Saturday Feb 4, 2017 (Workshops and Tutorials)

7:30–8:30am: Breakfast (616AB) and 10:00–10:30am: Break (616AB)
PHOTONICS: Photonics-Optics Technology Oriented Networking, Information, and Computing Systems (414)
SCAW: Sensor to Cloud Architectures Workshop (417A)
First Workshop on Pioneering Processor Paradigms (415B)
12:00–1:30pm: Lunch (self) and 3:00–3:30pm: Break (616AB)
PHOTONICS: Photonics-Optics Technology Oriented Networking, Information, and Computing Systems (414)
SCAW: Sensor to Cloud Architectures Workshop (417A)
First Workshop on Pioneering Processor Paradigms (415B)
5:00pm: End

Sunday Feb 5, 2017 (Workshops and Tutorials)

7:30–8:30am: Breakfast (616AB) and 10:00–10:30am: Break (616AB)
Accelerating Big Data Processing with Hadoop, Spark, and Memcached on Datacenters with Modern Architectures (415A)


Learning gem5 Tutorial and Coding Sprint (615B)


HiPINEB: Interconnection Networks in the Exascale and Big-Data Era (417A)

12:00–1:30pm: Lunch (self) and 3:00–3:30pm: Break (616AB)
Learning gem5 Tutorial and Coding Sprint (615B)


HiPINEB: Interconnection Networks in the Exascale and Big-Data Era (417A)


An Introduction to OpenPiton, a Manycore Open Source Processor  (617)

5:00pm: End
6pm: HPCA/CGO/PPoPP Welcome Reception and Poster Session (Salon H – 6th Floor)

Main Conference Program

HPCA/CGO/PPoPP 2017 Program Schedule (Feb 4-8, 2017), Austin, TX

Monday Feb 6, 2017 (Main Program)



7:30-8:30am: Breakfast (Salon H Prefunction)

8:30-8:45am: Opening (Salon H – 6th Floor)

8:45-9:55am (Salon H – 6th Floor) – Keynote: Guy Steele (Oracle Labs): It’s Time for a New Old Language

9:55-10:20am: Break (Salon H Prefunction)




10:20-11:45am (Salon FG – 6th Floor)

HPCA Session 1: Lightning Rounds

Session Chair: Daniel A. Jiménez (Texas A&M)

10:20-11:45am (Salon J – 6th Floor)

CGO Session 1: Shared Memory

Session Chair: Evelyn Duesterwald (IBM)

Legato: End-to-End Bounded Region Serializability Using Commodity Hardware Transactional Memory

Automatic Detection of Extended Data-Race-Free Regions

FinePar: Irregularity-Aware Fine-Grained Workload Partitioning on Integrated Architectures

10:20-11:45am (400/402)

PPoPP Session 1: GPU I

Session Chair: Keshav Pingali (UT Austin)

EffiSha: A Software Framework for Enabling Efficient Preemptive Scheduling of GPU

Layout Lock: A Scalable Locking Paradigm for Concurrent Data Layout Modifications

Understanding the GPU Microarchitecture

to Achieve Bare-Metal Performance Tuning

11:45-1:15pm: Lunch (Salon H – 6th Floor)

1:15-2:55pm (Salon FG – 6th Floor)

HPCA Session 2: Best Paper Nominees

Session Chair: Yale N. Patt (UT Austin)

Towards Pervasive and User Satisfactory CNN across GPU Microarchitectures

Near-Optimal Access Partitioning for Memory Hierarchies with Multiple Heterogeneous Bandwidth Sources

NCAP: Network-Driven, Packet Context-Aware Power Management for Client-Server Architecture

Supporting Address Translation for Accelerator-Centric Architectures

1:15-2:55pm (Salon J – 6th Floor)

CGO Session 2: GPU Optimization

Session Chair: Naveen Kumar (Google)

TwinKernels: An Execution Model to Improve GPU Hardware Scheduling at Compile Time

Taming Warp Divergence

Dynamic Buffer Overflow Detection for GPGPUs

Lift: A Functional Data-Parallel IR for High-Performance GPU Code Generation

1:15-2:55pm (400/402)

PPoPP Session 2: Concurrency

Session Chair: Michael Scott (Univ. of Rochester)

Checking Concurrent Data Structures Under the C/C++11 Memory Model

Hierarchical MCS Locks with Timeout

Contention in Structured Concurrency: Provably Efficient Dynamic Non-Zero Indicators for Nested Parallelism

