Code center > Suggestions

Regarding the enviroment grid!

<< < (2/2)

EricL:
Completely agree that threading will help in cases where the program does I/O either to disk or to the network or as we add architectual artifacts which create global critical sections.  The affordability of MP machines is also improving, so it won't be long before threading helps main-line perf as well for those with such machines.  Didn't mean to sound negative on threads, just wanted to make the point that they arn't magic or free.  Multi-threaded bugs are a bitch to find sometimes...

As far as the grid, I don't know enough about it to comment intelligently, but I'm all for a more complex environment.

Greven:

--- Quote ---Multi-threaded bugs are a bitch to find sometimes...
--- End quote ---


Oh you bet. I have just made an exam project about diskscheduling in C on linux/unix machines, with thread, mutex' you name it, I have used it.
My program compiled perfectly etc., but when run I got an error ---> killed (in 1 out of 10)! The program just returned the text "killed" instead of some error or something!!!! I were able to track it down to just one line of code, a free()! When I removed this single line the program did run fine, excepted that the program then had a memory leak then... It took me well of 6 hours to track this shit down.... But I settled for some mixed approach, indicating it in my paper... Well back to topic... right

Numsgil:

--- Quote from: EricL ---Multi-threaded bugs are a bitch to find sometimes...
--- End quote ---

 Amen, hehe.


When I finish the C++ version to a stable, releasable version my 5 top priorities are:

1.  Speed
2.  Environment Grid
3.  Metabolism
4.  Specialization and Speciation
5.  Sexual Reproduction

in that order.

Greven:
OK sounds good to me....

Navigation

[0] Message Index

[*] Previous page