2,977 research outputs found
Compiler Infrastructure for Specializing Domain-Specific Memory Templates
No full textSpecialized hardware accelerators are becoming important for more and more
applications. Thanks to specialization, they can achieve high performance and
energy efficiency but their design is complex and time consuming. This problem
is exacerbated when large amounts of data must be processed, like in modern big
data and machine learning applications. The designer has not only to optimize
the accelerator logic but also produce efficient memory architectures. To
simplify this process, we propose a multi-level compilation flow that
specializes a domain-specific memory template to match data, application, and
technology requirements
Stephanie Mathson interviews poet and author Judith Kerman
No full textPoet and author Judith Kerman talks about her experience as a Fulbright scholar in the Dominican Republic, her work translating poems by Cuban poet Dulce Mar\ueda Loynaz, learning Spanish, translating poems from Spanish, and her book "Retrofitting Blade Runner". Kerman is interviewed by Stephanie Mathson of the Michigan State University Libraries. Part of the MSU Libraries' Michigan Writers Series
Stephanie Mathson interviews poet and author Jack Ridl
No full textPoet and author Jack Ridl explains how he began writing, the writer series at Hope College, his coach poems, his chapbook "Against elegies," how working and living in Michigan shapes his work, and works in progress. Ridl is interviewed by Stephanie Mathson of the Michigan State University Libraries. Part of the MSU Libraries' Michigan Writers Series
Stephanie Mathson interviews poet and author Josie Kearns
No full textPoet and author Josie Kearns, professor of creative writing and literature at the University of Michigan, talks about teaching and writing, natural scenery in Michigan, her editorship of the book "New Poems From the Third Coast", her book "New Numbers", and other works in process. Kearns is interviewed by Stephanie Mathson from the Michigan State University Libraries for the MSU Libraries' Michigan Writers Series
Kara Gust interviews author and bioregionalist Stephanie Mills
No full textAuthor and ecologist Stephanie Mills talks about how she started writing and publishing, writing on nature and the environment, the challenges of being a writer, the influence of Michigan on her work, bio-regionalism, and a new book she is working on. Mills is interviewed by Michigan State University Librarian Kara Gust for the Michigan State University Libraries' Michigan Writers Series
Making a market for Miscanthus: Can new contract designs solve the biofuel investment hold-up problem?
No full textWe present designs for optimal contracts to solve the investment hold-up problem for perennial crops for the biofuel industry. A fixed-price contract is ex-ante efficient but renegotiation-proof for a limited range of discount parameters. A perfectly- indexed contract is both renegotiation-proof and ex-post efficient. Provided long-run land prices are stationary, the expected cost for both contracts converges to the long-run expected price of land for a risk-neutral farmer.Biofuels, Miscanthus, contract theory, industrial organization, renegotiation-proof contract, Marketing,
Author and bioregionalist Stephanie Mills reads her selected works at the Michigan Writers Series
No full textAuthor and ecologist Stephanie Mills reads from her first book "Whatever happened to ecology?" and from "Tough little beauties," then answers questions from the audience. The event is convened by Peter Berg, head of Michigan State University Libraries' Special Collections. Part of the Michigan State University Libraries' Michigan Writers Series. Held in the Main Library
Automatic Creation of High-Bandwidth Memory Architectures from Domain-Specific Languages: The Case of Computational Fluid Dynamics
No full textNumerical simulations can help solve complex problems. Most of these
algorithms are massively parallel and thus good candidates for FPGA
acceleration thanks to spatial parallelism. Modern FPGA devices can leverage
high-bandwidth memory technologies, but when applications are memory-bound
designers must craft advanced communication and memory architectures for
efficient data movement and on-chip storage. This development process requires
hardware design skills that are uncommon in domain-specific experts. In this
paper, we propose an automated tool flow from a domain-specific language (DSL)
for tensor expressions to generate massively-parallel accelerators on
HBM-equipped FPGAs. Designers can use this flow to integrate and evaluate
various compiler or hardware optimizations. We use computational fluid dynamics
(CFD) as a paradigmatic example. Our flow starts from the high-level
specification of tensor operations and combines an MLIR-based compiler with an
in-house hardware generation flow to generate systems with parallel
accelerators and a specialized memory architecture that moves data efficiently,
aiming at fully exploiting the available CPU-FPGA bandwidth. We simulated
applications with millions of elements, achieving up to 103 GFLOPS with one
compute unit and custom precision when targeting a Xilinx Alveo U280. Our FPGA
implementation is up to 25x more energy-efficient than expert-crafted Intel CPU
implementations.Comment: Accepted for publication in ACM Transactions on Reconfigurable
Technology and Systems (TRETS
From Domain-Specific Languages to Memory-Optimized Accelerators for Fluid Dynamics
No full textMany applications are increasingly requiring numerical simulations for solving complex problems. Most of these numerical algorithms are massively parallel and often implemented on parallel high-performance computers. However, classic CPU-based platforms suffer due to the demand for higher resolutions and the exponential growth of data. FPGAs offer a powerful and flexible alternative that can host accelerators to complement such platforms. Developing such application-specific accelerators is still challenging because it is hard to provide efficient code for hardware synthesis. In this paper, we study the challenges of porting a numerical simulation kernel onto FPGA.
We propose an automated tool flow from a domain-specific language (DSL) to generate accelerators for computational fluid dynamics on FPGA. Our DSL-based flow simplifies the exploration of parameters and constraints such as on-chip memory usage.
We also propose a decoupled optimization of memory and logic resources, which allows us to better use the limited FPGA resources.
In our preliminary evaluation, this enabled doubling the number of parallel kernels, increasing the accelerator speedup versus ARM execution from 7 to 12 times
Stephanie Mathson interviews essayist and memoirist Robert Root
No full textEssayist and memoirist Robert Root, professor of English at Central Michigan University, talks about his book "Recovering Ruth" and the genealogical research research in his work and his role as both a university professor and an author. He also shares his views on creative nonfiction, Michigan as a source of inspiration, and works in progress. Root is interviewed by Stephanie Mathson of the Michigan State University Libraries for the MSU Libraries' Michigan Writers Series
- …