Noise Injection Techniques for Reproducing Subtle and Unintended Message Races

2:55-3:15pm: Break (Salon H Prefunction)

3:15-4:55pm (Salon F – 6th Floor)

HPCA Session 3A: Industrial Session

Session Chair: Chris Wilkerson (Intel)

Vulnerabilities in MLC NAND Flash Memory Programming: Experimental Analysis, Exploits, and Mitigation Techniques

Defect Analysis and Cost Effective Resilience Architecture for Future DRAM Devices

Architecting an Energy Efficient DRAM System for GPUs

Design and Analysis of an APU for Exascale Computing

BRAVO: Balanced Reliability Aware Voltage Optimization

3:15-4:55pm (Salon G – 6th Floor)

HPCA Session 3B: Cache

Session Chair: Paul Gratz (Texas A&M)

Maximizing Cache Performance Under Uncertainty

SWAP: Effective Fine-Grain Management of Shared Last-Level Caches with Minimum Hardware Support

A Split Cache Hierarchy for Enabling Data-oriented Optimizations

Fast and Accurate Exploration of Multi-Level Caches Using Hierarchical Reuse Distance

3:15-4:55pm (Salon J – 6th Floor)

CGO Session 3: Best Paper Nominees

Session Chair: Aaron Smith (Microsoft)

Synthesizing Benchmarks for Predictive Modeling

Formalizing the Concurrency Semantics of an LLVM Fragment

ThinLTO: Scalable and Incremental LTO

Automatic Generation of Fast BLAS3-GEMM: A Portable Compiler Approach

3:15-4:55pm (400/402)

PPoPP Session 3: Tools

Session Chair: Milind Chabbi (HPE)

Thread Data Sharing in Cache: Theory and Measurement

Exploiting Vector and Multicore Parallelism for Recursive Data- and Task-Parallel Programs

Isoefficiency in Practice: Configuring and Understanding the Performance of Task-based Applications

Processor-Oblivious Record and Replay

4:55-5:15pm: Break (Salon H Prefunction)

5:15-6:55pm (Salon F – 6th Floor)

HPCA Session 4A: Power, Energy & Large-Scale Computing

Session Chair: Benjamin Lee (Duke)

Enabling Effective Module-oblivious Power Gating for Embedded Processors

Application-Specific Performance-Aware Energy Optimization on Android Mobile Devices

Fast decentralized power capping for Server Clusters

Random Folded Clos Topologies for Datacenter Networks

5:15-6:55pm (Salon G – 6th Floor)

HPCA Session 4B: Memory

Session Chair: Mike Ferdman (Stony Brook)

Tiny Directory: Efficient Shared Memory in Many-core Systems with Ultra-low-overhead Coherence Tracking

Partial Row Activation for Low-Power DRAM System

Understanding and Optimizing Power Consumption in Memory Networks

SoftMC: A Flexible and Practical Open-Source Infrastructure for Enabling Experimental DRAM Studies

5:15-6:15pm (Salon J – 6th Floor)

CGO ACM Student Research Competition (SRC) Presentations

5:15-5:45pm (400/402)

CGO and PPoPP Joint Session: Artifact Evaluation Discussion

7:30-8:30pm (Salon F – 6th Floor):

HPCA Business Meeting

6:30-7:30pm (Salon J – 6th Floor): CGO Business Meeting

6:30-7:30pm (400/402):

PPoPP Business Meeting

Tuesday Feb 7, 2017 (Main Program)



7:30-8:00am: Breakfast (Salon H Prefunction – 6th Floor)




8:00-9:40am (Salon F – 6th Floor)

HPCA Session 5A: NOC

Session Chair: Vijay Nagarajan (University of Edinburgh)

Static Bubble: A Framework for Deadlock-free Irregular On-chip Topologies

Designing Low-power, Low-latency Networks-on-Chip by Optimally Combining Electrical and Optical Links

Near-Ideal Networks-on-Chip for Servers

Design and Evaluation of AWGR-based Photonic NoC Architectures for 2.5D Integrated High Performance Computing Systems

8:00-9:40am (Salon G – 6th Floor)

HPCA Session 5B: Security

Session Chair: Calvin Lin (UT Austin)


Secure Dynamic Memory Scheduling Against Timing Channel Attacks

 Cold Boot Attacks are Still Hot: Security Analysis of Memory Scramblers in Modern Processors

