Hi Everyone,
So for the background, I have an API for retrieving data from a database.
For an endpoint there exists two function calls:
- SelectCollections - for retrieving the collection records
- HydrateCollection - for hydrating each collections (the second function
is called on each results of the SelectCollections function).
Previously we run the HydrateCollections sequentially using a for loop, and
it results in the following:
2018/05/22 10:59:05 HydrateCollection took 41.497134ms
2018/05/22 10:59:05 HydrateCollection took 42.695152ms
2018/05/22 10:59:05 HydrateCollection took 22.870913ms
2018/05/22 10:59:05 HydrateCollection took 17.833512ms
2018/05/22 10:59:05 HydrateCollection took 18.765984ms
2018/05/22 10:59:05 HydrateCollection took 16.338906ms
2018/05/22 10:59:05 HydrateCollection took 14.660654ms
2018/05/22 10:59:05 HydrateCollection took 20.818716ms
2018/05/22 10:59:05 HydrateCollection took 18.57054ms
2018/05/22 10:59:05 Resolving Items took 172.9182ms
2018/05/22 10:59:05 SelectCollections took 173.992051ms
After using goroutines to run HydrateCollections now the results is as
follows:
2018/05/22 11:21:36 HydrateCollection took 9.187861ms
2018/05/22 11:21:36 HydrateCollection took 63.651507ms
2018/05/22 11:21:36 HydrateCollection took 79.199976ms
2018/05/22 11:21:36 HydrateCollection took 122.349986ms
2018/05/22 11:21:37 HydrateCollection took 150.627746ms
2018/05/22 11:21:37 HydrateCollection took 168.432517ms
2018/05/22 11:21:37 HydrateCollection took 171.602705ms
2018/05/22 11:21:37 HydrateCollection took 179.127794ms
2018/05/22 11:21:37 HydrateCollection took 185.137562ms
2018/05/22 11:21:37 Resolving items took 185.434821ms
2018/05/22 11:21:37 SelectCollections took 187.358141ms
Why does the HydrateCollection take a significant performance hit now that
I'm using goroutine?
HydrateCollection function takes a pointer to Collection struct, and
mutating it inside the goroutine, so I think that it may be a problem. Also
the HydrateCollection itself is a method of a Repository pointer which
holds a database connection, the same Repository pointer is used by
SelectCollections.
hydrateChan := make(chan error, len(collections))
for index := range collections {
go func(coll *Collection) {
hydrateChan <- r.hydrateCollection(coll)
}(&collections[index])
}
for i := 0; i < len(collections); i++ {
err := <-hydrateChan
if err != nil {
logrus.Error(err)
}
}
Is there something that I need to address in the previous code snippet?
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
