So, it looks like a stupid mistake I made way back invalidates most of the graphs I've made for the thesis. Sorting it all out kind of swallowed yesterday and the day before. One bright side is that, since it doesn't affect timing at all, those graphs can stay. And because I already have that data, I was able to switch the code back to the old multi-process model instead of multi-threaded. Python does not multi-thread very well at all, but it allowed me to get more precise timing results. So rerunning everything is going considerably faster, now that I'm back to really slamming the 8-core research servers.
I did manage to make it to a Internet Anniversary party last night. Happy birthday, you insane exaggeration of all the best and the worst of humanity! I made this coffee cake for it, which is actually pretty easy and quite yum. But I used the streussel from this crumble instead of the one specified. There is very little I find as tasty as oatmeal streussel, and the cinnamon one was a bit overpowering last time I made it. I think it could use up to 50% more, but overall it worked very well. I'll certainly make it again -- finding good coffee cake in Vancouver is surprisingly hard.
I did manage to make it to a Internet Anniversary party last night. Happy birthday, you insane exaggeration of all the best and the worst of humanity! I made this coffee cake for it, which is actually pretty easy and quite yum. But I used the streussel from this crumble instead of the one specified. There is very little I find as tasty as oatmeal streussel, and the cinnamon one was a bit overpowering last time I made it. I think it could use up to 50% more, but overall it worked very well. I'll certainly make it again -- finding good coffee cake in Vancouver is surprisingly hard.
no subject
(no subject)
(no subject)
(no subject)
(no subject)
no subject
(no subject)
(no subject)