research
Publications, talks, and posters by categories in reversed chronological order.
Publications
2024
- Retargeting and Respecializing GPU Workloads for Performance PortabilityIn 2024 IEEE/ACM International Symposium on Code Generation and Optimization (CGO), Mar 2024
2023
- High-Performance GPU-to-CPU Transpilation and Optimization via High-Level Parallel ConstructsIn Proceedings of the 28th ACM SIGPLAN Annual Symposium on Principles and Practice of Parallel Programming, Mar 2023
2019
- HyperX Topology: First at-Scale Implementation and Comparison to the Fat-TreeIn Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, Mar 2019
Talks
Apr. 2024
I. R. Ivanov and W. S. Moses. Automatic Retuning of Floating-Point Precision -Ivan R. Ivanov, William S. Moses
Apr. 2024
Ivan R. Ivanov, Aiden Grossman, Ludger Paehler, William S. Moses, Johannes Doerfert Automatic Proxy App Generation through Input Capture and Generation
Mar. 2024
I. R. Ivanov, O. Zinenko, J. Domke, T. Endo, and W. S. Moses. Retargeting and Respecializing GPU Workloads for Performance Portability. CGO ‘24
Mar. 2024
I. R. Ivanov, J. Domke, T. Endo, J. Doerfert. Automatic Parallelization and OpenMP Offloading of Fortran. CGO ‘24 LLVM Performance Workshop
Oct. 2023
I. R. Ivanov, O. Zinenko, J. Domke, T. Endo, and W. S. Moses. Optimization of CUDA GPU Kernels and Translation to AMDGPU in Polygeist/MLIR. 2023 LLVM Developers’ Meeting. Student Talk
Oct. 2023
I. R. Ivanov, O. Zinenko, J. Domke, T. Endo, J. Doerfert, and W. S. Moses. GPU Kernel Compilation in Polygeist/MLIR. 2023 LLVM Developers’ Meeting GPU Offloading Workshop. Lightning Talk
Nov. 2022
W. S. Moses, I. R. Ivanov, J. Domke, T. Endo, J. Doerfert, and O. Zinenko. High-Performance GPU-to-CPU Transpilation and Optimization via High-Level Parallel Constructs in Polygeist/MLIR. 2022 LLVM Developers’ Meeting. Lightning Talk
Feb. 2022
I. R. Ivanov, J. Domke, and T. Endo. Automatic translation of CUDA code into high performance CPU code using LLVM IR transformations. The 4th R-CCS Internation Symposium. Lightning Talk
Feb. 2021
I. R. Ivanov, J. Domke, A. Nomura and T. Endo. Improved failover for HPC interconnects through localised routing restoration. The 3rd R-CCS Internation Symposium. Lightning Talk
Posters
Apr. 2024
I. R. Ivanov, J. Domke, T. Endo, and J. Doerfert. Automatic Parallelization and OpenMP Offloading of Fortran. JLESC 16
Mar. 2024
I. R. Ivanov, J. Domke, T. Endo, and J. Doerfert. Dynamic Thread Coarsening for OpenMP Offloading. CGO ‘24 Student Research Competition.
Mar. 2024
S. Burak, I. R. Ivanov, J. Domke, M. Mueller. Unifying SPMD and Multi-Value IR - Use Case: Static Verification of Collective Communication. CGO ‘24 Student Research Competition.
Sep. 2023
I. R. Ivanov, O. Zinenko, J. Domke, T. Endo, and W. S. Moses. Performance Portability of C/C++ CUDA Code via High-Level Intermediate Representation. 2023 RIKEN Summer School
May. 2023
Ryan Barton, Mohamed Wahib, Jens Domke, Ivan R. Ivanov, Toshio Endo, Satoshi Matsuoka. BITFLEX - An HPC User-Driven Automatic Toolchain for Precision Manipulation and Approximate Computing. ISC High Performance 2023
Feb. 2023
Ivan R. Ivanov, William S. Moses, Jens Domke, Toshio Endo. Parallel Optimizations and Transformations of GPU Kernels Using a High-Level representation in MLIR/Polygeist. CGO ‘23 Student Research Competition
Nov. 2022
W. S. Moses, I. R. Ivanov, J. Domke, T. Endo, J. Doerfert, and O. Zinenko. High-Performance GPU-to-CPU Transpilation and Optimization via High-Level Parallel Constructs in Polygeist/MLIR. 2022 LLVM Developers’ Meeting.
Theses
Mar 2023 - Master’s Thesis
Optimizations and Transformations of Parallel Code via High Level Intermediate Representation
Mar 2021 - Bachelor’s Thesis
Improved failover for HPC interconnects through localised routing restoration