analyst B 8Mhz                                      23'
ColossusChessX (atari ST) sel.                      32'32"
Mephisto Mirage                                     35'
Sphinx Dominator                                    38'
Simultano                                           44'
Supermondial                                        45'
Mephisto II                                         65'
Turbostar432                                        90'
Mephisto IIIS (16bit, 12mhz)                   ca.2h30'
ParExcellence                                  ca.6h
Mondial                                          18h30'
Constellation                                    22h
MMIV                                             23h30'
Mephisto III (8bit)                             nicht in 42h
Ameria 16Bit                                    nicht in 1h
Academy                                         nicht in 1h
Excell 68000                                    nicht in 1h
Super Forte                                     nicht in 1h
i can test the 4mb this coming weekend..not before..sorry

what is the winning move?
BxB??

Steve
Thor

you will be happy to learn there is nothing wrong with your Sparc module..
as promised..

Sparc 1-MB Finds Rxc3 in 11'54" .. 8 ply
Sparc 4-MB Finds Rxc3 in 07'44".. 8 ply
Testing Regards
Steve

Ah...One of my favorite positions of all time due to its beauty!

Rubenstein Sends his Regards!

Terry McCracken · Post by **Terry McCracken** » Fri Jul 13, 2007 9:32 am

mclane wrote:thank you. so the ranking is now

Code: Select all

Novag Star Diamond ext.ON                               9"
Mephisto Magellan                                    1'33"
Novag Star Diamond                                   2'15"
Novag Super Forte B 5mhz   sel.7                     3'40"
Saitek Sparc 4MB                                     7'44"
Novag Super Forte B 5mhz   sel.3                     7'54"
Mephisto London 68020                               10'15" 
Saitek Sparc 1MB                                    11'54"
Fidelity MachIII Master                             15'56"
Vancouver 68000                                     17'12" 
Analyst B 8Mhz                                      23'
MM5 18mhz                                           30'15"
ColossusChessX (atari ST) sel.                      32'32"
Mephisto Mirage                                     35'
Sphinx Dominator                                    38'
Simultano                                           44'
Supermondial                                        45'
RISC I                                              46'45" 
Mephisto II                                         65'
Turbostar432                                        90'
Mephisto IIIS (16bit, 12mhz)                   ca.2h30'
ParExcellence                                  ca.6h
Mondial                                          18h30'
Constellation                                    22h
MMIV                                             23h30'
Mephisto III (8bit)                             nicht in 42h
Almeria 16Bit                                   nicht in 1h
Academy                                         nicht in 1h
Excell 68000                                    nicht in 1h
Super Forte                                     nicht in 1h

Hmmm....No one tried the Novag Super Forte C? It's about 8 min. on select 5 but maybe faster on a lower or higher selective search...still not bad for an old slow machine

The Scorpio should be tested too, it should be fairly fast on this position.

Selective and Dedicated Regards,
Terry

Robert Weck · Post by **Robert Weck** » Fri Jul 13, 2007 12:54 pm

Steve B wrote: Sparc 1-MB Finds Rxc3 in 11'54" .. 8 ply
Sparc 4-MB Finds Rxc3 in 07'44".. 8 ply

this big difference surprises me a little bit!

a matter of the long thinking time?

But in the BT2450 test suite the difference between the 512KB and 2 MB Risc 2500 is marginal...

surprising Sparc regards,
Robert

mclane · Post by **mclane** » Fri Jul 13, 2007 1:01 pm

maybe it has to do with the position.
in schachcomputer.info we had also BIG differences when somebody posted the results of his mobile with hiarcs installed.
it seems hiarcs also reacts sensible here with different hash sizes:

hiarcs with hash

Code: Select all

4096kb Hash Tables
1 sec: -3,74 (HV 1.-Txc3 2. Lxb7 Txg3 3. Tf3)
7 sec: -4,01
9 sec: -4,13
17 sec: -4,63
26 sec: -6,58 (HV 1.-Txc3 2. Lxb7 Txg3 3. Tac1)
42 sec: -6,26
51 sec: -6,21
97 sec: -6,53
148 sec: -7,20

hiarcs without hash

Code: Select all

 1 sec: -1,8 (HV 1.-Lxe4+ 2. Sxe4 Dxh2+)
4 sec: -1,71
5 sec: -1,72 (HV 1.-Txc3)
7 sec: -3,00
13 sec: -0,96 (HV 1.-Lxe4+)
24 sec: -1,83 (HV 1.-Dxh2+ 2. Dxh2 Lxe4+)
32 sec: -1,94
39 sec: -1,95 (HV 1.-Txc3)
44 sec: -2,15
50 sec: -2,27 (HV 1.-Txc3 2. gxh4 Td2)
75 sec: -2,6
107 sec: -2,59
177 sec: -2.92
275 sec: -6,35 (HV 1.-Txc3 2. Lxb7 Txg3)

software/ hardware was
Hiarcs 9.6 / Treo 650 (312 Mhz Intel XScale)

i did not put it into the table because it is no dedicated chess computer
but a mobile

Robert Weck · Post by **Robert Weck** » Fri Jul 13, 2007 1:46 pm

mclane wrote:maybe it has to do with the position.
in schachcomputer.info we had also BIG differences when somebody posted the results of his mobile with hiarcs installed.
it seems hiarcs also reacts sensible here with different hash sizes:

hiarcs with hash
Code: Select all
4096kb Hash Tables
1 sec: -3,74 (HV 1.-Txc3 2. Lxb7 Txg3 3. Tf3)
7 sec: -4,01
9 sec: -4,13
17 sec: -4,63
26 sec: -6,58 (HV 1.-Txc3 2. Lxb7 Txg3 3. Tac1)
42 sec: -6,26
51 sec: -6,21
97 sec: -6,53
148 sec: -7,20
hiarcs without hash
Code: Select all
 1 sec: -1,8 (HV 1.-Lxe4+ 2. Sxe4 Dxh2+)
4 sec: -1,71
5 sec: -1,72 (HV 1.-Txc3)
7 sec: -3,00
13 sec: -0,96 (HV 1.-Lxe4+)
24 sec: -1,83 (HV 1.-Dxh2+ 2. Dxh2 Lxe4+)
32 sec: -1,94
39 sec: -1,95 (HV 1.-Txc3)
44 sec: -2,15
50 sec: -2,27 (HV 1.-Txc3 2. gxh4 Td2)
75 sec: -2,6
107 sec: -2,59
177 sec: -2.92
275 sec: -6,35 (HV 1.-Txc3 2. Lxb7 Txg3)
software/ hardware was
Hiarcs 9.6 / Treo 650 (312 Mhz Intel XScale)

i did not put it into the table because it is no dedicated chess computer
but a mobile

yes, but this was 4MB compared to no hash!

In the posting on .info it was also said, that even with just 16 KB hash there was no big difference to 4 MB!

Robert

mclane · Post by **mclane** » Fri Jul 13, 2007 2:20 pm

we don't know what strange mysterious things the spracklen have realized in their SPARC module

maybe a time machine ?

Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known

Re: Phantom: everything about it finally known