Hybrid-core systems speedup applications by offloading certain compute operations that
can run faster on hardware accelerators. However, such systems require significant
programming and porting effort to gain a performance benefit from the accelerators.
Therefore, prior to porting it is prudent to investigate the predicted performance benefit
of accelerators for a given workload. To address this problem we present a
performance-modeling framework that predicts the application performance rapidly and
accurately for hybrid-core systems. We present predictions for two full-scale HPC
applications—HYCOM and Milc. Our results for two accelerators (GPU and FPGA) show that
gather/scatter and stream operations can speedup by as much as a factor of 15 and overall
compute time of Milc and HYCOM improve by 3.4% and 20%, respectively. We also show that in
order to benefit from the accelerators, 70% of the latency of data transfer time between
the CPU and the accelerators needs to be overcome.
AlamSBhatiaNVetterJ (2007) An exploration of performance
attributes for symbolic modeling of emerging processing devices. In:
Proceedings of the 3 rd international high performance computing and
communications (HPCC) (eds Perrot.), Houston, USA, 2007, pp.
683–694. Springer:
New York.
2.
AlexandrovAIonescuMFSchauserKE (1997) LogGP: incorporating long messages into
the LogP Model. Journal of Parallel and Distributed
Computing44(1): 71–79.
3.
AsanovicKBodikRCatanzaroBC. (2006) The landscape of parallel computing
research: a view from Berkeley. Technical Report
UCB/EECS-2006-183, EECS Department, University of
California, Berkeley.
4.
BaileyDHBarszczEBartonJT (1991) The NAS parallel
benchmarks. International Journal of Supercomputer
Applications5(3): 66–73.
5.
BakhodaAYuanGLFungWWL (2009) Analyzing CUDA Workloads using a
Detailed GPU Simulator. In: IEEE international symposium on
performance analysis of systems and software,
ISPASS, Boston, USA,
2009.
6.
BakosJ (2010) High-performance heterogeneous
computing with the Convey HC-1. Computing in Science and
Engineering12(6): 80–87.
7.
BrewerTM (2010) Instruction Set Innovations for the
Convey HC-1 Computer. IEEE Micro30: 70–79.
8.
BinkertNBeckmannBBlackG (2011) The gem5 simulator.
SIGARCH Computer Architecture News39(2): 1–7.
9.
BurgerDAustinTM (1997) The SimpleScalar tool set, version
2.0. SIGARCH Computer Architecture News25(3): 13–25.
10.
CarringtonLLaurenzanoMSnavelyA. (2005) How well can simple metrics represent
the performance of HPC applications? In: Proceedings of the 2005
ACM/IEEE conference on high performance networking and computing (SC’05),
Seattle, USA.
11.
CarringtonLTikirMMOlschanowskyC. (2011) An idiom-finding tool for increasing
productivity of accelerators. In: Proceedings of the 2011
international conference on supercomputing (ICS 2011),
Tucson,
USA.
12.
CullerDEKarpRMPattersonD (1996) LogP: a practical model of parallel
computation. Communications of the ACM39(11): 78–85.
13.
GibbonsPBMatiasYRamachandranV (1998) The Queue-Read Queue-Write PRAM Model:
accounting for contention in parallel algorithms. SIAM Journal
of Computing28(2): 733–769.
14.
GovindarajuNLarsonJGrayJ. (2006) A memory model for scientific
algorithms on graphics processors. In: Proceedings of the 2006
ACM/IEEE conference on high performance networking and computing (SC’06),
Tampa,
Florida.
15.
Graph500 (2012) Brief
introduction to Graph500. Available at: www.graph500.org.
16.
GreggCHazelwoodK (2011) Where is the data? Why You cannot
debate GPU vs. CPU performance without the answer. In:
Proceedings of the international symposium on performance analysis of systems
and software (ISPASS), Austin, USA.
17.
HeJSnavelyAEVan Der WijngaartRF (2011) Automatic recognition of performance
idioms in scientific applications. In: Proceedings of the 25th
IEEE international parallel and distributed processing symposium (IPDPS),
2011, Anchorage,
USA.
