Houdini 1.03

Houdini 1.03 Multi-PV, searchmove and large page support. Improved evaluation function.


Currently the following UCI options are available:

* Hash: Amount of hash table memory used by the program in MB. Default 128, min 4, max 1024/16384 (32/64-bit).
* Threads: Maximum number of cores used by the analysis. Default is version-dependent, min 1, max 8.
* Split_Depth: Minimum depth at which work will be split between cores. Default 10, min 8, max 99.
* Ponder: [game play only] Have the engine think during its opponent's time. True (1) or false (0).
* Analysis_Contempt: [position analysis only] Make Houdini try to avoid draws. True (1) or false (0).
* MultiPV: [position analysis only] Number of principal variations shown. Default 1, min 1, max 16.
* Large_Pages: Activate large memory pages, see description below. True (1) or false (0).

32 bit
Can be used on any not too ancient Windows version. From release 1.02 on these compiles no longer depend on SSE2 instructions and will also run on older processors that don't support the SSE2 instruction set.

# Single core: Houdini_w32_1CPU http://www.multiupload.com/EY8T38WVGC
# Dual core: Houdini_w32_2CPU http://www.multiupload.com/NTV9A1OZ0H
# Quad core: Houdini_w32_4CPU http://www.multiupload.com/GQB4PCSVRT
# 6 or 8 core: Houdini_w32_8CPU http://www.multiupload.com/NLGQ1XO1V2


Recommended for 64-bit Windows, the speed increase is about 30% compared to the 32-bit version.

* Single core: Houdini_x64_1CPU http://www.multiupload.com/JWQVX49MB7
* Dual core: Houdini_x64_2CPU http://www.multiupload.com/HA76FPZ59Y
* Quad core: Houdini_x64_4CPU http://www.multiupload.com/62BLZ14RB2
* 6 or 8 core: Houdini_x64_8CPU http://www.multiupload.com/B6UO8FYTXB

On 64-bit hardware that supports SSE4.2 (e.g. Intel Core i5 and i7) you can use the "POPCNT" version that runs slightly faster than the standard version.

* Quad core: Houdini_x64_POPCNT_4CPU http://www.multiupload.com/OYVJRISDXS

* 6 or 8 core: Houdini_x64_POPCNT_8CPU http://www.multiupload.com/U81K5TTVF0

clair wrote:

thanks ill try this engine ,, im looking for ivanhoe b63 mob12b i hear its very strong [i]

i think that was one the girls tested and peter pants vetoed there suggestions to post it rather than mod12c
i could be wrong and have been before , only Angel will know for sure but shes still recovering from her operation, but even if correct i doubt it would be as fast as some of this months releases


Houdini 1.03 is as a slug to ivanhoe , all the pr about ever being faster is pure @@s. its evaluations are its strong point but without bases its just a contender

I am haveing problems with Houdini 1.03 x64 popcnt 4 cpu locking up and the clock looping back and forth at end of games on Fritz , does anyone else get this problem ?

many thanks for the update, i take it
have a nice day.

Hello
This crashes as often as Firebird :laf gal: .is no good to use at all , a mega and makes me so sad i

brenster wrote:

I agree, even though hash options and Multi PV has made improvements, this engine still has some bugs and crashed several times during play already....

Sadly we are recieving messages that it is equally as bad for intell pc users as it is for AMD users
we can not understand why simple tests prior to releases are not done , a few hours (at most ,minutes possibly ) would show such obvious problems

we get called "haters " for informing members of such issues and not "worshiping " and feeding egos
we are very gratefull and apprecative of members comments




~~~~~~~~~~~~~~~~~~~~~~~~~~~

it seems it is link with 64 and popcount because with x32 and 2 cpu, no problem under arena.

Intel Xeon X5355 Octa (8x2660 MHz).
Time 3'+0". Hash 128. Ponder off. GUI Fritz 12. Tablebases 5 pieces.
book: HS-7moves.ctg.

Deep Rybka 4 - Houdini 1.03 x64 8_CPU : +16 -10 =24 .

OK,Rybka is clear stronger but finally after long time a match with less than 50% boring draws!.

Sandi


~~~~~~~~

Seems for 8 cores it works ok but lose to DR4

ok if your a well of captalist i guess :pale:

Robert Houdart posts

This is a bug which several users have reported, and thanks to Herbert L on CCC forum I've been able to reproduce it on the development platform.
The correction should be fairly straightforward (a single line of code to add); I'll provide a fixed release tomorrow Saturday 17th July.

I'll keep you informed, sorry to all users for the inconvenience.

Robert