summaryrefslogtreecommitdiffstats
path: root/benchmark
AgeCommit message (Collapse)AuthorFilesLines
2014-01-07Switch raw data field to uint64 to fit all values and update benchmark.Christopher Peplin1-4/+2
2014-01-07Remove timestamp from trace JSON before benchmarking - oops!Christopher Peplin1-2/+4
The size savings is now about 35%, not 55%.
2014-01-07Update benchmark tool for latest protobuf format.Christopher Peplin1-9/+5
2014-01-07Add generated protobuf implementations.Christopher Peplin2-4/+0
2014-01-07Move proto definition to top level.Christopher Peplin2-36/+1
2014-01-07Use a container type to be able to stream protobufs.Christopher Peplin2-7/+26
2014-01-07Benchmark how much overhead each encoding adds vs raw CAN.Christopher Peplin1-4/+16
2014-01-07Benchmark encoding translated messages as binary, too.Christopher Peplin2-11/+48
Results for all of my trace files: > ./compare_sizes.py ~/ownCloud/OpenXC\ Traces/peplin/**/*.json 1 For the 89 trace files given... Total transferred raw JSON size is 13.3MB Total transferred raw binary size is 2.1MB Total transferred translated JSON size is 791.0MB Total transferred translated binary size is 318.6MB Total transferred JSON size is 804.3MB Total transferred binary size is 320.6MB Binary encoding is 84.328494% smaller than JSON for raw messages Binary encoding is 59.727489% smaller than JSON for translated messages Binary encoding is 60.133196% smaller than JSON overall
2014-01-07Add benchmarking tests for binary encoding (moved from cantranslator).Christopher Peplin5-0/+57