Zobrazeno 1 - 10
of 15
pro vyhledávání: '"Artjoms Šinkarovs"'
Autor:
Artjoms Šinkarovs, Sven-Bodo Scholz
Publikováno v:
Henriksen, T. (ed.), ARRAY '22: 8th ACM SIGPLAN International Workshop on Libraries, Languages and Compilers for Array Programming San Diego CA USA 13 June 2022, pp. 1–11
Contains fulltext : 253558.pdf (Publisher’s version ) (Open Access) ARRAY '22
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::ed211ef09c697f9b264b972dce333c80
https://repository.ubn.ru.nl/handle/2066/253558
https://repository.ubn.ru.nl/handle/2066/253558
Publikováno v:
ARRAY@PLDI
Low, T. Meng (ed.), ARRAY 2021: Proceedings of the 7th ACM SIGPLAN International Workshop on Libraries, Languages and Compilers for Array Programming, Virtual Canada 21 June 2021, pp. 39–50
Low, T. Meng (ed.), ARRAY 2021: Proceedings of the 7th ACM SIGPLAN International Workshop on Libraries, Languages and Compilers for Array Programming, Virtual Canada 21 June 2021, pp. 39–50
Modern machine learning frameworks are complex: they are typically organised in multiple layers each of which is written in a different language and they depend on a number of external libraries, but at their core they mainly consist of tensor operat
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::5ed45cbb6ef44225ec1f3abd36cb98f2
https://doi.org/10.1145/3460944.3464312
https://doi.org/10.1145/3460944.3464312
Autor:
Artjoms Šinkarovs
Publikováno v:
MSFP@ETAPS
We explore a data structure that generalises rectangular multi-dimensional arrays. The shape of an n-dimensional array is typically given by a tuple of n natural numbers. Each element in that tuple defines the length of the corresponding axis. If we
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::b50a5f588d585339324f65da20840271
http://arxiv.org/abs/2005.00198
http://arxiv.org/abs/2005.00198
Publikováno v:
Euro-Par 2019: Parallel Processing Workshops ISBN: 9783030483395
Euro-Par Workshops
Euro-Par Workshops
This paper proposes a new approach to checkpointing MPI applications that use long-running CUDA kernels. It becomes possible to take snapshots of data residing on the GPUs without waiting for kernels to complete. The proposed technique is implemented
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::1e356acf9d366fd1434afd1d89f00105
https://doi.org/10.1007/978-3-030-48340-1_53
https://doi.org/10.1007/978-3-030-48340-1_53
Autor:
Artjoms Šinkarovs, Sven-Bodo Scholz
Publikováno v:
IFL
We propose a new notation for data parallel operators on multi-dimensional arrays named tensor comprehensions. This notation combines the basic principle of array-comprehensions with syntactical shortcuts very close to those found in the so-called Te
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::95db93df6325ad463b4e8e22ab920802
https://doi.org/10.1145/3412932.3412947
https://doi.org/10.1145/3412932.3412947
Publikováno v:
ARRAY@PLDI
This paper shows how a Convolutional Neural Network (CNN) can be implemented in APL. Its first-class array support ideally fits that domain, and the operations of APL facilitate rapid and concise creation of generically reusable building blocks. For
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::efeb52fd7d47817c2ef971ce324de717
https://doi.org/10.1145/3315454.3329960
https://doi.org/10.1145/3315454.3329960
Publikováno v:
Lecture Notes in Computer Science ISBN: 9783030105488
Euro-Par Workshops
Euro-Par Workshops
Resilience for HPC applications typically is implemented as a CPU-based rollback-recovery technique. In this context, long running accelerator computations on GPUs pose a major challenge as these devices usually do not offer any means of interrupt. T
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::f708618ecba67d20ce5cb1f9d835d9b5
https://doi.org/10.1007/978-3-030-10549-5_64
https://doi.org/10.1007/978-3-030-10549-5_64
Publikováno v:
IFL
In this paper we present an optimisation for reference counting based garbage collection. The optimisation aims at reducing the total number of calls to the heap manager while preserving the key benefits of reference counting, i.e. the opportunities
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::cfc8839c7a3b2c77b6d594c613216898
https://doi.org/10.1145/3310232.3310242
https://doi.org/10.1145/3310232.3310242
Publikováno v:
ARRAY@PLDI
This paper aims to foster cross-fertilisation between programming language and compiler research performed on different array programming language infrastructures. We study how to enable better comparability of concepts and techniques by looking into
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::f011c15a459e745659f0eff728c278e2
https://doi.org/10.1145/3219753.3219754
https://doi.org/10.1145/3219753.3219754
Autor:
Sven-Bodo Scholz, Artjoms Šinkarovs
Publikováno v:
Concurrency and Computation: Practice and Experience. 28:2092-2119
Vector instructions of modern CPUs are crucially important for the performance of compute-intensive algorithms. Auto-vectorisation often fails because of an unfortunate choice of data layout by the programmer. This paper proposes a data layout infere
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::be9edde2fbe813b435096797b8bd2850
https://doi.org/10.1002/cpe.3501
https://doi.org/10.1002/cpe.3501