JuliaGPU/CUDA.jl

Option for deterministic sparse matrix vector multiplication

Open

#1,607 opened on Sep 30, 2022

View on GitHub
 (2 comments) (0 reactions) (0 assignees)Julia (274 forks)batch import
cuda librariesenhancementgood first issue

Repository metrics

Stars
 (1,408 stars)
PR merge metrics
 (Avg merge 5d 5h) (16 merged PRs in 30d)

Description

I would like to request deterministic sparse matrix vector multiplications. Essentially, the following code should always return the same result:

using CUDA
using Random

Random.seed!(546784)

A = cu(sprand(Float64, 1000, 1000, 0.6))
x = cu(rand(Float64, 1000))

y1 = A * x
y2 = A * x

maximum(abs.(y2 - y1))

A similar issue is #938.

What I found out so far

We need to use CUSPARSE_SPMV_COO_ALG2 or CUSPARSE_SPMV_CSR_ALG2. This will probably be related to changing the argument algo of mv!(): https://github.com/JuliaGPU/CUDA.jl/blob/33a71872b0aebe13f8210229e1265b779f90b78e/lib/cusparse/generic.jl#L112-L113

Unfortunately, the algorithm information seems to get lost in the wrapper function: https://github.com/JuliaGPU/CUDA.jl/blob/99f962abb8296e36f58963c80a4b52cc116e6560/lib/cusparse/interfaces.jl#L6-L9

The wrapper seems to be used for implementing the operators: https://github.com/JuliaGPU/CUDA.jl/blob/99f962abb8296e36f58963c80a4b52cc116e6560/lib/cusparse/interfaces.jl#L38-L60

The remaining question for the maintainers is now: Where do we pass in the algorithm?

Contributor guide

Research direction
Examine the mv! function in generic.jl and the wrapper in interfaces.jl to understand how to pass a deterministic algorithm flag (e.g., CUSPARSE SPMV CSR ALG2) through the multiplication operators. Implement a keyword argument for the algorithm and propagate it to the underlying cusparse routines.
Tech stack
None
Domain
api
Issue type
Feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
JuliaCUDA
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
30