18.
HongSKimH (2009) An analytical model for a GPU
architecture with memory-level and thread-level parallelism awareness.
In: Proceedings of the 36th international symposium on computer architecture
(ISCA), 2009, Austin,
Texas.
KerbysonDJHoisieAWassermanHJ (2003) Modelling the performance of large
scale systems. Keynote paper,
UK performance engineering workshop, July
2003.
22.
LaurenzanoMATikirMMCarringtonL (2010) PEBIL efficient static binary
instrumentation for Linux. In: Proceedings of the international
symposium on performance analysis of systems and software (ISPASS),
2010, White Plains,
USA.
23.
LuszczekPDongarraJKoesterD (2005) Introduction to the HPC Challenge
Benchmark Suite. Available at: http://icl.cs.utk.edu/hpcc/pubs.
24.
MagnussonPChristenssonMEskilsonJ (2002) Simics: A full system simulation
platform. Computer35(2): 50–58.
25.
MendesCLReedDA (1998) Integrated compilation and scalability
analysis for parallel systems. In: Proceedings of the 1998
international conference on parallel architectures and compilation techniques (PACT
'98), IEEE Computer Society,
Washington, USA.
OlschanowskyCSnavelyAMeswaniM. (2010) PIR: a static idiom
recognizer. In: Proceedings of the first international workshop
on parallel software tools and tool infrastructures (PSTI), San
Diego, USA.
29.
PerelmanEHamerlyGBiesbrouckMV. (2003) Using SimPoint for accurate and
efficient simulation. In: Proceedings of ACM SIGMETRICS the
international conference on measurement and modeling of computer systems,
June 2003.
30.
PllanaSBrandicIBenknerS (2007) Performance modeling and prediction of
parallel and distributed computing systems: a survey of the state of the
art. In: Proceedings of the first international conference on
complex, intelligent and software intensive systems (CISIS '07),
IEEE Computer Society, Washington,
USA, pp. 279–284.
31.
SaavedraRHSmithAJ (1995) Measuring cache and TLB performance and
their effect on benchmark run times. IEEE Transactions on
Computers44(10):
1223–1235.
32.
SaavedraRHSmithAJ (1996) Analysis of benchmark characteristics
and benchmark performance prediction. TOCS14(4): 344–384.
33.
SnavelyACarringtonLWolterN. (2002) A framework for application
performance modeling and prediction. In: Proceedings of the 2002
ACM/IEEE conference on high performance networking and computing
(SC’02).
34.
SvobodovaL (1976) Computer System Performance Measurement and
Evaluation Methods: Analysis and Applications. New
York: Elsevier.
35.
TikirMMCarringtonLSnavelyA. (2007) Genetic algorithm approach to modeling
the performance of memory-bound codes. In: Proceeding of the
2007 ACM/IEEE Conference on High Performance Networking and Computing
(SC’07).
36.
TikirMMLaurenzanoMCarringtonC. (2009) PSINS: an open source event tracer and
execution simulator for MPI applications. In: Proceedings of the
European Conference on Parallel Computing (EuroPar).
37.
VachharajaniMVachharajaniNPenryD. (2002) Microarchitectural exploration with
liberty. In: Proceedings of the international symposium on
microarchitecture.
38.
WaltersJPQudahBChaudharyV (2006) Accelerating the HMMER sequence
analysis suite using conventional processors. In: Proceedings of
the 20th international conference on advanced information networking and applications
(AINA'06).
39.
WilliamsSWatermanAPattersonD (2009) Roofline: an insightful visual
performance model for multicore architectures. Communications of
the ACM52(4): 65–76.
40.
XuZZhangXSunL (1996) Semi-empirical multiprocessor
performance predictions. Journal of Parallel and Distributed
Computing39: 14–28.
41.
YiJJLiljaDJ (2006) Simulation of computer architectures:
simulators, benchmarks, methodologies, and recommendations. IEEE
Transactions on Computers55(3): 268–280.
