We carry out a comparative performance study of multi-core CPUs, GPUs and Intel Xeon Phi (Many Integrated Core (MIC)) with a microscopy image analysis application. We experimentally evaluate the performance of computing devices on core operations of the application. We correlate the observed performance with the characteristics of computing devices and data access patterns, computation complexities, and parallelization forms of the operations. The results show a significant variability in the performance of operations with respect to the device used. The performances of operations with regular data access are comparable or sometimes better on a MIC than that on a GPU. GPUs are more efficient than MICs for operations that access data irregularly, because of the lower bandwidth of the MIC for random data accesses. We propose new performance-aware scheduling strategies that consider variabilities in operation speedups. Our scheduling strategies significantly improve application performance compared with classic strategies in hybrid configurations.
AugonnetCAumageOFurmentoNNamystRThibaultS (2012) StarPU-MPI: task programming over clusters of machines enhanced with accelerators. In: The 19th European MPI Users’ Group meeting (EuroMPI 2012).
2.
AugonnetCThibaultSNamystRWacrenierPA (2009) Starpu: A unified platform for task scheduling on heterogeneous multicore architectures. In: Euro-Par’09: proceedings of the 15th international Euro-Par conference on parallel processing, pp. 863–874. DOI:10.1007/978-3-642-03869-3_80.
3.
BosilcaGBouteillerAHeraultTLemarinierPSaengpatsaNTomovSDongarraJ (2011) Performance portability of a GPU enabled factorization with the DAGuE framework. In: IEEE international conference on cluster computing, pp. 395–402. DOI:10.1109/CLUSTER.2011.51.
BuenoJPlanasJDuranABadiaRMartorellXAyguadeELabartaJ (2012) Productive programming of GPU clusters with OmpSs. In: 2012 IEEE 26th international parallel distributed processing symposium (IPDPS), pp. 557–568. DOI:10.1109/IPDPS.2012.58.
6.
ChandolaVVatsavaiRR (2011) A scalable Gaussian process analysis algorithm for biomass monitoring. Statistical Analysis and Data Mining4(4): 430–445. DOI:10.1002/sam.10129.
7.
CooperLKongJGutmanDWangFCholletiSPanTWidenerPSharmaAMikkelsenTFlandersARubinDVan MeirEKurcTMorenoCBratDSaltzJ (2010) An integrative approach for in silico glioma research. IEEE Transactions on Biomedical Engineering57(10): 2617–2621. DOI:10.1109/TBME.2010.2060338.
8.
CooperLAKongJGutmanDAWangFGaoJAppinCCholletiSPanTSharmaAScarpaceLMikkelsenTKurcTMorenoCSBratDJSaltzJH (2012) Integrated morphologic analysis for the identification and characterization of disease subtypes. Journal of the American Medical Information Association2: 317–323.
9.
CramerTSchmidlDKlemmMan MeyD (2012) OpenMP Programming on Intel Xeon Phi coprocessors: an early performance comparison. In: Many-core applications research community symposium.
10.
DeanJGhemawatS (2004) MapReduce: Simplified data processing on large clusters. In: The sixth symposium on operating system design and implementation, pp. 137–150.
11.
DiamosGFYalamanchiliS (2008) Harmony: an execution model and runtime for heterogeneous many core systems. In: Proceedings of the 17th international symposium on high-performance distributed computing (HPDC). DOI:10.1145/1383422.1383447.
12.
EisenlohrJHudakDETomkoKPrinceTC (2012) Dense linear algebra factorization in OpenMP and Cilk Plus on Intel’s MIC architecture: development experiences and performance analysis. In: TACC-Intel highly parallel computing symposium
13.
HamidoucheKPotluriSSubramoniHKandallaKPandaDK (2013) MIC-RO: enabling efficient remote offload on heterogeneous many integrated core (MIC) clusters with InfiniBand. In: 27th international ACM international conference on supercomputing (ICS’13), pp. 399–408. DOI:10.1145/2464996.2465445.
14.
HartleyTCatalyurekURuizAUjaldonMIgualFMayoR (2008) Biomedical image analysis on a cooperative cluster of GPUs and multicores. In: 22nd ACM international conference on supercomputing.
