NDLI: DjiNN and Tonic:DNN as a service and its implications for future warehouse scale computers

Please wait, while we are loading the content...

Volume 43, Issue 5, December 2015

Volume 43, Issue 4, September 2015

Volume 43, Issue 3, May 2015

Volume 43, Issue 3, June 2015

A case for core-assisted bottleneck acceleration in GPUs

Redundant memory mappings for fast access to large memories

Quantitative comparison of hardware transactional memory for Blue Gene/Q, zEnterprise EC12, Intel Core, and POWER8

Cost-effective speculative scheduling in high performance processors

Exploring the potential of heterogeneous von neumann/dataflow execution models

PIM-enabled instructions

Thermal time shifting

A variable warp size architecture

Manycore network interfaces for in-memory rack-scale computing

Towards sustainable in-situ server systems in the big data era

A scalable processing-in-memory accelerator for parallel graph processing

Profiling a warehouse-scale computer

A fully associative, tagless DRAM cache

LaZy superscalar

SHRINK: Reducing the ISA complexity via instruction recycling

Reducing world switches in virtualized environment with flexible cross-world calls

Warped-compression

Unified address translation for memory-mapped SSDs with FlashMap

Accelerating asynchronous programs through event sneak peek

Coherence protocol for transparent management of scratchpad memories in shared memory manycore architectures

DjiNN and Tonic

Efficient execution of memory access phases using dataflow specialization

Computer performance microscopy with Shim

Multiple clone row DRAM

The load slice core microarchitecture

Branch vanguard

Data reorganization in memory using 3D-stacked DRAM

Flexible software profiling of GPU architectures

Flexible auto-refresh

Semantic locality and context-based prefetching using reinforcement learning

Architecting to achieve a billion requests per second throughput on a single key-value store server platform

Dynamic thread block launch

Volume 43, Issue 1, March 2015

DjiNN and Tonic:DNN as a service and its implications for future warehouse scale computers

Content Provider	ACM Digital Library
Author	Mudge, Trevor Mars, Jason Tang, Lingjia Chen, Quan Dreslinski, Ronald G. Hauswald, Johann Li, Cheng Laurenzano, Michael A. Kang, Yiping
Abstract	As applications such as Apple Siri, Google Now, Microsoft Cortana, and Amazon Echo continue to gain traction, web-service companies are adopting large deep neural networks (DNN) for machine learning challenges such as image processing, speech recognition, natural language processing, among others. A number of open questions arise as to the design of a server platform specialized for DNN and how modern warehouse scale computers (WSCs) should be outfitted to provide DNN as a service for these applications. In this paper, we present DjiNN, an open infrastructure for DNN as a service in WSCs, and Tonic Suite, a suite of 7 end-to-end applications that span image, speech, and language processing. We use DjiNN to design a high throughput DNN system based on massive GPU server designs and provide insights as to the varying characteristics across applications. After studying the throughput, bandwidth, and power properties of DjiNN and Tonic Suite, we investigate several design points for future WSC architectures. We investigate the total cost of ownership implications of having a WSC with a disaggregated GPU pool versus a WSC composed of homogeneous integrated GPU servers. We improve DNN throughput by over 120x for all but one application (40x for Facial Recognition) on an NVIDIA K40 GPU. On a GPU server composed of 8 NVIDIA K40s, we achieve near-linear scaling (around 1000x throughput improvement) for 3 of the 7 applications. Through our analysis, we also find that GPU-enabled WSCs improve total cost of ownership over CPU-only designs by 4-20x, depending on the composition of the workload
Starting Page	27
Ending Page	40
Page Count	14
File Format	PDF
ISSN	01635964
DOI	10.1145/2872887.2749472
Journal	ACM SIGARCH Computer Architecture News (CARN)
Volume Number	43
Issue Number	3
Language	English
Publisher	Association for Computing Machinery (ACM)
Publisher Date	1981-04-01
Publisher Place	New York
Access Restriction	One Nation One Subscription (ONOS)
Content Type	Text
Resource Type	Article

Central Library (ISO-9001:2015 Certified)
Indian Institute of Technology Kharagpur
Kharagpur, West Bengal, India | PIN - 721302

See location in the Map
03222 282435
Mail: support@ndl.gov.in

Sl.	Authority	Responsibilities	Communication Details
1	Ministry of Education (GoI), Department of Higher Education	Sanctioning Authority	https://www.education.gov.in/ict-initiatives
2	Indian Institute of Technology Kharagpur	Host Institute of the Project: The host institute of the project is responsible for providing infrastructure support and hosting the project	https://www.iitkgp.ac.in
3	National Digital Library of India Office, Indian Institute of Technology Kharagpur	The administrative and infrastructural headquarters of the project	Dr. B. Sutradhar bsutra@ndl.gov.in
4	Project PI / Joint PI	Principal Investigator and Joint Principal Investigators of the project	Dr. B. Sutradhar bsutra@ndl.gov.in Prof. Saswat Chakrabarti will be added soon
5	Website/Portal (Helpdesk)	Queries regarding NDLI and its services	support@ndl.gov.in
6	Contents and Copyright Issues	Queries related to content curation and copyright issues	content@ndl.gov.in
7	National Digital Library of India Club (NDLI Club)	Queries related to NDLI Club formation, support, user awareness program, seminar/symposium, collaboration, social media, promotion, and outreach	clubsupport@ndl.gov.in
8	Digital Preservation Centre (DPC)	Assistance with digitizing and archiving copyright-free printed books	dpc@ndl.gov.in
9	IDR Setup or Support	Queries related to establishment and support of Institutional Digital Repository (IDR) and IDR workshops	idr@ndl.gov.in