YetiSim Blog

A response on the TBB forum has revealed that the cache_aligned_allocator will use the scalable_allocator when present. With all the possible choices of memory allocators, I’m a bit puzzled as to which ones to use where. This is something that will have to be examined in greater detail in future. The scalable_allocator seems like the best choice right now, since the cache does not seem to be a major bottleneck at the moment. Each data structure will have unique memory requirements, depending upon its role within the simulation. It is most likely that each component will have to be carefully considered, and the appropriate allocator chosen. But this is something that can happen later… and perhaps facilities will be provided to the user to adjust these defaults themselves at runtime to tune their particular simulation.

I would like very much for the wiki to be completely open, however some pages have been filled with spam.  As a result, I have taken some actions to deal with it.  Basically, edits are now limited to those with accounts and confirmed email addresses.  A human test will also be required, basically one of those annoying pictures that you have to read the text from.  It is unfortunate, but I do not want to waste time cleaning up the spam from the wiki.  Hopefully this will reduce the problem.

I migrated my laptop, and hence my primary development system to Kubuntu over the weekend. I did this, because I could not get VTune to work on Gentoo. Kubuntu has come a long way since I last tried it, and I don’t mind it so far. I have successfully installed VTune, and it looks like a pretty nice tool.

For some reason however, VTune crashes when analyzing my test YetiSim simulation. I’m completely lost as to what is causing this. When I call a function linked from YetiSim, VTune crashes abruptly. I have posted to the VTune forum, and provided instructions on how to download YetiSim since that is the source of the problems.

At first I suspected that the source of the problems was the singleton MasterScheduler, because a source file with a single scheduler() call causes the fault. However, the creation of a single state machine also causes the fault. I suspect that the problem is that YetiSim is a static library, and that perhaps VTune would like a shared library more… although I have no idea what the problem is.

I’m going to ask someone else to run my code on their system, to see if they can replicate the problem.

Well, this is my first blog and I’m not sure what to expect. It’s late, and it’s been a long day of getting YetiSim’s new site ready for deployment. I successfully migrated YetiSim from my personal SVN server to its new home on the eyetap server at U of T. That was a bit of an ordeal, but it’s good now. I got websvn up, and this pretty blog today as well. I also have the wiki setup, which will be the main CMS for the site. I owe a huge thanks to Raymond Lo for all of his help in getting YetiSim up on EyeTap, and to Steve Mann for offering the EyeTap server for hosting. I appreciate all of their help, I know it can be a pain at times to have a new user who wants to ask a lot of questions and get a lot of help :-)

I’m excited to be able to start blogging, both about Threading Building Blocks and YetiSim. Over the past couple of months there have been a lot of things that I wanted to blog about, and well, now I finally can. Did I mention it was exciting? Well, school is very busy still. Not much work will get done on YetiSim until my break during Christmas. However, I will work on documentation for the site and of course blog ideas. My goal is to get a stable release of YetiSim out for the new year.

Well, I’m off to the local pub after a long day of work getting YetiSim up on EyeTap.