By utilizing diverse heterogeneous hardware resources, developers can significantly improve the performance of their applications. Currently, in order to determine which parts of an application suit a particular type of hardware accelerator better, an offline analysis that uses a priori knowledge of the target hardware configuration is necessary. To make matters worse, the above process has to be repeated every time the application or the hardware configuration changes.
This paper introduces TornadoVM, a virtual machine capable of reconfiguring applications, at runtime, for hardware acceleration based on the currently available hardware resources. Through TornadoVM, we introduce a new level of compilation in which applications can benefit from heterogeneous hardware. We showcase the capabilities of TornadoVM by executing a complex computer vision application and six
benchmarks on a heterogeneous system that includes a CPU, an FPGA, and a GPU. Our evaluation shows that by using dynamic reconfiguration, we achieve an average of 7.7× speedup over the statically-configured accelerated code.
Sun 14 AprDisplayed time zone: Eastern Time (US & Canada) change
16:00 - 18:05 | |||
16:00 25mTalk | QuickCheck: Using Speculation to Reduce the Overhead of Checks in NVM Frameworks Research Papers Thomas Shull University of Illinois at Urbana-Champaign, Jian Huang University of Illinois at Urbana-Champaign, Josep Torrellas University of Illinois at Urbana-Champaign | ||
16:25 25mTalk | Tail Latency in Node.js: Energy Efficient Turbo Boosting for Long Tail Requests in JavaScript Research Papers Wenzhi Cui Google, Daniel Richins The University of Texas at Austin, Yuhao Zhu University of Rochester, Vijay Janapa Janapa Reddi Harvard University | ||
16:50 25mTalk | Dynamic Application Reconfiguration on Heterogeneous Hardware Research Papers Juan Fumero University of Manchester, UK, Michail Papadimitriou University of Manchester, UK, Foivos S. Zakkak University of Manchester, UK, Maria Xekalaki University of Manchester, UK, James Clarkson University of Manchester, UK, Christos Kotselidis University of Manchester, UK DOI Authorizer link | ||
17:15 25mTalk | vSocket: Virtual Socket Interface for RDMA in Public Clouds Research Papers Dongyang Wang University of Science and Technology of China, China, Binzhang Fu Huawei Technologies, n.n., Gang Lu Huawei Technologies, n.n., Kun Tan Huawei Technologies, n.n., Bei Hua Huawei Technologies, n.n. / University of Science and Technology of China, China | ||
17:40 25mTalk | vCPU as a Container: Towards Accurate CPU Allocation for VMs Research Papers Li Liu George Mason University, USA, Haoliang Wang Adobe Research, USA, An Wang Case Western Reserve University, USA, Mengbai Xiao Ohio State University, USA, Yue Cheng George Mason University, USA, Songqing Chen George Mason University, USA | ||