REVIEW 2 cited by
Ace-TN: GPU-Accelerated Corner-Transfer-Matrix Renormalization of Infinite Projected Entangled-Pair States
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Ace-TN: GPU-Accelerated Corner-Transfer-Matrix Renormalization of Infinite Projected Entangled-Pair States
read the original abstract
The infinite projected entangled-pair state (iPEPS) ansatz is a powerful tensor-network approximation of an infinite two-dimensional quantum many-body state. Tensor-based calculations are particularly well-suited to utilize the high parallel efficiency of modern GPUs. We present Ace-TN, a modular and easily extendable open-source library developed to address the current need for an iPEPS framework focused on GPU acceleration. We demonstrate the advantage of using GPUs for the core iPEPS simulation methods and present a simple parallelization scheme for efficient multi-GPU execution. The latest distribution of Ace-TN can be obtained at https://github.com/ace-tn/ace-tn.
Forward citations
Cited by 2 Pith papers
-
Fast two-dimensional tensor-network contraction via subspace iteration
A new CTMRG variant, SI-CTMRG, substitutes QR-based subspace iteration for the dominant large SVD, shifting cost to tensor contractions and enabling state-of-the-art iPEPS calculations on a single H100 GPU.
-
Accelerating two-dimensional tensor network contractions using QR decompositions
A QR-based CTMRG variant accelerates iPEPS contractions by up to two orders of magnitude on GPUs with no accuracy loss for the Heisenberg and J1-J2 models.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.