2013 Fiscal Year Final Research Report
Development of Auto-Tuning framework using pragmas for optimization of GPU program
| Project/Area Number |
24700024
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Software
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2012-04-01 – 2014-03-31
|
| Keywords | GPGPU / 自動チューニング |
|---|
| Research Abstract |
This project researched to reduce the difficulty and labor of optimization of GPU program. Especially, in order to make the product useful for users, this project attached a high value to achieve the purpose by using simple pragmas and API functions.
1. In order to make performance evaluation and utilization of the result of the performance evaluation, this project proposed and developed a framework to utilize some information of the profiler.
2. This project studied and developed a framework which confirm the computation of GPU and reduce the cost of testing by focusing on the data transfer between CPU and GPU.
|
