You can see the new version that I give here that is what I changed from the initial version.

If you do not like it you can start from the version that passed stage 2 that is functionally equivalent.
uriblass/Stockfish@7ed15af...460d93d

That's a lot of work, thanks Uri! Congrats on another successful patch as well. Maybe the best plan is to apply your patch that passed the framework first, then do the cleanup patch as a followup, since it's a fairly big one, and we want to make sure it's right.

Thoughts?

Let's wait and see the cleanup. I think the cleanup is not so simple...

I did part of the clean up and you can also see the patch without the
cleanup so if you want to apply the patch without the cleanup first you can
do it.

I stopped the cleanup also because I was not sure if it is right what I did
when I replaced Depth type by int in search.cpp and thought maybe you
prefer it to be still a special type even without ONE_PLY

On Sat, Sep 27, 2014 at 6:05 PM, lucasart notifications@github.com wrote:

Let's wait and see the cleanup. I think the cleanup is not so simple...

—
Reply to this email directly or view it on GitHub
#55 (comment)
.

It's fine to commit this I think.

ELO: 3.90 +-3.1 at STC strongly indicates that this even a slight improvement itself.

@gary: If you agree, let's commit this straight away. Then we can start thinking about clean up. Marco already promised to take care of it.

And I'm talking about Uri's two original patches. Not the third patch!

I hope the third patch is going to help marco even if it is not a starting
point because he can see some new formulas that I use to make the code
equivalent and maybe he can use part of them without needing to think about
the right mathematical formula

for example
int d=((depthx-2)*2-(PvNode ? 0 : depthx/2))/2;
instead of a formula to calculate d and later to reduce 1 in case that you
get odd number.

On Sat, Sep 27, 2014 at 10:51 PM, Joona Kiiski notifications@github.com
wrote:

And I'm talking about Uri's two original patches. Not the third patch!

—
Reply to this email directly or view it on GitHub
#55 (comment)
.

@zamar, agreed, I made some cosmetic tweaks and pushed the original two patches.

@mcostalba, that's a big project!! Thanks for taking a look at it :)

@uriblass, congratulations on another nice patch!

Note that it is not only the result at short time control but also the
result at long time control
that I think is enough to pass also SPRT(0,6)

LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 7647 W: 1356 L: 1214 D: 5077

For comparison some earlier test with a similar number of games passed with
a similiar result 137 wins more than losses instead of 142 wins more than
losses in my patch

http://tests.stockfishchess.org/tests/view/526e08900ebc5936609edf7f

LLR: 2.95 (-2.94,2.94) [0.00,6.00]
Total: 7312 W: 1276 L: 1139 D: 4897

On Sat, Sep 27, 2014 at 10:50 PM, Joona Kiiski notifications@github.com
wrote:

It's fine to commit this I think.

ELO: 3.90 +-3.1 at STC strongly indicates that this even a slight
improvement itself.

@gary https://github.com/Gary: If you agree, let's commit this straight
away. Then we can start thinking about clean up. Marco already promised to
take care of it.

—
Reply to this email directly or view it on GitHub
#55 (comment)
.