{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/108496"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/108496","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Heterogeneous parallel virtual machine: A portable program representation and compiler for performance and energy optimizations on heterogeneous parallel systems","abstract":"\"Programming heterogeneous parallel systems, such as the SoCs (System-on-Chip) on mobile and edge devices is extremely difficult; the diverse parallel hardware they contain exposes vastly different hardware instruction sets, parallelism models and memory systems. Moreover, a wide range of diverse hardware and software approximation techniques are available for applications targeting heterogeneous SoCs, further exacerbating the programmability challenges. In this thesis, we alleviate the programmability challenges of such systems using flexible compiler intermediate representation solutions, in order to benefit from the performance and superior energy efficiency of heterogeneous systems. First, we develop Heterogeneous Parallel Virtual Machine (HPVM), a parallel program representation for heterogeneous systems, designed to enable functional and performance portability across popular parallel hardware. HPVM is based on a hierarchical dataflow graph with side effects. HPVM successfully supports three important capabilities for programming heterogeneous systems: a compiler intermediate representation (IR), a virtual instruction set (ISA), and a basis for runtime scheduling. We use the HPVM representation to implement an HPVM prototype, defining the HPVM IR as an extension of the Low Level Virtual Machine (LLVM) IR. Our results show comparable performance with optimized OpenCL kernels for the target hardware from a single HPVM representation using translators from HPVM virtual ISA to native code, IR optimizations operating directly on the HPVM representation, and the capability for supporting flexible runtime scheduling schemes from a single HPVM representation. We extend HPVM to ApproxHPVM, introducing hardware-independent approximation metrics in the IR to enable maintaining accuracy information at the IR level and mapping of application-level end-to-end quality metrics to system level \"\"knobs\"\". The approximation metrics quantify the acceptable accuracy loss for individual computations. Application programmers only need to specify high-level, and end-to-end, quality metrics, instead of detailed parameters for individual approximation methods. The ApproxHPVM system then automatically tunes the accuracy requirements of individual computations and maps them to approximate hardware when possible. ApproxHPVM results show significant performance and energy improvements for popular deep learning benchmarks. Finally, we extend to ApproxHPVM to ApproxTuner, a compiler and runtime system for approximation. ApproxTuner extends ApproxHPVM with a wide range of hardware and software approximation techniques. It uses a three step approximation tuning strategy, a combination of development-time, install-time, and dynamic tuning. Our strategy ensures software portability, even though approximations have highly hardware-dependent performance, and enables efficient dynamic approximation tuning despite the expensive offline steps. ApproxTuner results show significant performance and energy improvements across 7 Deep Neural Networks and 3 image processing benchmarks, and ensures that high-level end-to-end quality specifications are satisfied during adaptive approximation tuning.\"","abstract_html":""Programming heterogeneous parallel systems, such as the SoCs (System-on-Chip) on mobile and edge devices is extremely difficult; the diverse parallel hardware they contain exposes vastly different hardware instruction sets, parallelism models and memory systems. Moreover, a wide range of diverse hardware and software approximation techniques are available for applications targeting heterogeneous SoCs, further exacerbating the programmability challenges. In this thesis, we alleviate the programmability challenges of such systems using flexible compiler intermediate representation solutions, in order to benefit from the performance and superior energy efficiency of heterogeneous systems. First, we develop Heterogeneous Parallel Virtual Machine (HPVM), a parallel program representation for heterogeneous systems, designed to enable functional and performance portability across popular parallel hardware. HPVM is based on a hierarchical dataflow graph with side effects. HPVM successfully supports three important capabilities for programming heterogeneous systems: a compiler intermediate representation (IR), a virtual instruction set (ISA), and a basis for runtime scheduling. We use the HPVM representation to implement an HPVM prototype, defining the HPVM IR as an extension of the Low Level Virtual Machine (LLVM) IR. Our results show comparable performance with optimized OpenCL kernels for the target hardware from a single HPVM representation using translators from HPVM virtual ISA to native code, IR optimizations operating directly on the HPVM representation, and the capability for supporting flexible runtime scheduling schemes from a single HPVM representation. We extend HPVM to ApproxHPVM, introducing hardware-independent approximation metrics in the IR to enable maintaining accuracy information at the IR level and mapping of application-level end-to-end quality metrics to system level ""knobs"". The approximation metrics quantify the acceptable accuracy loss for individual computations. Application programmers only need to specify high-level, and end-to-end, quality metrics, instead of