WCCC 2013: Junior wins

[Mike S.]

With a 32-bit or 64-bit operating system ?

This info is missing, but I think we can savely assume that it is 64-bit. Hiarcs leads anyway.

[Harvey Williamson]

With a 32-bit or 64-bit operating system ?

This info is missing, but I think we can savely assume that it is 64-bit. Hiarcs leads anyway.

64

[Harvey Williamson]

In the WCSC, there are still 3 rounds to go which are scheduled for Sunday. Currently, Hiarcs leads with 4.5/7.

http://icga.uvt.nl/?page_id=633

Hiarcs is the defender of this title, having won it in 2011 against the very same opponents, except Merlin.

Actually we have 6.5/7 in the WCSC junior has 5.5

One piece of information on Jonny on its cluster. At the start of the games it was getting 2 billion nodes per second and reaching depth 40+

[Modern Times]

OK very good, so this year the equal hardware was a quad running a 64-bit O/S. That is fair to all competitors, and is the type of hardware easily affordable to most. Pretty much all PCs bought today have 64-bit Windows installed.

[Daniel Shawul]

I think Jonny's attempt to use all that hardware is more exciting than the tournament results. That is if it is an 'honest' attempt, by that I mean really trying to use it rather than show off with it using a poor algorithm that skyrockets the nps but doesn't actually help play. If I take Harvey's numbers 2billion/2400 ~ 0.85million nps which seems to suggest a ver good NPS scaling hence probably a poor algorithm. It is exciting nonetheless and I hope some details are revealed for some of us that are curious.

[Modern Times]

Yes, I'm fascinated by chess engines playing on huge hardware, no matter what the results.

[Uri Blass]

In the WCSC, there are still 3 rounds to go which are scheduled for Sunday. Currently, Hiarcs leads with 4.5/7.

http://icga.uvt.nl/?page_id=633

Hiarcs is the defender of this title, having won it in 2011 against the very same opponents, except Merlin.

Actually we have 6.5/7 in the WCSC junior has 5.5

One piece of information on Jonny on its cluster. At the start of the games it was getting 2 billion nodes per second and reaching depth 40+

With 2 draws and 5 wins you can have only 6 out of 7
Junior has 5 out of 7.

[Harvey Williamson]

In the WCSC, there are still 3 rounds to go which are scheduled for Sunday. Currently, Hiarcs leads with 4.5/7.

http://icga.uvt.nl/?page_id=633

Hiarcs is the defender of this title, having won it in 2011 against the very same opponents, except Merlin.

Actually we have 6.5/7 in the WCSC junior has 5.5

One piece of information on Jonny on its cluster. At the start of the games it was getting 2 billion nodes per second and reaching depth 40+

With 2 draws and 5 wins you can have only 6 out of 7
Junior has 5 out of 7.

Did we have 2 draws it is all merging into 1 you are probably right!

[Uri Blass]

I think Jonny's attempt to use all that hardware is more exciting than the tournament results. That is if it is an 'honest' attempt, by that I mean really trying to use it rather than show off with it using a poor algorithm that skyrockets the nps but doesn't actually help play. If I take Harvey's numbers 2billion/2400 ~ 0.85million nps which seems to suggest a ver good NPS scaling hence probably a poor algorithm. It is exciting nonetheless and I hope some details are revealed for some of us that are curious.

The results clearly suggest that the hardware helped jonny because with equal hardware it got only 3.5 out of 7 when in WCCC they got 7 out of 10.

If we ignore games against merlin that lost all the game we get
5 out of 8 in WCCC against 2.5 out of 6 in WCSC and it suggest more than 100 elo improvement relative to equal hardware.

[Laskos]

I think Jonny's attempt to use all that hardware is more exciting than the tournament results. That is if it is an 'honest' attempt, by that I mean really trying to use it rather than show off with it using a poor algorithm that skyrockets the nps but doesn't actually help play. If I take Harvey's numbers 2billion/2400 ~ 0.85million nps which seems to suggest a ver good NPS scaling hence probably a poor algorithm. It is exciting nonetheless and I hope some details are revealed for some of us that are curious.

The results clearly suggest that the hardware helped jonny because with equal hardware it got only 3.5 out of 7 when in WCCC they got 7 out of 10.

If we ignore games against merlin that lost all the game we get
5 out of 8 in WCCC against 2.5 out of 6 in WCSC and it suggest more than 100 elo improvement relative to equal hardware.

Still, only 4-6x effective speedup from 16 cores (competition) to 2400 cores.