It's available here
It is based on vibe-0.7.29.
Some notes:
It would be nice to tackle these before final round 13 will come out.
I would look into ldc setup, but I'm on vacation next week, so volunteers welcome :)
Am 01.09.2016 um 13:31 schrieb tchaloupka:
It's available here
It is based on vibe-0.7.29.
Some notes:
- vibe-ldc setup is failing for some reason
- fortunes test is also failing (some validation test)
- I would expect the overal results to be somewhat better
It would be nice to tackle these before final round 13 will come out.
I would look into ldc setup, but I'm on vacation next week, so volunteers welcome :)
One obvious improvement would be to implement MySQL or Postgres support,
because MongoDB is somewhat slow and this affects most tests. I'll try
to have a look at the fortunes benchmark, but the plain text one also
looks really odd.
On Fri, 2 Sep 2016 14:45:23 +0200, Sönke Ludwig wrote:
Am 01.09.2016 um 13:31 schrieb tchaloupka:
It's available here
It is based on vibe-0.7.29.
Some notes:
- vibe-ldc setup is failing for some reason
- fortunes test is also failing (some validation test)
- I would expect the overal results to be somewhat better
It would be nice to tackle these before final round 13 will come out.
I would look into ldc setup, but I'm on vacation next week, so volunteers welcome :)One obvious improvement would be to implement MySQL or Postgres support,
because MongoDB is somewhat slow and this affects most tests. I'll try
to have a look at the fortunes benchmark, but the plain text one also
looks really odd.
SQL databases came with single reason, to avoid lag of fetching and searching around stacked data.
Am 28.09.2016 um 04:55 schrieb EMRE TEMELKURAN:
On Fri, 2 Sep 2016 14:45:23 +0200, Sönke Ludwig wrote:
Am 01.09.2016 um 13:31 schrieb tchaloupka:
It's available here
It is based on vibe-0.7.29.
Some notes:
- vibe-ldc setup is failing for some reason
- fortunes test is also failing (some validation test)
- I would expect the overal results to be somewhat better
It would be nice to tackle these before final round 13 will come out.
I would look into ldc setup, but I'm on vacation next week, so volunteers welcome :)One obvious improvement would be to implement MySQL or Postgres support,
because MongoDB is somewhat slow and this affects most tests. I'll try
to have a look at the fortunes benchmark, but the plain text one also
looks really odd.SQL databases came with single reason, to avoid lag of fetching and searching around stacked data.
I'd say that the primary reason was to provide a powerful standard query
language, with the specialty of being a relational. In theory, MongoDB
(without being relational) could also store pre-compiled queries and
could have equal performance if the proper indexes are there, but the
implementation simply doesn't seem to be very fast.
Just as a background, when I first started to implement those
benchmarks, I actually started to use mysql-native, but got distracted
before it was finished, because progress was hindered by the lack of
documentation. Then someone was faster and submitted the MongoDB based
benchmarks (which is of course much better than nothing at all!), which
is where we are now.
On Wed, 28 Sep 2016 12:53:12 +0200, Sönke Ludwig wrote:
Am 28.09.2016 um 04:55 schrieb EMRE TEMELKURAN:
On Fri, 2 Sep 2016 14:45:23 +0200, Sönke Ludwig wrote:
Am 01.09.2016 um 13:31 schrieb tchaloupka:
It's available here
It is based on vibe-0.7.29.
Some notes:
- vibe-ldc setup is failing for some reason
- fortunes test is also failing (some validation test)
- I would expect the overal results to be somewhat better
It would be nice to tackle these before final round 13 will come out.
I would look into ldc setup, but I'm on vacation next week, so volunteers welcome :)One obvious improvement would be to implement MySQL or Postgres support,
because MongoDB is somewhat slow and this affects most tests. I'll try
to have a look at the fortunes benchmark, but the plain text one also
looks really odd.SQL databases came with single reason, to avoid lag of fetching and searching around stacked data.
I'd say that the primary reason was to provide a powerful standard query
language, with the specialty of being a relational. In theory, MongoDB
(without being relational) could also store pre-compiled queries and
could have equal performance if the proper indexes are there, but the
implementation simply doesn't seem to be very fast.Just as a background, when I first started to implement those
benchmarks, I actually started to use mysql-native, but got distracted
before it was finished, because progress was hindered by the lack of
documentation. Then someone was faster and submitted the MongoDB based
benchmarks (which is of course much better than nothing at all!), which
is where we are now.
Whoops. None of the vibe.d benchmarks are completing at the moment!