15.
HartleyTSauleECatalyurekU (2010) Automatic dataflow application tuning for heterogeneous systems. In: HiPC. IEEE, pp. 1–10.
16.
HeBFangWLuoQGovindarajuNKWangT (2008) Mars: A MapReduce framework on graphics processors. In: Proceedings of the 17th International Conference on Parallel Architectures and Compilation Techniques, PACT '08, pp. 260–269.
17.
HeineckeAVaidyanathanKSmelyanskiyM (2013) Design and implementation of the Linpack benchmark for single and multi-node systems based on Intel Xeon Phi coprocessor. In: The 27th IEEE international symposium on parallel distributed processing.
18.
HuoXRaviVAgrawalG (2011) Porting irregular reductions on heterogeneous CPU–GPU configurations. In: 18th international conference on high performance computing (HiPC), pp. 1–10. DOI:10.1109/HiPC.2011.6152715.
19.
JoóBKalamkarDVaidyanathanKSmelyanskiyMPamnanyKLeeVDubeyPWatsonIW (2013) Lattice QCD on Intel Xeon Phi coprocessors. In: Supercomputing (Lecture Notes in Computer Science, vol. 7905). New York: Springer, pp. 40–54. DOI:10.1007/978-3-642-38750-0_4.
20.
KongJCooperLAWangFGaoJTeodoroGScarpaceL (2013a) Machine-based morphologic analysis of glioblastoma using whole-slide pathology images uncovers clinically relevant molecular correlates. PloS One8(11): e81049.
21.
KongJWangFTeodoroGCooperLADMorenoCSKurçTMPanTSaltzJHBratDJ (2013b) High-performance computational analysis of glioblastoma pathology images with database support identifies molecular and survival correlates. In: 2013 IEEE international conference on bioinformatics and biomedicine, Shanghai, China, 18–21 December 2013, pp. 229–236. DOI:10.1109/BIBM.2013.6732495.
22.
KumarSKimDSmelyanskiyMChenYKChhuganiJ. (2008) Atomic vector operations on chip multiprocessors. In: 35th international symposium on computer architecture. DOI:10.1109/ISCA.2008.38.
23.
LindermanMDCollinsJDWangHMengTH (2008) Merge: a programming model for heterogeneous multi-core systems. SIGPLAN Notices43(3): 287–296. DOI:10.1145/1353536.1346318.
24.
LukCKHongSKimH (2009) Qilin: Exploiting parallelism on heterogeneous multiprocessors with adaptive mapping. In: 42nd international symposium on microarchitecture.
25.
McCalpinJD (1995) Memory bandwidth and machine balance in current high performance computers. IEEE Computer Society Technical Committee on Computer Architecture Newsletter, pp. 19–25.
OliveiraVMAde Alencar LotufoR (2010) A study on connected components labeling algorithms using GPUs. In: SIBGRAPI.
28.
ParasharMMuralidharRLeeWArnoldDCDongarraJWheelerMF (2005) Enabling interactive and collaborative oil reservoir simulations on the grid. Concurrency - Practice and Experience17: 1387–1414.
29.
RaviVMaWChiuDAgrawalG (2010) Compiler and runtime support for enabling generalized reduction computations on heterogeneous parallel configurations. In: The 24th ACM international conference on supercomputing, pp. 137–146.
30.
RossbachCJCurreyJSilbersteinMRayBWitchelE (2011) PTask: operating system abstractions to manage GPUs as compute devices. In: The 23rd ACM symposium on operating systems principles, pp. 233–248. DOI:10.1145/2043556.2043579.
31.
RuifrokACJohnstonDA (2001) Quantification of histochemical staining by color deconvolution. Analytical and Quantitative Cytology and Histology23(4): 291–299.
32.
SaltzJTeodoroGPanTCooperLKongJKlaskySKurcT (2013) Feature-based analysis of large-scale spatio-temporal sensor data on hybrid architectures. International Journal of High Performance Computing Applications27(3): 263–272.
33.
