際際滷shows by User: jasonriedy / http://www.slideshare.net/images/logo.gif 際際滷shows by User: jasonriedy / Thu, 14 Oct 2021 00:52:50 GMT 際際滷Share feed for 際際滷shows by User: jasonriedy Lucata at the HPEC GraphBLAS BoF /slideshow/lucata-at-the-hpec-graphblas-bof-250439305/250439305 2021-graphblas-bof-lucata-211014005251
A five minute piece to introduce the Lucata architecture to the GraphBLAS BoF.]]>
A five minute piece to introduce the Lucata architecture to the GraphBLAS BoF.]]> Thu, 14 Oct 2021 00:52:50 GMT /slideshow/lucata-at-the-hpec-graphblas-bof-250439305/250439305 jasonriedy@slideshare.net(jasonriedy) Lucata at the HPEC GraphBLAS BoF jasonriedy A five minute piece to introduce the Lucata architecture to the GraphBLAS BoF. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/2021-graphblas-bof-lucata-211014005251-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> A five minute piece to introduce the Lucata architecture to the GraphBLAS BoF.
Lucata at the HPEC GraphBLAS BoF from Jason Riedy
]]> 88 0 https://cdn.slidesharecdn.com/ss_thumbnails/2021-graphblas-bof-lucata-211014005251-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 LAGraph 2021-10-13 /slideshow/lagraph-20211013/250439292 lagraph-2021-10-13-211014005046
Introducing folks to the Lucata platform for high-performance, massive-scale graph analysis.]]>
Introducing folks to the Lucata platform for high-performance, massive-scale graph analysis.]]> Thu, 14 Oct 2021 00:50:45 GMT /slideshow/lagraph-20211013/250439292 jasonriedy@slideshare.net(jasonriedy) LAGraph 2021-10-13 jasonriedy Introducing folks to the Lucata platform for high-performance, massive-scale graph analysis. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/lagraph-2021-10-13-211014005046-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Introducing folks to the Lucata platform for high-performance, massive-scale graph analysis.
LAGraph 2021-10-13 from Jason Riedy
]]> 123 0 https://cdn.slidesharecdn.com/ss_thumbnails/lagraph-2021-10-13-211014005046-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Lucata at the HPEC GraphBLAS BoF /slideshow/lucata-at-the-hpec-graphblas-bof/250439277 2021-graphblas-bof-lucata-211014004830
A five minute set from the HPEC 2021 GraphBLAS BoF.]]>
A five minute set from the HPEC 2021 GraphBLAS BoF.]]> Thu, 14 Oct 2021 00:48:29 GMT /slideshow/lucata-at-the-hpec-graphblas-bof/250439277 jasonriedy@slideshare.net(jasonriedy) Lucata at the HPEC GraphBLAS BoF jasonriedy A five minute set from the HPEC 2021 GraphBLAS BoF. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/2021-graphblas-bof-lucata-211014004830-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> A five minute set from the HPEC 2021 GraphBLAS BoF.
Lucata at the HPEC GraphBLAS BoF from Jason Riedy
]]> 90 0 https://cdn.slidesharecdn.com/ss_thumbnails/2021-graphblas-bof-lucata-211014004830-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Graph analysis and novel architectures /slideshow/graph-analysis-and-novel-architectures/239447058 graphanalysisandnovelarchitectures3-201124223536
Presented at CERFACS Sparse Days, 24 Nov 2020]]>
Presented at CERFACS Sparse Days, 24 Nov 2020]]> Tue, 24 Nov 2020 22:35:36 GMT /slideshow/graph-analysis-and-novel-architectures/239447058 jasonriedy@slideshare.net(jasonriedy) Graph analysis and novel architectures jasonriedy Presented at CERFACS Sparse Days, 24 Nov 2020 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/graphanalysisandnovelarchitectures3-201124223536-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Presented at CERFACS Sparse Days, 24 Nov 2020
Graph analysis and novel architectures from Jason Riedy
]]> 102 0 https://cdn.slidesharecdn.com/ss_thumbnails/graphanalysisandnovelarchitectures3-201124223536-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 GraphBLAS and Emus /slideshow/graphblas-and-emus/239447040 main-201124223352
Given at IEEE HPEC 2020 GraphBLAS BoF]]>
Given at IEEE HPEC 2020 GraphBLAS BoF]]> Tue, 24 Nov 2020 22:33:52 GMT /slideshow/graphblas-and-emus/239447040 jasonriedy@slideshare.net(jasonriedy) GraphBLAS and Emus jasonriedy Given at IEEE HPEC 2020 GraphBLAS BoF <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/main-201124223352-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Given at IEEE HPEC 2020 GraphBLAS BoF
GraphBLAS and Emus from Jason Riedy
]]> 93 0 https://cdn.slidesharecdn.com/ss_thumbnails/main-201124223352-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Reproducible Linear Algebra from Application to Architecture /slideshow/reproducible-linear-algebra-from-application-to-architecture-228263588/228263588 siampp2020-out-200216201834
All computing must be parallel to take advantage of modern systems like multicore processors, GPUs, and distributed systems. Results that are not bit-wise reproducible introduce doubt on many levels. Sometimes that is appropriate. Reproducibility limitations occur because underlying libraries do not specify their reproducibility requirements. New advances in interfaces, algorithms, and architectures allow selecting among those requirements in the future. This talk covers many of the upcoming options and their trade-offs.]]>
