Skip to content
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

Improvements to sort functions #1373

Merged
merged 28 commits into from
Apr 24, 2016
Merged

Improvements to sort functions #1373

merged 28 commits into from
Apr 24, 2016

Conversation

shehzan10
Copy link
Member

@shehzan10 shehzan10 commented Mar 31, 2016

This pull request adds:

  • PERF: Improve the performance of batched sort along 0th dimension
  • FEAT: Add sorting along higher dimensions
  • Enable larger tests for sorting functions
  • FEAT: Add multi dimension support for media
  • Add multi dimensions tests for median

Fixes #291 #394

This allows complex sorting based on a key, which can be the abs, real, imag
etc of the value array. So the user can choose which metric they want to use.
* Sort now allows all dimensions
* Sort if much faster by using batched mode. This takes up more memory though.
* Enabled large tests for sort
* Added tests for sorting on dim1 and dim2
@shehzan10 shehzan10 added this to the 3.3.2 milestone Mar 31, 2016
@shehzan10 shehzan10 changed the title WIP: Improvements to sort fucntions WIP: Improvements to sort functions Apr 1, 2016
@shehzan10
Copy link
Member Author

build arrayfire windows ci

@shehzan10
Copy link
Member Author

@pavanky updated sort_index code for CPU and CUDA

@shehzan10
Copy link
Member Author

build arrayfire tegrak1 ci
build arrayfire tegrax1 ci
build arrayfire linux-03 ci

@shehzan10
Copy link
Member Author

build arrayfire tegrak1 ci

@shehzan10 shehzan10 changed the title WIP: Improvements to sort functions Improvements to sort functions Apr 21, 2016
@shehzan10
Copy link
Member Author

build arrayfire linux-03 ci

cuda::kernel::iota<uint>(pSeq, seqDims, tileDims);

// Make pkey, pVal into a pair
thrust::device_vector<IndexPair<Tk, Tv> > X(inDims.elements());
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Allocate using arrayfire. This is going to be costly otherwise.

@shehzan10
Copy link
Member Author

@pavanky @9prady9 Can you guys review f6eae07

@9prady9
Copy link
Member

9prady9 commented Apr 23, 2016

@shehzan10 Tests look fine.

@pavanky
Copy link
Member

pavanky commented Apr 23, 2016

looks good 👍

@pavanky
Copy link
Member

pavanky commented Apr 23, 2016

build arrayfire ci

1 similar comment
@shehzan10
Copy link
Member Author

build arrayfire ci

@pavanky pavanky merged commit 9c4cad3 into arrayfire:hotfix-3.3.2 Apr 24, 2016
@shehzan10 shehzan10 deleted the sort branch April 25, 2016 21:52
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants