What is it about?
This work defines a new way for GPUs to handle data in persistent memory (PM), allowing faster and more flexible recovery from system crashes. It creates a model that balances performance and consistency by leveraging GPU-specific capabilities like scopes and buffers.
Featured Image
Photo by Andrey Matveev on Unsplash
Why is it important?
This paper marks a step in combining the persistence of NVM with the parallelism of GPUs to improve applications which could rely on both.
Read the Original
This page is a summary of: Scoped Buffered Persistency Model for GPUs, January 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3575693.3575749.
You can read the full text:
Contributors
The following have contributed to this page