All computing must be parallel to take advantage of modern systems like multicore processors, GPUs, and distributed systems. Results that are not bit-wise reproducible introduce doubt on many levels. Sometimes that is appropriate. Reproducibility limitations occur because underlying libraries do not specify their reproducibility requirements. New advances in interfaces, algorithms, and architectures allow selecting among those requirements in the future. This talk covers many of the upcoming options and their trade-offs.]]> Sun, 16 Feb 2020 20:18:34 GMT /slideshow/reproducible-linear-algebra-from-application-to-architecture-228263588/228263588 jasonriedy@slideshare.net(jasonriedy) Reproducible Linear Algebra from Application to Architecture jasonriedy All computing must be parallel to take advantage of modern systems like multicore processors, GPUs, and distributed systems. Results that are not bit-wise reproducible introduce doubt on many levels. Sometimes that is appropriate. Reproducibility limitations occur because underlying libraries do not specify their reproducibility requirements. New advances in interfaces, algorithms, and architectures allow selecting among those requirements in the future. This talk covers many of the upcoming options and their trade-offs. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/siampp2020-out-200216201834-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> All computing must be parallel to take advantage of modern systems like multicore processors, GPUs, and distributed systems. Results that are not bit-wise reproducible introduce doubt on many levels. Sometimes that is appropriate. Reproducibility limitations occur because underlying libraries do not specify their reproducibility requirements. New advances in interfaces, algorithms, and architectures allow selecting among those requirements in the future. This talk covers many of the upcoming options and their trade-offs.
Reproducible Linear Algebra from Application to Architecture from Jason Riedy
]]> 179 0 https://cdn.slidesharecdn.com/ss_thumbnails/siampp2020-out-200216201834-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 PEARC19: Wrangling Rogues: A Case Study on Managing Experimental Post-Moore Architectures /slideshow/wrangling-rogues-a-case-study-on-managing-experimental-postmoore-architectures/159845556 rogues-gallery-pearc19-190801164004
The Rogues Gallery is a new experimental testbed that is focused on tackling "rogue'' architectures for the Post-Moore era of computing. While some of these devices have roots in the embedded and high-performance computing spaces, managing current and emerging technologies provides a challenge for system administration that are not always foreseen in traditional data center environments. We present an overview of the motivations and design of the initial Rogues Gallery testbed and cover some of the unique challenges that we have seen and foresee with upcoming hardware prototypes for future post-Moore research. Specifically, we cover the networking, identity management, scheduling of resources, and tools and sensor access aspects of the Rogues Gallery and techniques we have developed to manage these new platforms. We argue that current tools like the Slurm resource manager can support new rogues without major infrastructure changes.]]>
The Rogues Gallery is a new experimental testbed that is focused on tackling "rogue'' architectures for the Post-Moore era of computing. While some of these devices have roots in the embedded and high-performance computing spaces, managing current and emerging technologies provides a challenge for system administration that are not always foreseen in traditional data center environments. We present an overview of the motivations and design of the initial Rogues Gallery testbed and cover some of the unique challenges that we have seen and foresee with upcoming hardware prototypes for future post-Moore research. Specifically, we cover the networking, identity management, scheduling of resources, and tools and sensor access aspects of the Rogues Gallery and techniques we have developed to manage these new platforms. We argue that current tools like the Slurm resource manager can support new rogues without major infrastructure changes.]]> Thu, 01 Aug 2019 16:40:04 GMT /slideshow/wrangling-rogues-a-case-study-on-managing-experimental-postmoore-architectures/159845556 jasonriedy@slideshare.net(jasonriedy) PEARC19: Wrangling Rogues: A Case Study on Managing Experimental Post-Moore Architectures jasonriedy The Rogues Gallery is a new experimental testbed that is focused on tackling "rogue'' architectures for the Post-Moore era of computing. While some of these devices have roots in the embedded and high-performance computing spaces, managing current and emerging technologies provides a challenge for system administration that are not always foreseen in traditional data center environments. We present an overview of the motivations and design of the initial Rogues Gallery testbed and cover some of the unique challenges that we have seen and foresee with upcoming hardware prototypes for future post-Moore research. Specifically, we cover the networking, identity management, scheduling of resources, and tools and sensor access aspects of the Rogues Gallery and techniques we have developed to manage these new platforms. We argue that current tools like the Slurm resource manager can support new rogues without major infrastructure changes. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/rogues-gallery-pearc19-190801164004-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> The Rogues Gallery is a new experimental testbed that is focused on tackling &quot;rogue&#39;&#39; architectures for the Post-Moore era of computing. While some of these devices have roots in the embedded and high-performance computing spaces, managing current and emerging technologies provides a challenge for system administration that are not always foreseen in traditional data center environments. We present an overview of the motivations and design of the initial Rogues Gallery testbed and cover some of the unique challenges that we have seen and foresee with upcoming hardware prototypes for future post-Moore research. Specifically, we cover the networking, identity management, scheduling of resources, and tools and sensor access aspects of the Rogues Gallery and techniques we have developed to manage these new platforms. We argue that current tools like the Slurm resource manager can support new rogues without major infrastructure changes.
