ACM Computing Surveys (CSUR)
Better I/O through byte-addressable, persistent memory
Proceedings of the ACM SIGOPS 22nd symposium on Operating systems principles
Proceedings of the 2010 ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis
High performance solid state storage under Linux
MSST '10 Proceedings of the 2010 IEEE 26th Symposium on Mass Storage Systems and Technologies (MSST)
Moneta: A High-Performance Storage Array Architecture for Next-Generation, Non-volatile Memories
MICRO '43 Proceedings of the 2010 43rd Annual IEEE/ACM International Symposium on Microarchitecture
Operating system implications of fast, cheap, non-volatile memory
HotOS'13 Proceedings of the 13th USENIX conference on Hot topics in operating systems
Onyx: a protoype phase change memory storage array
HotStorage'11 Proceedings of the 3rd USENIX conference on Hot topics in storage and file systems
SCMFS: a file system for storage class memory
Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis
DiDi: Mitigating the Performance Impact of TLB Shootdowns Using a Shared TLB Directory
PACT '11 Proceedings of the 2011 International Conference on Parallel Architectures and Compilation Techniques
When poll is better than interrupt
FAST'12 Proceedings of the 10th USENIX conference on File and Storage Technologies
Hi-index | 0.00 |
Many I/O subsystems of general-purpose operating systems are optimized for the hard disk drive. The "writeback" feature is being used to overcome the mechanical limitation of the hard disk. But the current "writeback" mechanism is not optimal for memory-based storage devices which do not use mechanical instruments to read and write data. We have investigated the current "writeback" mechanism of Linux operating system and attempted to optimize it for memory-based storage devices. The current "writeback" handles bios one by one based on the mechanical limitation of the hard disk even though there may be many bios that could be handled simultaneously; the bios could be handled differently for memory-based storage devices. We have modified "writeback" so that multiple bios can be processed at a time, leading to improvements by more than 140% in write performance.