detailed parameters for individual approximation methods. The ApproxHPVM system then automatically tunes the accuracy requirements of individual computations and maps them to approximate hardware when possible. ApproxHPVM results show significant performance and energy improvements for popular deep learning benchmarks. Finally, we extend to ApproxHPVM to ApproxTuner, a compiler and runtime system for approximation. ApproxTuner extends ApproxHPVM with a wide range of hardware and software approximation techniques. It uses a three step approximation tuning strategy, a combination of development-time, install-time, and dynamic tuning. Our strategy ensures software portability, even though approximations have highly hardware-dependent performance, and enables efficient dynamic approximation tuning despite the expensive offline steps. ApproxTuner results show significant performance and energy improvements across 7 Deep Neural Networks and 3 image processing benchmarks, and ensures that high-level end-to-end quality specifications are satisfied during adaptive approximation tuning."","abstract_has_math":false,"creators":["Kotsifakou, Maria"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":["Adve, Vikram S","Adve, Sarita V","Misailovic, Sasa","Brooks, David M"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2020,"date_issued":"2020-10-07T20:59:54Z","date_published":"2020-10-07T20:59:54Z","updated_at":"2026-07-22T22:24:48Z","subjects":["Virtual ISA","Compiler","Parallel IR","Heterogeneous Systems","GPU","Vector SIMD","Approximate Computing","Deep Neural Networks","DNNs","Approximation Tuning"],"languages":["en"],"rights":["Copyright 2020 Maria Kotsifakou"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/108496","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Adve, Vikram S","Adve, Sarita V","Misailovic, Sasa","Brooks, David M"]},{"key":"dc:creator","label":"Author","values":["Kotsifakou, Maria"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2020-10-07T20:59:54Z","2020-07-15","2020-08"]},{"key":"dc:type","label":"Dc Type","values":["text","Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Virtual ISA","Compiler","Parallel IR","Heterogeneous Systems","GPU","Vector SIMD","Approximate Computing","Deep Neural Networks","DNNs","Approximation Tuning"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2020 Maria Kotsifakou"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/108496"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["\"Programming heterogeneous parallel systems, such as the SoCs (System-on-Chip) on mobile and edge devices is extremely difficult; the diverse parallel hardware they contain exposes vastly different hardware instruction sets, parallelism models and memory systems. Moreover, a wide range of diverse hardware and software approximation techniques are available for applications targeting heterogeneous SoCs, further exacerbating the programmability challenges. In this thesis, we alleviate the programmability challenges of such systems using flexible compiler intermediate representation solutions, in order to benefit from the performance and superior energy efficiency of heterogeneous systems. First, we develop Heterogeneous Parallel Virtual Machine (HPVM), a parallel program representation for heterogeneous systems, designed to enable functional and performance portability across popular parallel hardware. HPVM is based on a hierarchical dataflow graph with side effects. HPVM successfully supports three important capabilities for programming heterogeneous systems: a compiler intermediate representation (IR), a virtual instruction set (ISA), and a basis for runtime scheduling. We use the HPVM representation to implement an HPVM prototype, defining the HPVM IR as an extension of the Low Level Virtual Machine (LLVM) IR. Our results show comparable performance with optimized OpenCL kernels for the target hardware from a single HPVM representation using translators from HPVM virtual ISA to native code, IR optimizations operating directly on the HPVM representation, and the capability for supporting flexible runtime scheduling schemes from a single HPVM representation. We extend HPVM to ApproxHPVM, introducing hardware-independent approximation metrics in the IR to enable maintaining accuracy information at the IR level and mapping of application-level end-to-end quality metrics to system level \"\"knobs\"\". The approximation metrics quantify the acceptable accuracy loss for individual computations. Application programmers only need to specify high-level, and end-to-end, quality metrics, instead of detailed parameters for individual approximation methods. The ApproxHPVM system then automatically tunes the accuracy requirements of individual computations and maps them to approximate hardware when possible. ApproxHPVM results show significant performance and energy improvements for popular deep learning benchmarks. Finally, we extend to ApproxHPVM to ApproxTuner, a compiler and runtime system for approximation. ApproxTuner extends ApproxHPVM with a wide range of hardware and software approximation techniques. It uses a three step approximation tuning strategy, a combination of development-time, install-time, and dynamic tuning. Our strategy ensures software portability, even though approximations have highly hardware-dependent performance, and enables efficient dynamic approximation tuning despite the expensive offline steps. ApproxTuner results show significant performance and energy improvements across 7 Deep Neural Networks and 3 image processing benchmarks, and ensures that high-level end-to-end quality specifications are satisfied during adaptive approximation tuning.