Jakob Truelsen
Jakob Truelsen
I guess we could trie to write a custom comparator to see if we can get the internal part up to speed, as for the number of phases, that is...
Running tpie/build/apps/siv_dump/siv_dump /tmp/dataset.siv faults while tpie/build/apps/siv_dump/siv_dump --recursive /tmp/dataset.siv succedes [example.zip](https://github.com/thomasmoelhave/tpie/files/1889056/example.zip)
I don't think this fix works. if i do for (auto it=tree.begin(); it != tree.end(); ++it) {...} It appears to fail as a new end node is constructed and deleted...
We should not use unordred maps to cache nodes. We should fix the comparision..
If you use the ppp branch, and include terrastream/pipe/chunk Then create a pipe like this generate_points() | chunk(1024*1024) | parallel(process(), false, 4, 1 ) | unchunk() | consume_points() Here generate_points,...
We alwayes use map() in terrastream, so I propose to deprecate or remove lambda.
I would definitly make sence to compress the blocks, instead of compressing the individual text strings. If @mortal has time perhaps he can tell us what the best approach will...
I think the main issue putting d>250 is that we run out of fds on linux if we do several parallel merges. However since this is only an issue for...
Do we want to remove the default case?
I do not plan on implementing this. Pull requests are however welcome.