Talk by Richard (Rich) Vuduc of Georgia Tech, presented at the Workshop on Visualization and Analysis of Performance of Large-scale Software, which was held on October 14.
- Talk slides: PDF (~ 6.3 MiB)
- Relevant tutorial (GPU-specific): MICRO'10 tutorial.
- Relevant paper 1, which is an example of scaling up: Rahimian et al., “Petascale direct numerical simulation of blood flow on 200k cores and heterogeneous architectures.” PDF preprint.
- Relevant paper 2, which is an example of diagnosis: Chandramowlishwaran et al., “Diagnosis, tuning, and redesign for multicore performance: A case study of the fast multipole method.;rdquo; PDF preprint.
- Relevant paper 3: Choi and Vuduc, “A roofline model of energy.” IPDPS'13 preprint | “Extended remix” (tech report).
- Relevant paper 4: Czechowski and Vuduc, “A theoretical framework for algorithm-architecture co-design.” IPDPS'13 preprint.