cd /news/machine-learning/sobek-streaming-equivariant-tensor-p… · home topics machine-learning article
[ARTICLE · art-66798] src=arxiv.org ↗ pub= topic=machine-learning verified=true sentiment=↑ positive

Sobek: Streaming Equivariant Tensor Product Convolutions

Researchers have developed Sobek, a generated-CUDA backend for streaming equivariant tensor product convolutions that eliminates edge-sized intermediates, achieving speedups of 1.2× to 49.7× and reducing peak memory by up to 99% across all 75 capacity-matched comparisons. The method, described in a paper submitted on July 20, 2026, enables workloads up to two orders of magnitude larger than OpenEquivariance while maintaining near-peak throughput.

read2 min views2 publishedJul 21, 2026
Sobek: Streaming Equivariant Tensor Product Convolutions
Image: source
[Submitted on 20 Jul 2026]


[View PDF](/pdf/2607.18074)

[HTML (experimental)](https://arxiv.org/html/2607.18074v1)

Abstract:Equivariant graph neural networks repeatedly apply edge-conditioned tensor-product convolutions over graph edges. Conventional implementations materialize edge-specific weights, messages, and adjoints, causing tensor-product workspace and memory traffic to grow rapidly with graph size and operator width. This limits feasible workloads and can prevent larger problems from fully utilizing the GPU.

We show that these edge-sized intermediates are artifacts of the execution schedule, not requirements of the equivariant operator. By reassociating radial projection, spherical-harmonic coupling, and graph aggregation, edge-local products can be consumed directly into bounded receiver-side state. The resulting streaming formulation preserves fully connected multiplicity mixing and extends through forward, backward, and double backward.

We implement this formulation in Sobek, a generated-CUDA backend, and evaluate it across edge-scaling regimes and varied feature structures. Across two operator families and all three differentiation orders, Sobek is faster in all 75 capacity-matched comparisons, with speedups ranging from $1.2\times$ to $49.7\times$, and reduces peak allocated memory by up to 99%. It also executes workloads up to two orders of magnitude beyond OpenEquivariance's capacity while retaining near-peak throughput. These results show that edge-scaled tensor-product workspace is a property of the conventional schedule, not of equivariant convolution itself.

Submission history #

From: Vladimir Chorošajev [[view email](/show-email/f6115ad5/2607.18074)]

**[v1]** Mon, 20 Jul 2026 15:43:21 UTC (113 KB)

References & Citations

...

Bibliographic Explorer

(What is the Explorer?) Connected Papers

(What is Connected Papers?) Litmaps

(What is Litmaps?) scite Smart Citations

(What are Smart Citations?)# Code, Data and Media Associated with this Article alphaXiv

(What is alphaXiv?) CatalyzeX Code Finder for Papers

(What is CatalyzeX?) DagsHub

(What is DagsHub?) Gotit.pub

(What is GotitPub?) Hugging Face

(What is Huggingface?) ScienceCast

(What is ScienceCast?)# Demos Influence Flower

(What are Influence Flowers?) CORE Recommender

(What is CORE?) IArxiv Recommender

(What is IArxiv?)# arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

── more in #machine-learning 4 stories · sorted by recency
── more on @sobek 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/sobek-streaming-equi…] indexed:0 read:2min 2026-07-21 ·