Dario Korolija
Dario Korolija has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Gustavo Alonso
ETH Zurich, Zurich, Switzerland
6 shared publications - Benjamin Ramhorst
ETH Zurich, Zurich, Switzerland
2 shared publications - Maximilian Jakob Heer
ETH Zurich, Zurich, Switzerland
2 shared publications - Abishek Ramdas
Systems Group, ETH Zurich
1 shared publication - Adam Turowski
Systems Group, ETH Zurich
1 shared publication - Anastasiia Ruzhanskaia
Systems Group, ETH Zurich
1 shared publication
Latest Publications
- Open Access
- 26 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1829-1846
RoPeerTo: A Datacenter-Scale Architecture for Peer-To-Peer DMA between GPUs and FPGAs
- 26 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1829-1846
Open AccessRoPeerTo: A Datacenter-Scale Architecture for Peer-To-Peer DMA between GPUs and FPGAs
- Open Access
- 12 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 639-654
Coyote v2: Raising the Level of Abstraction for Data Center FPGAs
- 12 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 639-654
Open AccessCoyote v2: Raising the Level of Abstraction for Data Center FPGAs
- Open Access
- 25 August 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computer Systems
CCKit: An open-source toolkit for cache coherent accelerators
The trend towards system specialization is leading to a proliferation of accelerators, exposing interconnects as serious bottlenecks, both in functionality and performance. As a result, several alternative approaches have been proposed which promise to expand the coherence domain beyond homogeneous sockets to rack scale heterogeneous systems. In parallel, GPU vendors have developed their own high bandwidth interconnects also aiming for heterogeneous coherence beyond the CPU. This expansion of the coherency domain raises many questions that remain unanswered, in particular, how devices other than CPUs will interact with the coherence protocol and whether applications can take advantage of these expanded domains. As protocols such as CXL are still evolving, it is important to explore alternative designs that go beyond what the commercial specifications dictate. For this purpose, we developed CCKit, an open-source, server-class toolkit comprising a complete cache coherency stack on reconfigurable accelerators. CCKit is more flexible than commercial products and its performance is highly competitive with hardware-based implementations, thus enabling important and novel application use-cases for expanded coherence domains. Experimental data from real workloads provide the ability to influence and expand future interconnects, protocols, and applications.
- 25 August 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computer Systems
Open AccessCCKit: An open-source toolkit for cache coherent accelerators
The trend towards system specialization is leading to a proliferation of accelerators, exposing interconnects as serious bottlenecks, both in functionality and performance. As a result, several alternative approaches have been proposed which promise to expand the coherence domain beyond homogeneous sockets to rack scale heterogeneous systems. In parallel, GPU vendors have developed their own high bandwidth interconnects also aiming for heterogeneous coherence beyond the CPU. This expansion of the coherency domain raises many questions that remain unanswered, in particular, how devices other than CPUs will interact with the coherence protocol and whether applications can take advantage of these expanded domains. As protocols such as CXL are still evolving, it is important to explore alternative designs that go beyond what the commercial specifications dictate. For this purpose, we developed CCKit, an open-source, server-class toolkit comprising a complete cache coherency stack on reconfigurable accelerators. CCKit is more flexible than commercial products and its performance is highly competitive with hardware-based implementations, thus enabling important and novel application use-cases for expanded coherence domains. Experimental data from real workloads provide the ability to influence and expand future interconnects, protocols, and applications.
- Open Access
- 4 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 77-82
Data Processing with FPGAs on Modern Architectures
Trends in hardware, the prevalence of the cloud, and the rise of highly demanding applications have ushered an era of specialization that is quickly changing the way data is processed at scale. These changes are likely to continue and accelerate in the next years as new technologies are adopted and deployed: smart NICs, smart storage, smart memory, disaggregated storage, disaggregated memory, specialized accelerators (GPUS, TPUs, FPGAs), as well as a wealth of ASICS specifically created to deal with computationally expensive tasks (e.g., cryptography or compression). In this tutorial we focus on data processing on FPGAs, a technology that has received less attention than, e.g., TPUs or GPUs but that is, however, increasingly being deployed in the cloud for data processing tasks due to the architectural flexibility of FPGAs and their ability to process data at line rate, something not possible with other type of processors or accelerators. In the tutorial we will cover what are FPGAs, their characteristics, their advantages and disadvantages over other design options, as well as examples from deployments in industry and how they are used in a variety of data processing tasks. Then we will provide a brief introduction to FPGA programming with High Level Synthesis (HLS) tools as well as briefly describe resources available to researchers in the form of academic clusters and open source systems that simplify the first steps. The tutorial will also include several case studies borrowed from research done in collaboration with companies that illustrate both the potential of FPGAs in data processing but also how software and hardware architectures are evolving to take advantage of the possibilities offered by FPGAs.
