Purpose: The purpose of this study was to investigate the dose-effect relation between acute esophageal toxicity (AET) and the dose-volume parameters of the esophagus after intensity modulated radiation therapy (IMRT) and concurrent chemotherapy for patients with non-small cell lung cancer (NSCLC). Patients and Methods: One hundred thirty-nine patients with inoperable NSCLC treated with IMRT and concurrent chemotherapy were prospectively analyzed. The fractionation scheme was 66 Gy in 24 fractions. All patients received concurrently a daily dose of cisplatin (6 mg/m Superscript-Two ). Maximum AET was scored according to Common Toxicity Criteria 3.0. Dose-volume parameters V5 to V70, D{sub mean} and D{sub max} of the esophagus were calculated. A logistic regression analysis was performed to analyze the dose-effect relation between these parameters and grade {>=}2 and grade {>=}3 AET. The outcome was compared with the clinically used esophagus V35 prediction model for grade {>=}2 after radical 3-dimensional conformal radiation therapy (3DCRT) treatment. Results: In our patient group, 9% did not experience AET, and 31% experienced grade 1 AET, 38% grade 2 AET, and 22% grade 3 AET. The incidence of grade 2 and grade 3 AET was not different from that in patients treated with CCRT using 3DCRT. The V50 turned out to be the most significant dosimetric predictor for grade {>=}3 AET (P=.012). The derived V50 model was shown to predict grade {>=}2 AET significantly better than the clinical V35 model (P<.001). Conclusions: For NSCLC patients treated with IMRT and concurrent chemotherapy, the V50 was identified as most accurate predictor of grade {>=}3 AET. There was no difference in the incidence of grade {>=}2 AET between 3DCRT and IMRT in patients treated with concurrent chemoradiation therapy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2012

Improving High-Performance Sparse Libraries Using Compiler-Assisted Specialization: A PETSc Case Study

Computers with hardware accelerators, also referred to as hybrid-core systems, speedup applications by offloading certain compute operations that can run faster on accelerators. Thus, it is not surprising that many of top500 supercomputers use accelerators. However, in addition to procurement cost, significant programming and porting effort is required to realize the potential benefit of such accelerators. Hence, before building such a system it is prudent to answer the question 'what is the projected performance benefit from accelerators for the workloads of interest?'. We address this question by way of a performance-modeling framework that predicts realizable application performance on accelerators rapidly and accurately without going to the considerable effort of porting and tuning. The modeling framework first automatically identifies commonly found compute patterns in scientific applications which we term idioms, which may benefit by accelerator technology. Next the framework models the predicted speedup of those idioms if they were to be ported to and run on hardware accelerators. As a proof of concept we characterize two kinds of accelerators 1) the FPGA accelerators on a Convey HC-1 system and 2) an NVIDIA FERMI GPU accelerator. We model performance of the idioms gather/scatter and stream and our predictions show that where these occur in two full-scale HPC applications, Milc and HYCOM, gather/scatter speeds up by as much as 15X, and stream by as much as 14X, whereas the overall compute time of Milc improves by 3.4% and HYCOM by 20%.

Computer Science↗