Any news on the A.I.?

[Greyshaft]

So we have:
* a test team (Patrice + 20 others)
* a test planning team (Nebert and I)
* a programming team (Steve, Dean, dhatchen, herwin plus others)
* a unit description team (capitan and his merry men)


Steve, you're going to have to start designing medals or campaign patches so when we all descend on your place for the post-launch party we can see with which campaigns we were each involved ...

"Ah, Borger! I see you wear the 'Order of Mapping' medal and the 'Tutorial 101' patch. You must have been in the Battle for Newbie Comprehension back in November '07 when we were fighting against those d@mned Acronyms armed with TLAs. They were the days. Lost a few good Testers in that battle."

[Shannon V. OKeets]

ORIGINAL: Greyshaft

So we have:
* a test team (Patrice + 20 others)
* a test planning team (Nebert and I)
* a programming team (Steve, Dean, dhatchen, herwin plus others)
* a unit description team (capitan and his merry men)


Steve, you're going to have to start designing medals or campaign patches so when we all descend on your place for the post-launch party we can see with which campaigns we were each involved ...

"Ah, Borger! I see you wear the 'Order of Mapping' medal and the 'Tutorial 101' patch. You must have been in the Battle for Newbie Comprehension back in November '07 when we were fighting against those d@mned Acronyms armed with TLAs. They were the days. Lost a few good Testers in that battle."

Actually, when I was managing an AI group back in the 1980's, I did fashion a set of ribbons for each of the AI projects and gave them out at the end of the year to the team members. Some people had 4 or 5 ribbons around their necks at the end of the celebration/party. Napoleon knew what he was doing.[:)]

[Anendrue]

ORIGINAL: Shannon V. OKeets

Dean,

Thanks. I will consider it. There are many problems though, one of which is I expect the AIO routines to be threaded so they run in the background while the human player is making decisions. That is so the response time from the AIO will be faster because it will have performed as many preparatory analytical routines as possible before it is the AIO's turn to 'move'.

Threaded, hopefully my quad core will be put to the test!

(bold in the quote added by me.)

[Shannon V. OKeets]

ORIGINAL: abj9562

ORIGINAL: Shannon V. OKeets

Dean,

Thanks. I will consider it. There are many problems though, one of which is I expect the AIO routines to be threaded so they run in the background while the human player is making decisions. That is so the response time from the AIO will be faster because it will have performed as many preparatory analytical routines as possible before it is the AIO's turn to 'move'.

Threaded, hopefully my quad core will be put to the test!

(bold in the quote added by me.)

Only if the Delphi compiler knows to make use of it. It is possible to have threaded applications on a single processor.

[Anendrue]

True you can thread on a single core and my wife has my old pentium 4 with HT capability. However most "new" compilers still will not thread across 4 cores, most will do 2 cores and almost all can thread on a single core. I am just wishing for the future before it gets here.

[SamuraiProgrmmr]

I think that it will already take advantage of multi core processors.   In fact, I a pretty sure that is the case.  Windows does not care about where a thread came from, only that it is a thread.
 
Where things get hairy for using multicore processors is when you try to do parallel processing.  That is to say when you have a loop that executes many times, each iteration of the loop is farmed out to a different processor and each step is independent of the other.
 
Already when you do threading, (if done properly) you have made the decisions of when one thread needs to interrupt the main thread to ask for access to non thread-safe resources and/or data.
 
At the very least, all of the other processes running on your machine (i.e. Windows, TCP/IP, Disk Access Routines, etc.) will run on the processor that is slowest.

[mavraamides]

ORIGINAL: SamuraiProgrammer

I think that it will already take advantage of multi core processors. In fact, I a pretty sure that is the case. Windows does not care about where a thread came from, only that it is a thread.

That's my understanding as well.

Windows will always choose the 'least busy' processor for the next thread. For example, I have a dual core here at work and when I'm testing a process bound program that I'm working on, it's main process thread may peg out one of the cores for multiple minutes while it plows through a few hundred billion fp ops. Meanwhile, every other thread that comes up, even from within the same process, gets run in the other core.

There is no special multi-core optimization that I need to set for the compiler.

[Shannon V. OKeets]

ORIGINAL: GordianKnot

ORIGINAL: SamuraiProgrammer

I think that it will already take advantage of multi core processors. In fact, I a pretty sure that is the case. Windows does not care about where a thread came from, only that it is a thread.

That's my understanding as well.

Windows will always choose the 'least busy' processor for the next thread. For example, I have a dual core here at work and when I'm testing a process bound program that I'm working on, it's main process thread may peg out one of the cores for multiple minutes while it plows through a few hundred billion fp ops. Meanwhile, every other thread that comes up, even from within the same process, gets run in the other core.

There is no special multi-core optimization that I need to set for the compiler.

Nice. This is good news. Thanks.

[SamuraiProgrmmr]

Of course, if you wanted to be Real Spiffy (TM), you could have the AIs running as separate applications and use DDE to exchange data between them. 
 
(RUN DUCK AND COVER)
 
(In all seriousness, it would prevent some of the problems associated with thread crashing and since you are already building a message handler ....)
 
Keep up the good work!!!
 
Have a great day!
 
 

[Shannon V. OKeets]

ORIGINAL: SamuraiProgrammer

Of course, if you wanted to be Real Spiffy (TM), you could have the AIs running as separate applications and use DDE to exchange data between them. 

(RUN DUCK AND COVER)

(In all seriousness, it would prevent some of the problems associated with thread crashing and since you are already building a message handler ....)

Keep up the good work!!!

Have a great day!

[Anendrue]

Thank you for updating my knowledge of multi core processing!