SauleEKayaKÇatalyürekÜV (2013) Performance Evaluation of Sparse Matrix Multiplication Kernels on Intel Xeon Phi. CoRR abs/1302.1078.
34.
SharpRRidgwayRIyengarSGulacyAWenzelPde BruinAMachirajuRHuangKLeoneGSaltzJH (2006) Registration and 3D visualization of large microscopy images. In: Proceedings of the SPIE annual medical imaging meetings, pp. 923–934.
35.
ShowalterPS (2001) Remote sensing in disaster research: A review. Disaster Prevention and Management10: 21–29.
36.
TarjanRE (1975) Efficiency of a good but not linear set union algorithm. Journal of the ACM22(2): 215–225. DOI:10.1145/321879.321884.
37.
TeodoroGHartleyTCatalyurekUFerreiraR (2010) Run-time optimizations for replicated dataflows on heterogeneous environments. In: The 19th ACM international symposium on high performance distributed computing, pp. 13–24.
TeodoroGKurcTKongJCooperLSaltzJ (2014) Comparative performance analysis of Intel (R) Xeon Phi (TM), GPU, and CPU: a case study from microscopy image analysis. In: 2014 IEEE 28th international parallel and distributed processing symposium (IPDPS’14), pp. 1063–1072. DOI:10.1109/IPDPS.2014.111.
40.
TeodoroGKurcTPanTCooperLKongJWidenerPSaltzJ (2012b) Accelerating large scale image analyses on parallel, CPU–GPU equipped systems. In: 26th IEEE international parallel and distributed processing symposium, pp. 1093–1104.
TeodoroGPanTKurcTMKongJCooperLAPodhorszkiNKlaskySSaltzJH (2013b) High-throughput analysis of large microscopy image datasets on CPU–GPU cluster platforms. In: IPDPS’13: proceedings of the 2013 IEEE international symposium on parallel and distributed processing.
43.
TeodoroGSachettoRSertelOGurcanMMeiraWJrCatalyurekUFerreiraR (2009) Coordinating the use of GPU and CPU for improving performance of compute intensive applications. In: Proceedings of the 2009 IEEE international conference on cluster computing, New Orleans, LA, 31 August–4 September 2009, pp. 1–10.
44.
TeodoroGValleEMarianoNTorresRMeira JWagnerSaltzJ (2013c) Approximate similarity search for online multimedia services on distributed CPU–GPU platforms. The VLDB Journal. DOI:10.1007/s00778-013-0329-7.
45.
UpadhyayGRaySPanigrahyS (2008) Derivation of crop phenological parameters using multi-date SPOT-VGT-NDVI data: A case study for Punjab. Journal of the Indian Society of Remote Sensing36: 37–50.
46.
VatsavaiRBhaduriB (2011a) A hybrid classification scheme for mining multisource geospatial data. GeoInformatica15(1): 29–47. DOI:10.1007/s10707-010-0113-4.
47.
VatsavaiRBhaduriB (2011b) A hybrid classification scheme for mining multisource geospatial data. Geoinformatica15: 29–47.
48.
VatsavaiRShekharSBhaduriBL (2008) A learning scheme for recognizing sub-classes from model trained on aggregate classes. In: Structural, Syntactic, and Statistical Pattern Recognition (Lecture Notes in Computer Science, vol. 5342). New York: Springer, pp. 967–976.
49.
VatsavaiRTuttleMBhaduriBBrightECheriyadatAChandolaVGraesserJ (2011) Rapid damage assessment using high-resolution remote sensing imagery: Tools and techniques. In: 2011 IEEE international geoscience and remote sensing symposium (IGARSS), pp. 1445–1448.
50.
VincentL (1993) Morphological grayscale reconstruction in image analysis: Applications and efficient algorithms. IEEE Transactions on Image Processing2: 176–201.
51.
WanZDozierJ (1996) A generalized split-window algorithm for retrieving land-surface temperature from space. IEEE Transactions on Geoscience and Remote Sensing34(4): 892–905. DOI:10.1109/36.508406.
52.
WilliamsSWatermanAPattersonD (2009) Roofline: an insightful visual performance model for multicore architectures. Communications of the ACM52(4): 65–76.
