TalkChess.com

geots wrote:But instead just run the exes and in console mode type in "go depth 22" and it will run thru 22 depths.

Codesquid wrote:

geots wrote:But instead just run the exes and in console mode type in "go depth 22" and it will run thru 22 depths.

That's an UCI command, so you'll either need to pass uci as argument to the executable or instead send uci as first line before you can issue further UCI commands.

Reason is that Octochess defaults to CECP unless instructed otherwise.

Jim Ablett wrote:

Look, it can't be true that I am ok either way with either his download or yours. Speaking of 64bit now. One of the 2 I will be better off using. No time to be modest. I really need to know. Which?

Hi George,

I didn't test my 64 bit bit version against Tim's this time, I was preoccupied just trying to a 32 bit version working.
I would load them both up and see which one has the highest knps and is the most stable & use that one.
Mine maybe slightly slower than normal as I just noticed I used my 32-bit edited source for the 64 bit compile.
This uses 'C' bitscan routines instead of 'Asm' which is a tad slower. Would be interesting to see if it is still faster than Tim's.

Jim.

Jim Ablett wrote:I played some test games against Nejmet and it seems of similar strength.

Jim.

Octochess r4558 x64 - Nejmet 3.07 : 22,0/27 19-2-6 (11=1111===111111111=110=110)  81%  +252

Werner wrote:

Jim Ablett wrote:I played some test games against Nejmet and it seems of similar strength.

Jim.

I have played some games too:

Code: Select all

Octochess r4558 x64 - Nejmet 3.07 : 22,0/27 19-2-6 (11=1111===111111111=110=110)  81%  +252

and I see here a quite different result

geots wrote:First, how did this version start off as 4529 and end up as 4558 after a few threads?

I really would like to run this engine- but I would never use a 1GB table just on principle.

octochess.exe --hash 256

Before I attempt to load the 64bit, could either Tim or Jim explain what the engine is going to do with 3GB of my RAM memory.

Codesquid wrote:

geots wrote:First, how did this version start off as 4529 and end up as 4558 after a few threads?

Looks like Jim built a development version straight out of Octochess' source repository. As it's not an officially released version, beware of bugs as I haven't done any QA on that particular revision.

I really would like to run this engine- but I would never use a 1GB table just on principle.

You can use the --hash argument to specify a different hash table size. Acceptable values are power of two and in megabytes.

Example for 256MB hash:

Code: Select all

octochess.exe --hash 256

Before I attempt to load the 64bit, could either Tim or Jim explain what the engine is going to do with 3GB of my RAM memory.

One gigantic hash table.

Codesquid wrote: Example for 256MB hash:

Code: Select all

octochess.exe --hash 256

Werner wrote:

Codesquid wrote: Example for 256MB hash:

Code: Select all

octochess.exe --hash 256

...can you please give an example how this works in Arena??

... and why does the engine sometimes in an endgame use more than 1CPU? I sometimes saw 100 % cpu usage on my quad here??

Codesquid wrote:

Werner wrote:

Codesquid wrote: Example for 256MB hash:

Code: Select all

octochess.exe --hash 256

...can you please give an example how this works in Arena??

I can't, I'm not familiar with Arena.

... and why does the engine sometimes in an endgame use more than 1CPU? I sometimes saw 100 % cpu usage on my quad here??

Octochess is multi-threaded. First root node is searched single-threaded, the other root nodes are searched in parallel. You can control the number of search threads with the --threads argument.