Porting a neuro-imaging application to a CPU-GPU cluster
Bibliographic record
Abstract
The ever increasing complexity of scientific applications has led to utilization of new HPC paradigms such as Graphical Processing Units (GPUs). However, modifying existing applications to enable them to be executed on GPU can be challenging. Furthermore, the considerable speedup achieved by execution of linear algebra operations on GPUs has added a huge heterogeneity to HPC clusters. In this work, we enabled NPAIRS, a neuro-imaging application, to be executed on GPU with slight modifications to its original code. This important feature of our implementation enables current users of NPAIRS, i.e. non-expert bio-medical scientists, to get benefit from GPU without having to apply fundamental changes to their existing application. As the second part of our research, we investigated the efficiency of several scheduling algorithms for a heterogeneous cluster that contains GPU nodes. Experimental results show that we achieved 7× speedup for NPAIRS. Moreover, although scheduling does not play an important role when there is no GPU node in the cluster, it can highly improve the makespan for a CPU-GPU cluster. We compared our scheduling results with Torque and MCT, two of the most commonly used schedulers in current HPC platforms. Our results show that the Sufferage scheduling can improve the makespan of Torque and MCT by 47% and 4% respectively.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".