Separable 2D Convolution with Polymorphic Register Files

Ciobanu, Catalin Bogdan; Gaydadjiev, Georgi

doi:10.1007/978-3-642-36424-2_27

Cited by 2 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A vectorized separable 2D convolution algorithm, which avoids strided memory accesses when accessing column-wise input data, is introduced in [12]. The input matrix contains MAX_V × MAX_H elements.…”

Section: Vectorized Separable 2d Convolutionmentioning

confidence: 99%

See 1 more Smart Citation

The Case for Polymorphic Registers in Dataflow Computing

Ciobanu

Gaydadjiev

Pilato

et al. 2017

Int J Parallel Prog

View full text Add to dashboard Cite

Heterogeneous systems are becoming increasingly popular, delivering high performance through hardware specialization. However, sequential data accesses may have a negative impact on performance. Data parallel solutions such as Polymorphic Register Files (PRFs) can potentially accelerate applications by facilitating high-speed, parallel access to performance-critical data. This article shows how PRFs can be integrated into dataflow computational platforms. Our semi-automatic, compiler-based methodology generates customized PRFs and modifies the computational kernels to efficiently exploit them. We use a separable 2D convolution case study to evaluate the impact of memory latency and bandwidth on performance compared to a state-of-theart NVIDIA Tesla C2050 GPU. We improve the throughput up to 56.17X and show that the PRF-augmented system outperforms the GPU for 9 × 9 or larger mask sizes, even in bandwidth-constrained systems.

show abstract

Section: Vectorized Separable 2d Convolutionmentioning

confidence: 99%

“…Furthermore, potential power and area savings were shown by employing fewer PRF cores compared to a system with Cell processors. A preliminary evaluation of the PRF 2D separable convolution performance showed that PRFs can outperform state-of-the-art GPUs in terms of throughput [12].…”

Section: Introductionmentioning

confidence: 99%

The Case for Polymorphic Registers in Dataflow Computing

Ciobanu

Gaydadjiev

Pilato

et al. 2017

Int J Parallel Prog

View full text Add to dashboard Cite

show abstract

“…A vectorized separable 2D convolution algorithm which avoids strided memory accesses when accessing column-wise input data is introduced in [11]. The input matrix contains MAX V × MAX H elements.…”

Section: Case Study and Analytical Validationmentioning

confidence: 99%

“…Furthermore, potential power and area savings were shown by employing fewer PRF cores compared to a Cell processors system. A preliminary evaluation of the PRF 2D separable convolution performance [11] showed that PRFs outperform state-of-the art GPUs in throughput for mask sizes of 9 × 9 elements of larger when bandwidth is constrained.…”

Section: Introductionmentioning

confidence: 99%

Dataflow computing with Polymorphic Registers

Ciobanu

Gaydadjiev

Pilato

et al. 2013

2013 International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation (SAMOS)

View full text Add to dashboard Cite

Heterogeneous systems are becoming increasingly popular for data processing. They improve performance of simple kernels applied to large amounts of data. However, sequential data loads may have negative impact. Data parallel solutions such as Polymorphic Register Files (PRFs) can potentially accelerate applications by facilitating high speed, parallel access to performance-critical data. Furthermore, by PRF customization, specific data path features are exposed to the programmer in a very convenient way. PRFs allow additional control over the registers dimensions, and the number of elements which can be simultaneously accessed by computational units. This paper shows how PRFs can be integrated in dataflow computational platforms. In particular, starting from an annotated source code, we present a compiler-based methodology that automatically generates the customized PRFs and the enhanced computational kernels that efficiently exploit them.

show abstract

Separable 2D Convolution with Polymorphic Register Files

Cited by 2 publications

References 16 publications

The Case for Polymorphic Registers in Dataflow Computing

The Case for Polymorphic Registers in Dataflow Computing

Dataflow computing with Polymorphic Registers

Contact Info

Product

Resources

About