TY - GEN
T1 - Leveraging GPU Tensor Cores for Double Precision Euclidean Distance Calculations
AU - Gallet, Benoit
AU - Gowanlock, Michael
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Tensor cores (TCs) are a type of Application-Specific Integrated Circuit (ASIC) and are a recent addition to Graphics Processing Unit (GPU) architectures. As such, TCs are purposefully designed to greatly improve the performance of Matrix Multiply-Accumulate (MMA) operations. While TCs are heavily studied for machine learning and closely related fields, where their high efficiency is undeniable, MMA operations are not unique to these fields. More generally, any computation that can be expressed as MMA operations can leverage TCs, and potentially benefit from their higher computational throughput compared to other general-purpose cores, such as CUDA cores on Nvidia GPUs. In this paper, we propose the first double precision (FP64) Euclidean distance calculation algorithm, which is expressed as MMA operations to leverage TCs on Nvidia GPUs, rather than the more commonly used CUDA cores. To show that the Euclidean distance can be accelerated in a real-world application, we evaluate our proposed TC algorithm on the distance similarity self-join problem, as the most computationally intensive part of the algorithm consists of computing distances in a multi-dimensional space. We find that the performance gain from using the tensor core algorithm over the CUDA core algorithm depends weakly on the dataset size and distribution, but is strongly dependent on data dimensionality. Overall, TCs are a compelling alternative to CUDA cores, particularly when the data dimensionality is low (≤ 4), as we achieve an average speedup of 1.28× and up to 2.23× against a state-of-the-art GPU distance similarity self-join algorithm. Furthermore, because this paper is among the first to explore the use of TCs for FP64 general-purpose computation, future research is promising.
AB - Tensor cores (TCs) are a type of Application-Specific Integrated Circuit (ASIC) and are a recent addition to Graphics Processing Unit (GPU) architectures. As such, TCs are purposefully designed to greatly improve the performance of Matrix Multiply-Accumulate (MMA) operations. While TCs are heavily studied for machine learning and closely related fields, where their high efficiency is undeniable, MMA operations are not unique to these fields. More generally, any computation that can be expressed as MMA operations can leverage TCs, and potentially benefit from their higher computational throughput compared to other general-purpose cores, such as CUDA cores on Nvidia GPUs. In this paper, we propose the first double precision (FP64) Euclidean distance calculation algorithm, which is expressed as MMA operations to leverage TCs on Nvidia GPUs, rather than the more commonly used CUDA cores. To show that the Euclidean distance can be accelerated in a real-world application, we evaluate our proposed TC algorithm on the distance similarity self-join problem, as the most computationally intensive part of the algorithm consists of computing distances in a multi-dimensional space. We find that the performance gain from using the tensor core algorithm over the CUDA core algorithm depends weakly on the dataset size and distribution, but is strongly dependent on data dimensionality. Overall, TCs are a compelling alternative to CUDA cores, particularly when the data dimensionality is low (≤ 4), as we achieve an average speedup of 1.28× and up to 2.23× against a state-of-the-art GPU distance similarity self-join algorithm. Furthermore, because this paper is among the first to explore the use of TCs for FP64 general-purpose computation, future research is promising.
KW - Euclidean Distance
KW - GPU
KW - Similarity Searches
KW - Tensor Cores
UR - http://www.scopus.com/inward/record.url?scp=85158164826&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85158164826&partnerID=8YFLogxK
U2 - 10.1109/HiPC56025.2022.00029
DO - 10.1109/HiPC56025.2022.00029
M3 - Conference contribution
AN - SCOPUS:85158164826
T3 - Proceedings - 2022 IEEE 29th International Conference on High Performance Computing, Data, and Analytics, HiPC 2022
SP - 135
EP - 144
BT - Proceedings - 2022 IEEE 29th International Conference on High Performance Computing, Data, and Analytics, HiPC 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 29th Annual IEEE International Conference on High Performance Computing, Data, and Analytics, HiPC 2022
Y2 - 18 December 2022 through 21 December 2022
ER -