okayzed

we can either store the mantissa separately from the integer portion (and the column size might get doubled or more this way) or we can store floats as a new...

Investigate fast & dynamic serialization libraries

2

https://github.com/alecthomas/go_serialization_benchmarks 50% of program time is spent in deserialization. we need faster

faster hashmaps?

2

https://github.com/cornelk/hashmap

[grpc] make more sybil commands friendly to repeated execution

2

currently, the sybil binary is designed to run once and then exit - we need to audit and fix the individual areas so we can ingest, digest and query multiple...

Make distributed queries a reality

1

# Distributed sybil This is an umbrella task for working towards distributed queries with sybil as an aggregator. ## Steps to get there ### Deployment * setting up docker /...

see if perf and accuracy is reasonable, too

add benchmarks on a commit by commit basis (tmc's idea) that will measure: * how long to save/load one block of ints, strs and sets * how long to query...

The DB lock files will no longer work as expected when using multiple queries or ingestions in a single sybil process, because they use the PID to coordinate and figure...

there are some places that i think can use improvement, this task is to list and improve them accuracy: * combining histograms between leaf nodes might lose accuracy, because the...

We should check in serialized results to the repo for past sybil versions and make sure that future sybil binaries are compatible with them