PEARC19: Wrangling Rogues: A Case Study on Managing Experimental Post-Moore Architectures from Jason Riedy
]]> 121 7 https://cdn.slidesharecdn.com/ss_thumbnails/rogues-gallery-pearc19-190801164004-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 ICIAM 2019: Reproducible Linear Algebra from Application to Architecture /slideshow/reproducible-linear-algebra-from-application-to-architecture/156809508 iciam-2019-reprola-out-190721203709
All computing must be parallel to take advantage of modern systems like multicore processors, GPUs, and distributed systems. Results that are not bit-wise reproducible introduce doubt on many levels. Sometimes that is appropriate. Reproducibility limitations occur because underlying libraries do not specify their reproducibility requirements. New advances in interfaces, algorithms, and architectures allow selecting among those requirements in the future. This talk covers many of the upcoming options and their trade-offs.]]>
All computing must be parallel to take advantage of modern systems like multicore processors, GPUs, and distributed systems. Results that are not bit-wise reproducible introduce doubt on many levels. Sometimes that is appropriate. Reproducibility limitations occur because underlying libraries do not specify their reproducibility requirements. New advances in interfaces, algorithms, and architectures allow selecting among those requirements in the future. This talk covers many of the upcoming options and their trade-offs.]]> Sun, 21 Jul 2019 20:37:09 GMT /slideshow/reproducible-linear-algebra-from-application-to-architecture/156809508 jasonriedy@slideshare.net(jasonriedy) ICIAM 2019: Reproducible Linear Algebra from Application to Architecture jasonriedy All computing must be parallel to take advantage of modern systems like multicore processors, GPUs, and distributed systems. Results that are not bit-wise reproducible introduce doubt on many levels. Sometimes that is appropriate. Reproducibility limitations occur because underlying libraries do not specify their reproducibility requirements. New advances in interfaces, algorithms, and architectures allow selecting among those requirements in the future. This talk covers many of the upcoming options and their trade-offs. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/iciam-2019-reprola-out-190721203709-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> All computing must be parallel to take advantage of modern systems like multicore processors, GPUs, and distributed systems. Results that are not bit-wise reproducible introduce doubt on many levels. Sometimes that is appropriate. Reproducibility limitations occur because underlying libraries do not specify their reproducibility requirements. New advances in interfaces, algorithms, and architectures allow selecting among those requirements in the future. This talk covers many of the upcoming options and their trade-offs.
ICIAM 2019: Reproducible Linear Algebra from Application to Architecture from Jason Riedy
]]> 247 9 https://cdn.slidesharecdn.com/ss_thumbnails/iciam-2019-reprola-out-190721203709-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 ICIAM 2019: A New Algorithm Model for Massive-Scale Streaming Graph Analysis /jasonriedy/a-new-algorithm-model-for-massivescale-streaming-graph-analysis-156808819 iciam-2019-graph-out-190721203329
Applications in many areas analyze an ever-changing environment. On billion vertices graphs, providing snapshots imposes a large performance cost. We propose the first formal model for graph analysis running concurrently with streaming data updates. We consider an algorithm valid if its output is correct for the initial graph plus some implicit subset of concurrent changes. We show theoretical properties of the model, demonstrate the model on various algorithms, and extend it to updating results incrementally.]]>
Applications in many areas analyze an ever-changing environment. On billion vertices graphs, providing snapshots imposes a large performance cost. We propose the first formal model for graph analysis running concurrently with streaming data updates. We consider an algorithm valid if its output is correct for the initial graph plus some implicit subset of concurrent changes. We show theoretical properties of the model, demonstrate the model on various algorithms, and extend it to updating results incrementally.]]> Sun, 21 Jul 2019 20:33:29 GMT /jasonriedy/a-new-algorithm-model-for-massivescale-streaming-graph-analysis-156808819 jasonriedy@slideshare.net(jasonriedy) ICIAM 2019: A New Algorithm Model for Massive-Scale Streaming Graph Analysis jasonriedy Applications in many areas analyze an ever-changing environment. On billion vertices graphs, providing snapshots imposes a large performance cost. We propose the first formal model for graph analysis running concurrently with streaming data updates. We consider an algorithm valid if its output is correct for the initial graph plus some implicit subset of concurrent changes. We show theoretical properties of the model, demonstrate the model on various algorithms, and extend it to updating results incrementally. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/iciam-2019-graph-out-190721203329-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Applications in many areas analyze an ever-changing environment. On billion vertices graphs, providing snapshots imposes a large performance cost. We propose the first formal model for graph analysis running concurrently with streaming data updates. We consider an algorithm valid if its output is correct for the initial graph plus some implicit subset of concurrent changes. We show theoretical properties of the model, demonstrate the model on various algorithms, and extend it to updating results incrementally.
