geots wrote:It would be almost impossible for me to assume I am the only one who sees Nemo stealing half it's opponents time- leaving him with only 50% of his allotted time. What can I say- I don't guess I will be able to use it.
george
This is correct!
Nemo uses 1CPU in a eng-eng match, ponder off even when the engine is idle. It does not ponder. I saw this last together with The King.
...and it looks extremly strong for a new engine (2800??)
Werner
!! Many thanks to Sven Schüle supporting the Nemo Project !! and for being Part of the Nemo-Team
* Also thanks to the team of chessprogamming wiki for the excellent information provided.
* Thx to the CCC members, who always provide some interesting discussions and topics in
the programming section. I am learning a lot from you, at this place.
* Thx to all other people, who supported me in one or the other way while developing Nemo.(Thx HGM).
* Thank you Jim Ablett for hosting Nemo!
Enjoy !
Looking forward to trying it out. Thanks!
Hi, Graham. I am happy to hear it will complement your large engine
collection.
Michael
It would be almost impossible for me to assume I am the only one who sees Nemo stealing half it's opponents time- leaving him with only 50% of his allotted time. What can I say- I don't guess I will be able to use it.
george
Hello George,
what do you mean with steal opponents time ?
Nemo definitely steals nothing.
By the way, we are already analyzing potential problems with the
memory allocation. Because you are so far the only one reporting
a problem i will contact you later by pm, if that is ok ?
Maybe with your help we are able to catch sth much faster.
Werner wrote:
Nemo uses 1CPU in a eng-eng match, ponder off even when the engine is idle. It does not ponder. I saw this last together with The King.
...and it looks extremly strong for a new engine (2800??)
Werner
... a "Sleep(1);" in the waiting loop in source code usually fixes this under Windows
!! Many thanks to Sven Schüle supporting the Nemo Project !! and for being Part of the Nemo-Team
* Also thanks to the team of chessprogamming wiki for the excellent information provided.
* Thx to the CCC members, who always provide some interesting discussions and topics in
the programming section. I am learning a lot from you, at this place.
* Thx to all other people, who supported me in one or the other way while developing Nemo.(Thx HGM).
* Thank you Jim Ablett for hosting Nemo!
Enjoy !
Looking forward to trying it out. Thanks!
Hi, Graham. I am happy to hear it will complement your large engine
collection.
Michael
It would be almost impossible for me to assume I am the only one who sees Nemo stealing half it's opponents time- leaving him with only 50% of his allotted time. What can I say- I don't guess I will be able to use it.
george
Hello George,
what do you mean with steal opponents time ?
Nemo definitely steals nothing.
By the way, we are already analyzing potential problems with the
memory allocation. Because you are so far the only one reporting
a problem i will contact you later by pm, if that is ok ?
Maybe with your help we are able to catch sth much faster.
Michael
Michael- steal time, use half his opponents' time when it should not- it is all semantics. Nothing personal. I have no idea how Nemo gets half the time his opponent should be using- meaning the opponent only get 50% of his time. May be he loans it to Nemo- who knows. HOW Nemo ends up with it is irrelevant. Until it is fixed, neither I nor Werner at CEGT can use it.
and here result of this test-match with 40/3
Nemo SP64n 1.0 Beta - Frenzee 3.5.19 x64 3.0 - 1.0 (4: +2, =2, -0) 75.00% Perf = +190 Elo
Thx Werner + George, Fix in work.
(but as long as the engines dont use the same core it shouldnt effect anything)
I haven't thought through that, but it would not happen for me anyway. On a dual, I would run 2 tournaments with ponder OFF. On a quad- 3 with ponder off and one for normal use.
I'm running a quick eng-eng vs Komodo on my quad, using nemo64o under fritz, ponder off.
I'm using a single cpu Hyperthr off for fritz, in order to have a clear picture of the problem:
from the task manager, when nemo thinks is correctly using 12% of cpu, but when komodo thinks he uses only 6%, and nemo the remaining 6%.