\"","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2020-10-02 without embargo terms","The student, Maria Kotsifakou, accepted the attached license on 2020-07-14 at 20:39.","The student, Maria Kotsifakou, submitted this Dissertation for approval on 2020-07-14 at 21:24.","This Dissertation was approved for publication on 2020-07-15 at 11:29.","DSpace SAF Submission Ingestion Package generated from Vireo submission #15620 on 2020-10-02 at 15:13:50","Made available in DSpace on 2020-10-07T20:59:54Z (GMT). No. of bitstreams: 2 KOTSIFAKOU-DISSERTATION-2020.pdf: 2615323 bytes, checksum: 0c7e2cdbb87aa33aec3c7a0ec8181225 (MD5) LICENSE.txt: 4213 bytes, checksum: 813a50220874aec5284c1abb4d1f7c3a (MD5) Previous issue date: 2020-07-15"]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Heterogeneous parallel virtual machine: A portable program representation and compiler for performance and energy optimizations on heterogeneous parallel systems"]}]}],"canonical_facts":{"dc:contributor":["Adve, Vikram S","Adve, Sarita V","Misailovic, Sasa","Brooks, David M"],"dc:creator":["Kotsifakou, Maria"],"dc:date":["2020-10-07T20:59:54Z","2020-07-15","2020-08"],"dc:description":["\"Programming heterogeneous parallel systems, such as the SoCs (System-on-Chip) on mobile and edge devices is extremely difficult; the diverse parallel hardware they contain exposes vastly different hardware instruction sets, parallelism models and memory systems. Moreover, a wide range of diverse hardware and software approximation techniques are available for applications targeting heterogeneous SoCs, further exacerbating the programmability challenges. In this thesis, we alleviate the programmability challenges of such systems using flexible compiler intermediate representation solutions, in order to benefit from the performance and superior energy efficiency of heterogeneous systems. First, we develop Heterogeneous Parallel Virtual Machine (HPVM), a parallel program representation for heterogeneous systems, designed to enable functional and performance portability across popular parallel hardware. HPVM is based on a hierarchical dataflow graph with side effects. HPVM successfully supports three important capabilities for programming heterogeneous systems: a compiler intermediate representation (IR), a virtual instruction set (ISA), and a basis for runtime scheduling. We use the HPVM representation to implement an HPVM prototype, defining the HPVM IR as an extension of the Low Level Virtual Machine (LLVM) IR. Our results show comparable performance with optimized OpenCL kernels for the target hardware from a single HPVM representation using translators from HPVM virtual ISA to native code, IR optimizations operating directly on the HPVM representation, and the capability for supporting flexible runtime scheduling schemes from a single HPVM representation. We extend HPVM to ApproxHPVM, introducing hardware-independent approximation metrics in the IR to enable maintaining accuracy information at the IR level and mapping of application-level end-to-end quality metrics to system level \"\"knobs\"\". The approximation metrics quantify the acceptable accuracy loss for individual computations. Application programmers only need to specify high-level, and end-to-end, quality metrics, instead of detailed parameters for individual approximation methods. The ApproxHPVM system then automatically tunes the accuracy requirements of individual computations and maps them to approximate hardware when possible. ApproxHPVM results show significant performance and energy improvements for popular deep learning benchmarks. Finally, we extend to ApproxHPVM to ApproxTuner, a compiler and runtime system for approximation. ApproxTuner extends ApproxHPVM with a wide range of hardware and software approximation techniques. It uses a three step approximation tuning strategy, a combination of development-time, install-time, and dynamic tuning. Our strategy ensures software portability, even though approximations have highly hardware-dependent performance, and enables efficient dynamic approximation tuning despite the expensive offline steps. ApproxTuner results show significant performance and energy improvements across 7 Deep Neural Networks and 3 image processing benchmarks, and ensures that high-level end-to-end quality specifications are satisfied during adaptive approximation tuning.\"","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2020-10-02 without embargo terms","The student, Maria Kotsifakou, accepted the attached license on 2020-07-14 at 20:39.","The student, Maria Kotsifakou, submitted this Dissertation for approval on 2020-07-14 at 21:24.","This Dissertation was approved for publication on 2020-07-15 at 11:29.","DSpace SAF Submission Ingestion Package generated from Vireo submission #15620 on 2020-10-02 at 15:13:50","Made available in DSpace on 2020-10-07T20:59:54Z (GMT). No. of bitstreams: 2 KOTSIFAKOU-DISSERTATION-2020.pdf: 2615323 bytes, checksum: 0c7e2cdbb87aa33aec3c7a0ec8181225 (MD5) LICENSE.txt: 4213 bytes, checksum: 813a50220874aec5284c1abb4d1f7c3a (MD5) Previous issue date: 2020-07-15"],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/2142/108496"],"dc:language":["en"],"dc:rights":["Copyright 2020 Maria Kotsifakou"],"dc:subject":["Virtual ISA","Compiler","Parallel IR","Heterogeneous Systems","GPU","Vector SIMD","Approximate Computing","Deep Neural Networks","DNNs","Approximation Tuning"],"dc:title":["Heterogeneous parallel virtual machine: A portable program representation and compiler for performance and energy optimizations on heterogeneous parallel systems"],"dc:type":["text","Thesis"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:48Z"}