ICIAM 2019: A New Algorithm Model for Massive-Scale Streaming Graph Analysis from Jason Riedy
]]> 222 12 https://cdn.slidesharecdn.com/ss_thumbnails/iciam-2019-graph-out-190721203329-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Novel Architectures for Applications in Data Science and Beyond /slideshow/novel-architectures-for-applications-in-data-science-and-beyond/133990610 rg-190301234648
Describing new HPDA applications and architectures with a focus on Georgia Tech's CRNCH Rogues Gallery]]>
Describing new HPDA applications and architectures with a focus on Georgia Tech's CRNCH Rogues Gallery]]> Fri, 01 Mar 2019 23:46:47 GMT /slideshow/novel-architectures-for-applications-in-data-science-and-beyond/133990610 jasonriedy@slideshare.net(jasonriedy) Novel Architectures for Applications in Data Science and Beyond jasonriedy Describing new HPDA applications and architectures with a focus on Georgia Tech's CRNCH Rogues Gallery <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/rg-190301234648-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Describing new HPDA applications and architectures with a focus on Georgia Tech&#39;s CRNCH Rogues Gallery
Novel Architectures for Applications in Data Science and Beyond from Jason Riedy
]]> 151 5 https://cdn.slidesharecdn.com/ss_thumbnails/rg-190301234648-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Characterization of Emu Chick with Microbenchmarks /jasonriedy/characterization-of-emu-chick-with-microbenchmarks gatech-riedy-190123154302
Presenting current and recent performance results on the Emu Chick looking towards graph and data analysis.]]>
Presenting current and recent performance results on the Emu Chick looking towards graph and data analysis.]]> Wed, 23 Jan 2019 15:43:01 GMT /jasonriedy/characterization-of-emu-chick-with-microbenchmarks jasonriedy@slideshare.net(jasonriedy) Characterization of Emu Chick with Microbenchmarks jasonriedy Presenting current and recent performance results on the Emu Chick looking towards graph and data analysis. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/gatech-riedy-190123154302-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Presenting current and recent performance results on the Emu Chick looking towards graph and data analysis.
Characterization of Emu Chick with Microbenchmarks from Jason Riedy
]]> 130 3 https://cdn.slidesharecdn.com/ss_thumbnails/gatech-riedy-190123154302-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 CRNCH 2018 Summit: Rogues Gallery Update /slideshow/crnch-2018-summit-rogues-gallery-update/121576405 rogues-gallery-pres-181102171156
In one classic sense a rogue is someone who goes their own way, who breaks away from the crowd. The CRNCH Rogues Gallery aims to support computer architecture rogues by being a physical and virtual space providing access to novel computing architectures. Researchers find applications, and architects discover what happens when their prototypes hit reality. Our goals are to help kick-start software ecosystems, train students in novel system evaluation and use, and provide rapid feedback to architects. By exposing students and researchers to this set of unique hardware, we foster cross-cutting discussions about hardware designs that will drive future performance improvements in computing long after the Moores Law era of cheap transistors ends. We provide a brief description of the current Rogues Gallery along with successes and research highlights over the last year. ]]>
In one classic sense a rogue is someone who goes their own way, who breaks away from the crowd. The CRNCH Rogues Gallery aims to support computer architecture rogues by being a physical and virtual space providing access to novel computing architectures. Researchers find applications, and architects discover what happens when their prototypes hit reality. Our goals are to help kick-start software ecosystems, train students in novel system evaluation and use, and provide rapid feedback to architects. By exposing students and researchers to this set of unique hardware, we foster cross-cutting discussions about hardware designs that will drive future performance improvements in computing long after the Moores Law era of cheap transistors ends. We provide a brief description of the current Rogues Gallery along with successes and research highlights over the last year. ]]> Fri, 02 Nov 2018 17:11:56 GMT /slideshow/crnch-2018-summit-rogues-gallery-update/121576405 jasonriedy@slideshare.net(jasonriedy) CRNCH 2018 Summit: Rogues Gallery Update jasonriedy In one classic sense a rogue is someone who goes their own way, who breaks away from the crowd. The CRNCH Rogues Gallery aims to support computer architecture rogues by being a physical and virtual space providing access to novel computing architectures. Researchers find applications, and architects discover what happens when their prototypes hit reality. Our goals are to help kick-start software ecosystems, train students in novel system evaluation and use, and provide rapid feedback to architects. By exposing students and researchers to this set of unique hardware, we foster cross-cutting discussions about hardware designs that will drive future performance improvements in computing long after the Moores Law era of cheap transistors ends. We provide a brief description of the current Rogues Gallery along with successes and research highlights over the last year. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/rogues-gallery-pres-181102171156-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> In one classic sense a rogue is someone who goes their own way, who breaks away from the crowd. The CRNCH Rogues Gallery aims to support computer architecture rogues by being a physical and virtual space providing access to novel computing architectures. Researchers find applications, and architects discover what happens when their prototypes hit reality. Our goals are to help kick-start software ecosystems, train students in novel system evaluation and use, and provide rapid feedback to architects. By exposing students and researchers to this set of unique hardware, we foster cross-cutting discussions about hardware designs that will drive future performance improvements in computing long after the Moores Law era of cheap transistors ends. We provide a brief description of the current Rogues Gallery along with successes and research highlights over the last year.
