-
Notifications
You must be signed in to change notification settings - Fork 66
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update to 0.7 #61
Update to 0.7 #61
Conversation
test/scratch.jl
Outdated
@@ -0,0 +1,189 @@ | |||
NearestNeighbors.knn_point!(tree, [.1,.2,.3], false, Vector{Float64}(5), Vector{Int}(5), NearestNeighbors.always_false) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Accidentally committed this file?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oops, yes! Removing...
Codecov Report
@@ Coverage Diff @@
## master #61 +/- ##
==========================================
+ Coverage 93.89% 99.35% +5.45%
==========================================
Files 14 13 -1
Lines 508 311 -197
==========================================
- Hits 477 309 -168
+ Misses 31 2 -29
Continue to review full report at Codecov.
|
A question: should I revert the |
Nah, just keep it in the same :) |
Currently blocked on StaticArrays 0.7 support. |
bfbe593
to
a0a77af
Compare
reinterpret
, make copies of data instead
Updated this. I couldn't get the |
a0a77af
to
900dc50
Compare
No worries, fine with me!
… Am 27.06.2018 um 14:22 schrieb Kristoffer Carlsson ***@***.***>:
Updated this.
I couldn't get the reorder thing to work with the data free trees so, for now, I removed it. @rened Is this feature important to you?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
|
Fix for #59, alternative approach to #60
Since
ReinterpretArray
seems to have an unacceptable indexing overhead here (see #60 and https://discourse.julialang.org/t/big-overhead-with-the-new-lazy-reshape-reinterpret/7635/1), @KristofferC advices tocopy
instead ofreshape(reinterpret(data...),...)
.This PR implements two parts required for this approach:
1- A
@noinline reinterpret_or_copy
function that usesreinterpret
for julia versions prior to the introduction ofReinterpretArray
lazy wrappers, and makes a reinterpreted copy otherwise.2- Change from the deprecated uninitialized contructor
Array{T}(::Int...)
to the new formArray{T}(uninitialized, ::Int...)
. Not doing this introduced a very large performance regression under julia master. To remain backward compatible I used@compat
, which adds a dependency to Compat.jl.With this, the timings for a few tests I tried were identical to the ones in the master branch under v0.6.