Auto-Tuning Complex Array Layouts for GPUs

dc.contributor.authorWeber, Nicolasen_US
dc.contributor.authorGoesele, Michaelen_US
dc.contributor.editorMargarita Amor and Markus Hadwigeren_US
dc.date.accessioned2014-12-16T07:31:19Z
dc.date.available2014-12-16T07:31:19Z
dc.date.issued2014en_US
dc.description.abstractThe continuing evolution of Graphics Processing Units (GPU) has shown rapid performance increases over the years. But with each new hardware generation, the constraints for programming them efficiently have changed. Programs have to be tuned towards one specific hardware to unleash the full potential. This is time consuming and costly as vendors tend to release a new generation every 18 months. It is therefore important to auto-tune GPU code to achieve GPU-specific improvements. Using either static or empirical profiling to adjust parameters or to change the kernel implementation. We introduce a new approach to automatically improve memory access on GPUs. Our system generates an application specific library which abstracts the memory access for complex arrays on the host and GPU side. This allows to optimize the code by exchanging the memory layout without recompiling the application, as all necessary layouts are pre-compiled into the library. Our implementation is able to speedup real-world applications up to an order of magnitude and even outperforms hand-tuned implementations.en_US
dc.description.seriesinformationEurographics Symposium on Parallel Graphics and Visualizationen_US
dc.identifier.isbn978-3-905674-59-0en_US
dc.identifier.issn1727-348Xen_US
dc.identifier.urihttps://doi.org/10.2312/pgv.20141085en_US
dc.identifier.urihttps://diglib.eg.org/handle/10.2312/pgv.20141085.057-064
dc.publisherThe Eurographics Associationen_US
dc.subjectD.3.3 [Programming Technique]en_US
dc.subjectLanguage Constructs and Featuresen_US
dc.subjectData types and structuresen_US
dc.subjectI.3.1 [Computer Graphics]en_US
dc.subjectHardware Architectureen_US
dc.subjectGraphics processorsen_US
dc.subjectI.3.6 [Computer Graphics]en_US
dc.subjectMethodology and Techniquesen_US
dc.subjectGraphics data structures and data typesen_US
dc.titleAuto-Tuning Complex Array Layouts for GPUsen_US
Files
Original bundle
Now showing 1 - 2 of 2
Loading...
Thumbnail Image
Name:
057-064.pdf
Size:
546.93 KB
Format:
Adobe Portable Document Format
Loading...
Thumbnail Image
Name:
paper1005_mm.pdf
Size:
1.18 MB
Format:
Adobe Portable Document Format