CRNCH 2018 Summit: Rogues Gallery Update from Jason Riedy
]]> 105 6 https://cdn.slidesharecdn.com/ss_thumbnails/rogues-gallery-pres-181102171156-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Augmented Arithmetic Operations Proposed for IEEE-754 2018 /slideshow/augmented-arithmetic-operations-proposed-for-ieee754-2018/103146048 augmentedoperation-ejr-180626152156
Algorithms for extending arithmetic precision through compensated summation or arithmetics like double-double rely on operations commonly called twoSum and twoProduct. The current draft of the IEEE 754 standard specifies these operations under the names augmentedAddition and augmentedMultiplication. These operations were included after three decades of experience because of a motivating new use: bitwise reproducible arithmetic. Standardizing the operations provides a hardware acceleration target that can provide at least a 33% speed improvements in reproducible dot product, placing reproducible dot product almost within a factor of two of common dot product. This paper provides history and motivation for standardizing these operations. We also define the operations, explain the rationale for all the specific choices, and provide parameterized test cases for new boundary behaviors. ]]>
Algorithms for extending arithmetic precision through compensated summation or arithmetics like double-double rely on operations commonly called twoSum and twoProduct. The current draft of the IEEE 754 standard specifies these operations under the names augmentedAddition and augmentedMultiplication. These operations were included after three decades of experience because of a motivating new use: bitwise reproducible arithmetic. Standardizing the operations provides a hardware acceleration target that can provide at least a 33% speed improvements in reproducible dot product, placing reproducible dot product almost within a factor of two of common dot product. This paper provides history and motivation for standardizing these operations. We also define the operations, explain the rationale for all the specific choices, and provide parameterized test cases for new boundary behaviors. ]]> Tue, 26 Jun 2018 15:21:56 GMT /slideshow/augmented-arithmetic-operations-proposed-for-ieee754-2018/103146048 jasonriedy@slideshare.net(jasonriedy) Augmented Arithmetic Operations Proposed for IEEE-754 2018 jasonriedy Algorithms for extending arithmetic precision through compensated summation or arithmetics like double-double rely on operations commonly called twoSum and twoProduct. The current draft of the IEEE 754 standard specifies these operations under the names augmentedAddition and augmentedMultiplication. These operations were included after three decades of experience because of a motivating new use: bitwise reproducible arithmetic. Standardizing the operations provides a hardware acceleration target that can provide at least a 33% speed improvements in reproducible dot product, placing reproducible dot product almost within a factor of two of common dot product. This paper provides history and motivation for standardizing these operations. We also define the operations, explain the rationale for all the specific choices, and provide parameterized test cases for new boundary behaviors. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/augmentedoperation-ejr-180626152156-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Algorithms for extending arithmetic precision through compensated summation or arithmetics like double-double rely on operations commonly called twoSum and twoProduct. The current draft of the IEEE 754 standard specifies these operations under the names augmentedAddition and augmentedMultiplication. These operations were included after three decades of experience because of a motivating new use: bitwise reproducible arithmetic. Standardizing the operations provides a hardware acceleration target that can provide at least a 33% speed improvements in reproducible dot product, placing reproducible dot product almost within a factor of two of common dot product. This paper provides history and motivation for standardizing these operations. We also define the operations, explain the rationale for all the specific choices, and provide parameterized test cases for new boundary behaviors.
