Greatly Accelerated Scaling of Streaming Problems with a Migrating Thread Architecture
Event Type
Workshop
Algorithms
Architectures
Big Data
Data Analytics
Memory Systems
Numerical Algorithms
W
TimeMonday, 15 November 202111am - 11:30am CST
Location224
DescriptionApplications where continuous streams of data are passed through large data structures are increasing in importance. However, their execution on conventional architectures, is highly inefficient. The primary issue is often the need to stream large numbers of disparate data items through the equivalent of very large hash tables distributed across many nodes. This paper builds on prior work on the Firehose streaming benchmark where an emerging architecture using threads that can migrate has shown to be much more efficient at such problems. This paper extends that work to use a second generation system to not only show that same improved efficiency (~10X) for larger core counts, but even significantly higher raw performance (with FPGA-based cores running at 1/10th the clock of conventional systems). Further, this additional data makes a reasonable projection that an architecture with current technology would lead to 10X performance gain on an apples-to-apples basis with conventional systems.