Commits on Sep 15, 2024

1. Add benchmarks for "insert many" operations …

   A couple quick benchmarks for the two "insert many" variants to help vet
   whether or not we should keep the "fast" route.
   
   There is definitely a speed difference, with the fast variant being
   about 25% faster as it can skip returning rows:
   
       go test -bench=BenchmarkClient
       goos: darwin
       goarch: arm64
       pkg: github.com/riverqueue/river
       cpu: Apple M1
       BenchmarkClient/JobInsertMany-8                      312           4156647 ns/op
       BenchmarkClient/JobInsertManyFast-8                  417           3054985 ns/op
       PASS
       ok      github.com/riverqueue/river     15.709s
   
   Can't 100% decide where that leaves us. There's definitely a speed
   advantage, but it definitely congests the API to a degree, and having
   more functions make implementing features like middleware trickier.

   

   brandur committed Sep 15, 2024
   Configuration menu

   • View commit details
   • Copy full SHA for 999ff03
   • Browse repository at this point

   Copy the full SHA

   999ff03 View commit details

   Browse the repository at this point in the history