- 4 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 77-82
Open AccessData Processing with FPGAs on Modern Architectures
Trends in hardware, the prevalence of the cloud, and the rise of highly demanding applications have ushered an era of specialization that is quickly changing the way data is processed at scale. These changes are likely to continue and accelerate in the next years as new technologies are adopted and deployed: smart NICs, smart storage, smart memory, disaggregated storage, disaggregated memory, specialized accelerators (GPUS, TPUs, FPGAs), as well as a wealth of ASICS specifically created to deal with computationally expensive tasks (e.g., cryptography or compression). In this tutorial we focus on data processing on FPGAs, a technology that has received less attention than, e.g., TPUs or GPUs but that is, however, increasingly being deployed in the cloud for data processing tasks due to the architectural flexibility of FPGAs and their ability to process data at line rate, something not possible with other type of processors or accelerators. In the tutorial we will cover what are FPGAs, their characteristics, their advantages and disadvantages over other design options, as well as examples from deployments in industry and how they are used in a variety of data processing tasks. Then we will provide a brief introduction to FPGA programming with High Level Synthesis (HLS) tools as well as briefly describe resources available to researchers in the form of academic clusters and open source systems that simplify the first steps. The tutorial will also include several case studies borrowed from research done in collaboration with companies that illustrate both the potential of FPGAs in data processing but also how software and hardware architectures are evolving to take advantage of the possibilities offered by FPGAs.
- Open Access
- 8 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-4
Serverless FPGA: Work-In-Progress
In this short paper we investigate the combination of two emerging technologies: the tight provisioning requirements of Serverless computing and the acceleration potential of FPGAs. Serverless platforms suffer from container overheads, notably cold start latency, while having to adapt to Function-as-a-Service (FaaS) workloads. By exploring re-configurability of FPGAs and their acceleration power, we propose an innovative light-weight Serverless platform for FPGA-based FaaS applications which aims to reduce these overheads. In this study, we explore the feasibility of the idea by implementing key elements of such platform onto the FPGA. Our initial results show potential for acceleration in all aspects of function invocation.
- 8 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-4
Open AccessServerless FPGA: Work-In-Progress
In this short paper we investigate the combination of two emerging technologies: the tight provisioning requirements of Serverless computing and the acceleration potential of FPGAs. Serverless platforms suffer from container overheads, notably cold start latency, while having to adapt to Function-as-a-Service (FaaS) workloads. By exploring re-configurability of FPGAs and their acceleration power, we propose an innovative light-weight Serverless platform for FPGA-based FaaS applications which aims to reduce these overheads. In this study, we explore the feasibility of the idea by implementing key elements of such platform onto the FPGA. Our initial results show potential for acceleration in all aspects of function invocation.
- Open Access
- 1 August 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 197-203
EasyNet: 100 Gbps Network for HLS
The massive deployment of FPGAs in data centers is opening up new opportunities for accelerating distributed applications. However, developing a distributed FPGA application remains difficult for two reasons. First, commonly available development frameworks (e.g., Xilinx Vitis) lack explicit support for networking. Developers are, thus, forced to build their own infrastructure to handle the data movement between the host, the FPGA, and the network. Second, distributed applications are made even more complex by using low level interfaces to access the network and process packets. Ideally, one needs to combine high performance with a simple interface for both point-to-point and collective operations. To overcome these inefficiencies and enable further research in networking and distributed application on FPGAs, we first show how to integrate an open-source 100 Gbps TCP/IP stack into a state-of-the-art FPGA development framework (Xilinx Vitis) without degrading its performance. Further, we provide a set of MPI-like communication primitives for both point-to-point and collective operations as a High Level Synthesis (HLS) library. Our point-to-point primitives saturate a 100 Gbps link and our collective primitives achieve low latency. With our approach, developers can write hardware kernels in high level languages with the network abstracted away behind standard interfaces. To evaluate the ease of use and performance in a real application, we distribute a K-Means algorithm with the new stack and achieve a 1.9X and 3.5X throughput increase with 2 FPGAs and 4 FPGAs respectively.