Cooperative Path-ORAM for Effective Memory Bandwidth Sharing in Server Settings

Camouflage: Memory Traffic Shaping to Mitigate Timing Attacks

8:25-9:40am (Salon J – 6th Floor)

CGO Session 4: Memory Dependencies

Session Chair: Ayal Zaks (Intel)

Pointer Disambiguation via Strict Inequalities

A Collaborative Dependence Analysis Framework

Characterizing Data Organization Effects on Heterogeneous Memory Architectures

8:00-9:40am (400/402)

PPoPP Session 4: GPU II

Session Chair: Angelina Lee (Washington Univ. St. Louis)

Simple, Accurate, Analytical Time Modeling and Optimal Tile Size Selection for GPGPU Stencils

Combining SIMD and Many/Multi-core Parallelism for Finite State Machines with Enumerative Speculation

S-Caffe: Co-designing MPI Runtimes and Caffe for Scalable Deep Learning on Modern GPU Clusters

Model-based Iterative CT Image Reconstruction on GPUs

9:40-10:05am: Break (Salon H Prefunction)

10:05-11:45am (Salon F – 6th Floor)

HPCA Session 6A: Emerging Storage

Session Chair: Samira Khan (University of Virginia)

SILC-FM: Subblocked InterLeaved Cache-Like Flat Memory Organization

ATOM: Atomic Durability in Non-volatile Memory through Hardware Logging

KAML: A Flexible, High-Performance Key-Value SSD

Balancing Performance and Lifetime of MLC PCM by Using a Region Retention Monitor

10:05-11:45am (Salon G – 6th Floor)

HPCA Session 6B: Scheduling

Session Chair: Miquel Pericàs (Chalmers)

Reliability-Aware Scheduling on Heterogeneous Multicore Processors

Hipster: Hybrid Task Manager for Latency-Critical Cloud Workloads

Cooper: Task Colocation with Cooperative Games

MemPod: A Clustered Architecture for Efficient and Scalable Migration in Flat Address Space Multi-Level Memories

10:05-11:45am (Salon J – 6th Floor)

CGO Session 5: Accelerators & Binary Translation

Session Chair: Milind Chabbi (HP)

Clairvoyance: Look-Ahead Compile-time Scheduling

Phase-Aware Optimization in Approximate Computing

A Space- and Energy-Efficient Code Compression/Decompression Technique for Coarse-Grained Reconfigurable Architectures

Cross-ISA Machine Emulation for Multicores

10:05-11:45am (400/402)

PPoPP Session 5: Best Paper Nominees

Session Chair: Lawrence Rauchwerger (Texas A&M Univ.)

Pagoda: Fine-Grained GPU Resource Virtualization for Narrow Tasks

Groute: An Asynchronous Multi-GPU Programming Model for Irregular Computations

Tapir: Embedding Fork-Join Parallelism into LLVM’s Intermediate Representation

A Multicore Path to Connectomics-on-Demand

11:45am-1:15pm: Lunch (Salon H – 6th Floor)

1:15-2:25pm (Salon H – 6th Floor) – Keynote: Steve Keckler (Nvidia): Everyone Needs High Performance Computing

2:25-2:50pm: Break (Salon H Prefunction)

2:50-4:30pm (Salon F – 6th Floor)

HPCA Session 7A: Novel Architectures

Session Chair: Carole-Jean Wu (Arizona State University)

Exploring Hyperdimensional Associative Memory

GraphPIM: Enabling Instruction-Level PIM Offloading in Graph Computing Frameworks

High-Bandwidth Low-Latency Approximate Interconnection Networks

Compute Caches

2:50-4:30pm (Salon G – 6th Floor)

HPCA Session 7B: Control-Flow and Microarchitecture

Session Chair: Daniel A. Jiménez (Texas A&M)

Boomerang: A Metadata-Free Architecture for Control Flow Delivery

PABST: Proportional Allocation of Bandwidth at the Source and Target

SOUP-N-SALAD: Allocation-oblivious Access Latency Reduction with Asymmetric DRAM Microarchitectures

Transparent and Efficient CFI Enforcement with Intel Processor Trace

2:50-4:30pm (Salon J – 6th Floor)

CGO Session 6: Feedback Directed and Whole Program Optimization

Session Chair: Alexandra Jimborean (Uppsala)

