Skip to content

issues Search Results · repo:finch-tensor/Finch.jl language:Julia

Filter by

297 results
 (59 ms)

297 results

infinch-tensor/Finch.jl (press backspace or delete to remove)

finch-tensor/Finch.jl
Error when computing negative and finite vector norm

Hi @willow-ahrens, I started porting vector norm and I get a RuleRewriteError for all negative values of p: using Finch using LinearAlgebra A_ref = [1 -2 -3; 2 3 -1] A = Tensor(Dense(SparseList(Element(0.0))), ...
bug
  • mtsokol
  • 1
  • Opened 
    7 days ago
  • #708

finch-tensor/Finch.jl
Galley MTTKRP performance in `v1.2.0`

Hi @kylebd99, I think this is the script that reproduces the Galley MTTKRP performance issue I mentioned for the 1.2.0 release: using Finch config = ( (100, 25, 100, 10, 0.001), (100, 25, 100, ...
  • mtsokol
  • Opened 
    24 days ago
  • #703

finch-tensor/Finch.jl
Implement squeeze (dropdims)

In python, the function is https://numpy.org/doc/2.2/reference/generated/numpy.squeeze.html, and in julia the function is https://docs.julialang.org/en/v1/base/arrays/#Base.dropdims. A related function ...
  • willow-ahrens
  • Opened 
    on Feb 12
  • #702

finch-tensor/Finch.jl
Implement argmin and argmax

The idea is to first use minby to compute the argmin, then return just the coordinates of the minimum. In python, we only need to handle a single axis (https://numpy.org/doc/2.1/reference/generated/numpy.argmin.html). ...
  • willow-ahrens
  • Opened 
    on Feb 12
  • #701

finch-tensor/Finch.jl
Investigate potential memory leak associated with the use of deferred nodes.

Deferred nodes now hold copies of the actual inputs, for use in the galley scheduler. If we hold on to the program with deferred nodes, we may introduce a memory leak as the inputs cannot be garbage collected. ...
  • willow-ahrens
  • 2
  • Opened 
    on Jan 7
  • #692

finch-tensor/Finch.jl
Investigate Galley call overhead for einsum interface.

As noticed in https://github.com/finch-tensor/Finch.jl/pull/679, Galley s adaptive scheduler appears to have a very large calling overhead when used with the einsum interface. Why is this the case? Can ...
  • willow-ahrens
  • Opened 
    on Jan 7
  • #691

finch-tensor/Finch.jl
- densifies zeros

A - A results in all the zero entries getting converted to -0.0 and being stored explicitly 😬 julia A = fsprand(5, 5, 0.5) 5×5 Tensor{SparseCOOLevel{2, Tuple{Int64, Int64}, Vector{Int64}, Tuple{Vector{Int64}, ...
  • willow-ahrens
  • 3
  • Opened 
    on Jan 3
  • #686

finch-tensor/Finch.jl
Add MWE for migration from SparseArrays

https://discourse.julialang.org/t/ann-finch-jl-sparse-and-structured-array-fusion/123178/16?u=willow
  • willow-ahrens
  • Opened 
    on Dec 30, 2024
  • #683

finch-tensor/Finch.jl
Add Gustavson's algorithm (workspace optimization) to the high-level scheduler.

https://discourse.julialang.org/t/ann-finch-jl-sparse-and-structured-array-fusion/123178/16?u=willow
  • willow-ahrens
  • Opened 
    on Dec 30, 2024
  • #682

finch-tensor/Finch.jl
Make Galley Adaptive To Inputs

The goal is to have galley keep a cache of compiled plans. These plans will be associated with a statistics object. To do this, we ll need to: 1. Make a separate GalleyExecutor. 2. Add a get_cannonical_stats(stats, ...
enhancement
  • kylebd99
  • Opened 
    on Dec 19, 2024
  • #678
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Press the
/
key to activate the search input again and adjust your query.
Issue origami icon

Learn how you can use GitHub Issues to plan and track your work.

Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub Issues
ProTip! 
Restrict your search to the title by using the in:title qualifier.
Issue search results · GitHub