- 1 August 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 197-203
Open AccessEasyNet: 100 Gbps Network for HLS
The massive deployment of FPGAs in data centers is opening up new opportunities for accelerating distributed applications. However, developing a distributed FPGA application remains difficult for two reasons. First, commonly available development frameworks (e.g., Xilinx Vitis) lack explicit support for networking. Developers are, thus, forced to build their own infrastructure to handle the data movement between the host, the FPGA, and the network. Second, distributed applications are made even more complex by using low level interfaces to access the network and process packets. Ideally, one needs to combine high performance with a simple interface for both point-to-point and collective operations. To overcome these inefficiencies and enable further research in networking and distributed application on FPGAs, we first show how to integrate an open-source 100 Gbps TCP/IP stack into a state-of-the-art FPGA development framework (Xilinx Vitis) without degrading its performance. Further, we provide a set of MPI-like communication primitives for both point-to-point and collective operations as a High Level Synthesis (HLS) library. Our point-to-point primitives saturate a 100 Gbps link and our collective primitives achieve low latency. With our approach, developers can write hardware kernels in high level languages with the network abstracted away behind standard interfaces. To evaluate the ease of use and performance in a real application, we distribute a K-Means algorithm with the new stack and achieve a 1.9X and 3.5X throughput increase with 2 FPGAs and 4 FPGAs respectively.
- 1 May 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 155-162
FPGA-Assisted Deterministic Routing for FPGAs
FPGA routing is one of the most time-consuming steps of FPGA compilation, often preventing fast edit-compile-test cycles in prototyping and development. There have been attempts to accelerate FPGA routing using algorithmic improvements, multi-core or multi-CPU platforms. Instead, we propose porting FPGA routing to a CPU+FPGA platform. Motivated by the approaches used in FPGA-accelerated graph processing, we propose and implement three acceleration strategies: (1) reducing the number of expensive random memory accesses, (2) parallel and pipelined computation, and (3) efficient hardware priority queues. To test and evaluate the router performance, we implement it on DE1-SoC, a mid-end ARM+FPGA platform of Intel. Our router works and produces good quality results. Moreover, we succeed in accelerating the software router running on the embedded ARM. However, when compared to the latest VPR router running on a powerful Intel Core-i5 CPU, our CPU+FPGA router is slower. This is not unexpected, given the limited performance of the chosen hardware platform. Since this design can easily be ported to newer and higher-end CPU+FPGA systems, we estimate the performance it could achieve; the results indicate that a non-negligible speedup over the software-only router could indeed be obtained.
- 1 May 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 155-162
FPGA-Assisted Deterministic Routing for FPGAs
FPGA routing is one of the most time-consuming steps of FPGA compilation, often preventing fast edit-compile-test cycles in prototyping and development. There have been attempts to accelerate FPGA routing using algorithmic improvements, multi-core or multi-CPU platforms. Instead, we propose porting FPGA routing to a CPU+FPGA platform. Motivated by the approaches used in FPGA-accelerated graph processing, we propose and implement three acceleration strategies: (1) reducing the number of expensive random memory accesses, (2) parallel and pipelined computation, and (3) efficient hardware priority queues. To test and evaluate the router performance, we implement it on DE1-SoC, a mid-end ARM+FPGA platform of Intel. Our router works and produces good quality results. Moreover, we succeed in accelerating the software router running on the embedded ARM. However, when compared to the latest VPR router running on a powerful Intel Core-i5 CPU, our CPU+FPGA router is slower. This is not unexpected, given the limited performance of the chosen hardware platform. Since this design can easily be ported to newer and higher-end CPU+FPGA systems, we estimate the performance it could achieve; the results indicate that a non-negligible speedup over the software-only router could indeed be obtained.