DPAD - new D client, any experience with it

[JonB]

Reading about the latest DPAD client version, it sounds like the "C" client will generate more MPts but has problems with its genetic algorithms, meaning its results aren't as good.

The "D" client supposedly fixes the science part.

I've started using the D client on two machines but not for long enough to compare.

Anyone examined the differences?

Jon

 

[waffleironhead]

some diiscussion about version here

Man you got me excited! But .d was new as of alomst a year ago.


the newer version uses an updated convergance algorithim. So the client is supposed to be smarter in the choices it makes so the efficiency gets better faster.

Also a point of note is that ANY results sent via .c version are not put into the topp 100 list and so are not IMO contributing to the project getting better. Stephen set up the sample files to weed out any .c results.

The .d version is not able to recreate the same results as the .c version exactly(they are alomst always lower). So if the .c version has the best result, the .d has a real hard time topping it as its results on the same trial are lower. Someone at the dpad forums has shown tho that some .d versions are better than .c versions(ie higher)

anyway....I would reccommend updating all clients to .d

 

[Assimilator1]

You got me going too (and I was thinking 'damn why did they have to release a new cleint in the middle of a SETI race) ,until I realised I'd long since switched to D

Afraid I haven't compared it to C.

 

[JonB]

Then I'm curious why one of my work clients went absolutely BERSERK telling me to upgrade to the new version.

I suppose I should've checked the dates.

 

[Assimilator1]

Strange ,I don't know ,I'd still be interested to know the differences between the clients.