Augmented Arithmetic Operations Proposed for IEEE-754 2018 from Jason Riedy
]]> 459 5 https://cdn.slidesharecdn.com/ss_thumbnails/augmentedoperation-ejr-180626152156-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Graph Analysis: New Algorithm Models, New Architectures /slideshow/graph-analysis-new-algorithm-models-new-architectures/96521603 computing-frontiers-ejr-out-180509140537
Invited talk at ACM Computing Frontiers 2018]]>
Invited talk at ACM Computing Frontiers 2018]]> Wed, 09 May 2018 14:05:37 GMT /slideshow/graph-analysis-new-algorithm-models-new-architectures/96521603 jasonriedy@slideshare.net(jasonriedy) Graph Analysis: New Algorithm Models, New Architectures jasonriedy Invited talk at ACM Computing Frontiers 2018 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/computing-frontiers-ejr-out-180509140537-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Invited talk at ACM Computing Frontiers 2018
Graph Analysis: New Algorithm Models, New Architectures from Jason Riedy
]]> 344 4 https://cdn.slidesharecdn.com/ss_thumbnails/computing-frontiers-ejr-out-180509140537-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 CRNCH Rogues Gallery: A Community Core for Novel Computing Platforms /slideshow/crnch-rogues-gallery-a-community-core-for-novel-computing-platforms-95720624/95720624 rogues-gallery-spp-180502161903
Version from the CSE Strategic Partner Program meeting]]>
Version from the CSE Strategic Partner Program meeting]]> Wed, 02 May 2018 16:19:03 GMT /slideshow/crnch-rogues-gallery-a-community-core-for-novel-computing-platforms-95720624/95720624 jasonriedy@slideshare.net(jasonriedy) CRNCH Rogues Gallery: A Community Core for Novel Computing Platforms jasonriedy Version from the CSE Strategic Partner Program meeting <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/rogues-gallery-spp-180502161903-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Version from the CSE Strategic Partner Program meeting
CRNCH Rogues Gallery: A Community Core for Novel Computing Platforms from Jason Riedy
]]> 89 3 https://cdn.slidesharecdn.com/ss_thumbnails/rogues-gallery-spp-180502161903-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 CRNCH Rogues Gallery: A Community Core for Novel Computing Platforms /slideshow/crnch-rogues-gallery-a-community-core-for-novel-computing-platforms/86117624 rogues-gallery-180113215823
The Rogues Gallery is a new concept focused on developing our understanding of next-generation hardware with a focus on unorthodox and uncommon technologies. This project, initiated by Georgia Tech's Center for Research into Novel Computing Hierarchies (CRNCH), will acquire new and unique hardware (ie, the aforementioned "rogues") from vendors, research labs, and startups and make this hardware available to students, faculty, and industry collaborators within a managed data center environment. By exposing students and researchers to this set of unique hardware, we hope to foster cross-cutting discussions about hardware designs that will drive future performance improvements in computing long after the Moore's Law era of "cheap transistors" ends. ]]>
The Rogues Gallery is a new concept focused on developing our understanding of next-generation hardware with a focus on unorthodox and uncommon technologies. This project, initiated by Georgia Tech's Center for Research into Novel Computing Hierarchies (CRNCH), will acquire new and unique hardware (ie, the aforementioned "rogues") from vendors, research labs, and startups and make this hardware available to students, faculty, and industry collaborators within a managed data center environment. By exposing students and researchers to this set of unique hardware, we hope to foster cross-cutting discussions about hardware designs that will drive future performance improvements in computing long after the Moore's Law era of "cheap transistors" ends. ]]> Sat, 13 Jan 2018 21:58:23 GMT /slideshow/crnch-rogues-gallery-a-community-core-for-novel-computing-platforms/86117624 jasonriedy@slideshare.net(jasonriedy) CRNCH Rogues Gallery: A Community Core for Novel Computing Platforms jasonriedy The Rogues Gallery is a new concept focused on developing our understanding of next-generation hardware with a focus on unorthodox and uncommon technologies. This project, initiated by Georgia Tech's Center for Research into Novel Computing Hierarchies (CRNCH), will acquire new and unique hardware (ie, the aforementioned "rogues") from vendors, research labs, and startups and make this hardware available to students, faculty, and industry collaborators within a managed data center environment. By exposing students and researchers to this set of unique hardware, we hope to foster cross-cutting discussions about hardware designs that will drive future performance improvements in computing long after the Moore's Law era of "cheap transistors" ends. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/rogues-gallery-180113215823-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> The Rogues Gallery is a new concept focused on developing our understanding of next-generation hardware with a focus on unorthodox and uncommon technologies. This project, initiated by Georgia Tech&#39;s Center for Research into Novel Computing Hierarchies (CRNCH), will acquire new and unique hardware (ie, the aforementioned &quot;rogues&quot;) from vendors, research labs, and startups and make this hardware available to students, faculty, and industry collaborators within a managed data center environment. By exposing students and researchers to this set of unique hardware, we hope to foster cross-cutting discussions about hardware designs that will drive future performance improvements in computing long after the Moore&#39;s Law era of &quot;cheap transistors&quot; ends.