Incremental Whole Program Optimization and Compilation

Optimizing Function Placement for Large-Scale Data-Center Applications

Minimizing the Cost of Iterative Compilation with Active Learning

Removing Checks in Dynamically Typed Languages through Efficient Profiling

2:50-4:30pm (400/402)

PPoPP Session 6: Languages & Compilers

Session Chair: Saday Sadayppan (Ohio State University)

SC-Haskell: Sequential Consistency in Languages that Minimize Mutable Shared Heap

Synchronized-by-Default Concurrency for Shared Memory Systems

Function Call Re-Vectorization

Optimizing the Four-Index Integral Transform Using Data Movement Lower Bounds Analysis

5:00 – 9:30pm: Excursion: Salt Lick BBQ (Vegetarians Welcome!)

Buses depart at 5pm and return at 9:30pm

Wednesday Feb 8, 2017 (Main Program)


7:30-8:15am: Breakfast (Salon H Prefunction – 6th Floor)

8:15-9:25am (Salon H – 6th Floor) – Keynote: Frank Seide (Microsoft): The Computer Science Behind the Microsoft Cognitive Toolkit — an Open Source Large-Scale Deep Learning Toolkit for Windows and Linux

9:25-9:50am: Break (Salon H Prefunction – 6th Floor)





9:50-11:30am (Salon F – 6th Floor)

HPCA Session 8A: Accelerators

Session Chair: Akanksha Jain (UT Austin)

PipeLayer: A Pipelined ReRAM-Based Accelerator for Deep Learning

FlexFlow: A Flexible Dataflow Accelerator Architecture for Convolutional Neural Network

Needle : Leveraging Program Analysis to Analyze and Extract Accelerators from Whole Programs

Radiation-Induced Error Criticality in Modern HPC Parallel Accelerators

9:50-11:30am (Salon G – 6th Floor)

HPCA Session 8B: GPU Power & Energy

Session Chair: David Kaeli (Northeastern)

Pilot Register File: Energy Efficient Register File for GPUs

G-Scalar: Cost-Effective Generalized Scalar Execution Architecture for Power-Efficient GPUs

Dynamic GPGPU Power Management using Adaptive Model Predictive Control

9:50-11:30am (Salon J – 6th Floor)

CGO Session 7: Reductions & Loops

Session Chair: Michael Laurenzano (Michigan)

Discovery and Exploitation of General Reductions: A Constraint Based Approach

Parallel Associative Reductions in Halide

Optimistic Loop Optimization

Software Prefetching for Indirect Memory Accesses

9:50-11:30am (400/402)

PPoPP Session 7: Data Analytics

Session Chair: Sam Midkiff (Purdue)

Using Butterfly-Patterned Partial Sums to Draw from Discrete Distributions

KiWi: A Key-Value Map for Scalable Real-Time Analytics

Grammar-aware Parallelization for Scalable XPath Querying

Eunomia: Scaling Concurrent Search Trees under Contention Using HTM

11:30-11:45am: Break (Salon H Prefunction – 6th Floor)

11:45am-1:00pm (Salon F – 6th Floor)

HPCA Session 9A: Best of CAL

Session Chair: Nam Sung Kim (Illinois)

Hardware Support for Privacy

Efficient Execution of Bursty Applications

Non-intrusive Persistence with a Backend NVM Controller

11:45am-1:00pm (Salon G – 6th Floor)

HPCA Session 9B: GPU

Session Chair: Abdullah Muzahid (UT San Antonio)

Efficient Sequential Consistency in GPUs with Relativistic Cache Coherence

Processing-in-Memory Enabled Graphics Processors for 3D Rendering

Controlled Kernel Launch for Dynamic Parallelism in GPUs

11:45am-12:35pm (400/402)

PPoPP Session 8: Fault Tolerance

Session Chair: E.N. Elnozahy (KAUST)

Self-Checkpoint: An In-Memory Checkpoint Method Using Less Space and Its Practice on Fault-Tolerant HPL

Silent Data Corruption Resilient Two-sided Matrix Factorizations

1:00pm-1:15pm (Salon F – 6th Floor):

HPCA Closing & Best Paper Award Announcement

11:30-11:45am (Salon J – 6th Floor):

CGO Closing & Best Paper Award Announcement

12:35-12:50pm (400/402)

PPoPP Closing & Best Paper Award Announcement