CRNCH Rogues Gallery: A Community Core for Novel Computing Platforms from Jason Riedy
]]> 180 5 https://cdn.slidesharecdn.com/ss_thumbnails/rogues-gallery-180113215823-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 A New Algorithm Model for Massive-Scale Streaming Graph Analysis /slideshow/a-new-algorithm-model-for-massivescale-streaming-graph-analysis/77886238 siam-ns-2017-ejr-170714190645
Applications in computer network security, social media analysis,and other areas rely on analyzing a changing environment. The data is rich in relationships and lends itself to graph analysis. Traditional static graph analysis cannot keep pace with network security applications analyzing nearly one million events per second and social networks like Facebook collecting 500 thousand comments per second. Streaming frameworks like STINGER support ingesting up three million of edge changes per second but there are few streaming analysis kernels that keep up with these rates. Here we present a new algorithm model for applying complex metrics to a changing graph. In this model, many more algorithms can be applied without having to stop the world.]]>
Applications in computer network security, social media analysis,and other areas rely on analyzing a changing environment. The data is rich in relationships and lends itself to graph analysis. Traditional static graph analysis cannot keep pace with network security applications analyzing nearly one million events per second and social networks like Facebook collecting 500 thousand comments per second. Streaming frameworks like STINGER support ingesting up three million of edge changes per second but there are few streaming analysis kernels that keep up with these rates. Here we present a new algorithm model for applying complex metrics to a changing graph. In this model, many more algorithms can be applied without having to stop the world.]]> Fri, 14 Jul 2017 19:06:45 GMT /slideshow/a-new-algorithm-model-for-massivescale-streaming-graph-analysis/77886238 jasonriedy@slideshare.net(jasonriedy) A New Algorithm Model for Massive-Scale Streaming Graph Analysis jasonriedy Applications in computer network security, social media analysis,and other areas rely on analyzing a changing environment. The data is rich in relationships and lends itself to graph analysis. Traditional static graph analysis cannot keep pace with network security applications analyzing nearly one million events per second and social networks like Facebook collecting 500 thousand comments per second. Streaming frameworks like STINGER support ingesting up three million of edge changes per second but there are few streaming analysis kernels that keep up with these rates. Here we present a new algorithm model for applying complex metrics to a changing graph. In this model, many more algorithms can be applied without having to stop the world. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/siam-ns-2017-ejr-170714190645-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Applications in computer network security, social media analysis,and other areas rely on analyzing a changing environment. The data is rich in relationships and lends itself to graph analysis. Traditional static graph analysis cannot keep pace with network security applications analyzing nearly one million events per second and social networks like Facebook collecting 500 thousand comments per second. Streaming frameworks like STINGER support ingesting up three million of edge changes per second but there are few streaming analysis kernels that keep up with these rates. Here we present a new algorithm model for applying complex metrics to a changing graph. In this model, many more algorithms can be applied without having to stop the world.
A New Algorithm Model for Massive-Scale Streaming Graph Analysis from Jason Riedy
]]> 455 2 https://cdn.slidesharecdn.com/ss_thumbnails/siam-ns-2017-ejr-170714190645-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 High-Performance Analysis of Streaming Graphs /slideshow/highperformance-analysis-of-streaming-graphs-77348572/77348572 acs-2017-ejr-170628191931
Graph-structured data in social networks, finance, network security, and others not only are massive but also under continual change. These changes often are scattered across the graph. Stopping the world to run a single, static query is infeasible. Repeating complex global analyses on massive snapshots to capture only what has changed is inefficient. We discuss requirements for single-shot queries on changing graphs as well as recent high-performance algorithms that update rather than recompute results. These algorithms are incorporated into our software framework for streaming graph analysis, STINGER.]]>
Graph-structured data in social networks, finance, network security, and others not only are massive but also under continual change. These changes often are scattered across the graph. Stopping the world to run a single, static query is infeasible. Repeating complex global analyses on massive snapshots to capture only what has changed is inefficient. We discuss requirements for single-shot queries on changing graphs as well as recent high-performance algorithms that update rather than recompute results. These algorithms are incorporated into our software framework for streaming graph analysis, STINGER.]]> Wed, 28 Jun 2017 19:19:31 GMT /slideshow/highperformance-analysis-of-streaming-graphs-77348572/77348572 jasonriedy@slideshare.net(jasonriedy) High-Performance Analysis of Streaming Graphs jasonriedy Graph-structured data in social networks, finance, network security, and others not only are massive but also under continual change. These changes often are scattered across the graph. Stopping the world to run a single, static query is infeasible. Repeating complex global analyses on massive snapshots to capture only what has changed is inefficient. We discuss requirements for single-shot queries on changing graphs as well as recent high-performance algorithms that update rather than recompute results. These algorithms are incorporated into our software framework for streaming graph analysis, STINGER. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/acs-2017-ejr-170628191931-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Graph-structured data in social networks, finance, network security, and others not only are massive but also under continual change. These changes often are scattered across the graph. Stopping the world to run a single, static query is infeasible. Repeating complex global analyses on massive snapshots to capture only what has changed is inefficient. We discuss requirements for single-shot queries on changing graphs as well as recent high-performance algorithms that update rather than recompute results. These algorithms are incorporated into our software framework for streaming graph analysis, STINGER.
High-Performance Analysis of Streaming Graphs from Jason Riedy
]]> 303 5 https://cdn.slidesharecdn.com/ss_thumbnails/acs-2017-ejr-170628191931-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 High-Performance Analysis of Streaming Graphs /slideshow/highperformance-analysis-of-streaming-graphs/72753916 siam-cse-2017-ejr-170302214500
Graph-structured data in social networks, finance, network security, and others not only are massive but also under continual change. These changes often are scattered across the graph. Stopping the world to run a single, static query is infeasible. Repeating complex global analyses on massive snapshots to capture only what has changed is inefficient. We discuss requirements for single-shot queries on changing graphs as well as recent high-performance algorithms that update rather than recompute results. These algorithms are incorporated into our software framework for streaming graph analysis, STING (Spatio-Temporal Interaction Networks and Graphs). ]]>
Graph-structured data in social networks, finance, network security, and others not only are massive but also under continual change. These changes often are scattered across the graph. Stopping the world to run a single, static query is infeasible. Repeating complex global analyses on massive snapshots to capture only what has changed is inefficient. We discuss requirements for single-shot queries on changing graphs as well as recent high-performance algorithms that update rather than recompute results. These algorithms are incorporated into our software framework for streaming graph analysis, STING (Spatio-Temporal Interaction Networks and Graphs). ]]> Thu, 02 Mar 2017 21:45:00 GMT /slideshow/highperformance-analysis-of-streaming-graphs/72753916 jasonriedy@slideshare.net(jasonriedy) High-Performance Analysis of Streaming Graphs jasonriedy Graph-structured data in social networks, finance, network security, and others not only are massive but also under continual change. These changes often are scattered across the graph. Stopping the world to run a single, static query is infeasible. Repeating complex global analyses on massive snapshots to capture only what has changed is inefficient. We discuss requirements for single-shot queries on changing graphs as well as recent high-performance algorithms that update rather than recompute results. These algorithms are incorporated into our software framework for streaming graph analysis, STING (Spatio-Temporal Interaction Networks and Graphs). <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/siam-cse-2017-ejr-170302214500-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Graph-structured data in social networks, finance, network security, and others not only are massive but also under continual change. These changes often are scattered across the graph. Stopping the world to run a single, static query is infeasible. Repeating complex global analyses on massive snapshots to capture only what has changed is inefficient. We discuss requirements for single-shot queries on changing graphs as well as recent high-performance algorithms that update rather than recompute results. These algorithms are incorporated into our software framework for streaming graph analysis, STING (Spatio-Temporal Interaction Networks and Graphs).
High-Performance Analysis of Streaming Graphs from Jason Riedy
]]> 351 3 https://cdn.slidesharecdn.com/ss_thumbnails/siam-cse-2017-ejr-170302214500-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Updating PageRank for Streaming Graphs /slideshow/updating-pagerank-for-streaming-graphs/62498337 gabb-2016-ejr-160528182532
Algorithm for efficiently and accurately updating PageRank as the graph changes from a stream of updates. Also includes needs from the upcoming GraphBLAS to support high-performance streaming graph analysis.]]>
Algorithm for efficiently and accurately updating PageRank as the graph changes from a stream of updates. Also includes needs from the upcoming GraphBLAS to support high-performance streaming graph analysis.]]> Sat, 28 May 2016 18:25:32 GMT /slideshow/updating-pagerank-for-streaming-graphs/62498337 jasonriedy@slideshare.net(jasonriedy) Updating PageRank for Streaming Graphs jasonriedy Algorithm for efficiently and accurately updating PageRank as the graph changes from a stream of updates. Also includes needs from the upcoming GraphBLAS to support high-performance streaming graph analysis. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/gabb-2016-ejr-160528182532-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Algorithm for efficiently and accurately updating PageRank as the graph changes from a stream of updates. Also includes needs from the upcoming GraphBLAS to support high-performance streaming graph analysis.
Updating PageRank for Streaming Graphs from Jason Riedy
]]> 513 6 https://cdn.slidesharecdn.com/ss_thumbnails/gabb-2016-ejr-160528182532-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 https://cdn.slidesharecdn.com/profile-photo-jasonriedy-48x48.jpg?cb=1684157881 Work email address is jason.riedy@cc.gatech.edu. My CV is available at http://lovesgoodfood.com/jason/cv/ , although typically a few months out of date. Some partial publication lists are at http://scholar.google.com/citations?user=1PpZ-r8AAAAJ , http://dl.acm.org/author_page.cfm?id=81327491132 , and http://www.informatik.uni-trier.de/~ley/pers/hd/r/Riedy:E=_Jason.html . Presentations are available either through my CV or http://www.slideshare.net/jasonriedy for now. Specialties: graph analysis, parallel irregular algorithms, data analysis, sparse linear algebra, dense linear algebra, parallel computing, floating-point arithmetic, high-performance computing, scalable services, free sof... lovesgoodfood.com/jason https://cdn.slidesharecdn.com/ss_thumbnails/2021-graphblas-bof-lucata-211014005251-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/lucata-at-the-hpec-graphblas-bof-250439305/250439305 Lucata at the HPEC Gra... https://cdn.slidesharecdn.com/ss_thumbnails/lagraph-2021-10-13-211014005046-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/lagraph-20211013/250439292 LAGraph 2021-10-13 https://cdn.slidesharecdn.com/ss_thumbnails/2021-graphblas-bof-lucata-211014004830-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/lucata-at-the-hpec-graphblas-bof/250439277 Lucata at the HPEC Gra...