Rakesh Kumar
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Rakesh Kumar has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Fatma Nisa Bostanci
ETH Zurich, Zurich, Switzerland
3 shared publications - Konstantinos Kanellopoulos
ETH Zurich, Zurich, Switzerland
3 shared publications - Mohammad Sadrosadati
ETH Zurich
3 shared publications - Onur Mutlu
ETH Zurich
3 shared publications - Rahul Bera
ETH Zurich, Zurich, Switzerland
3 shared publications - Nandita Vijaykumar
University of Toronto, Toronto, ON, Canada
2 shared publications
Latest Publications
- 11 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Computer Architecture Letters
- Vol. 25 (1) , 57-60
Driving the Core Frontend With LiteBTB
Branch target buffer (BTB) is a central component of high performance core front-ends as it not only steers instruction fetch by uncovering upcoming control flow but also enables highly effective fetch-directed instruction prefetching. However, the massive instruction footprints of modern server applications far exceed the capacities of moderately sized BTBs, resulting in frequent misses that inevitably hurt performance. While commercial CPUs deploy large BTBs to mitigate this problem, they incur high storage and area overheads. Prior efforts to reduce BTB storage have primarily targeted branch targets, which has proven highly effective — so much that the tag storage now dominates the BTB storage budget. We make a key observation that BTBs exhibit a large degree of tag redundancy, i.e. only a small fraction of entries contain unique tags, and this fraction falls sharply as BTB capacity grows. Leveraging this insight, we propose LiteBTB, which employs a dedicated hardware structure to store unique tags only once and replaces per-entry tags in BTB with compact tag pointers. To avoid latency overheads, LiteBTB accesses the tag storage and BTB in parallel. Our evaluation shows that LiteBTB reduces storage by up to 13.1% compared to the state-of-the-art BTB design, called BTB-X, while maintaining equivalent performance. Alternatively, with the same storage budget, LiteBTB accommodates up to 1.125× more branches, yielding up to 2.7% performance improvement.
- 11 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Computer Architecture Letters
- Vol. 25 (1) , 57-60
Driving the Core Frontend With LiteBTB
Branch target buffer (BTB) is a central component of high performance core front-ends as it not only steers instruction fetch by uncovering upcoming control flow but also enables highly effective fetch-directed instruction prefetching. However, the massive instruction footprints of modern server applications far exceed the capacities of moderately sized BTBs, resulting in frequent misses that inevitably hurt performance. While commercial CPUs deploy large BTBs to mitigate this problem, they incur high storage and area overheads. Prior efforts to reduce BTB storage have primarily targeted branch targets, which has proven highly effective — so much that the tag storage now dominates the BTB storage budget. We make a key observation that BTBs exhibit a large degree of tag redundancy, i.e. only a small fraction of entries contain unique tags, and this fraction falls sharply as BTB capacity grows. Leveraging this insight, we propose LiteBTB, which employs a dedicated hardware structure to store unique tags only once and replaces per-entry tags in BTB with compact tag pointers. To avoid latency overheads, LiteBTB accesses the tag storage and BTB in parallel. Our evaluation shows that LiteBTB reduces storage by up to 13.1% compared to the state-of-the-art BTB design, called BTB-X, while maintaining equivalent performance. Alternatively, with the same storage budget, LiteBTB accommodates up to 1.125× more branches, yielding up to 2.7% performance improvement.
- Open Access
- 30 March 2025
- Published by Association for Computing Machinery (ACM)
- p. 1400-1421
Virtuoso: Enabling Fast and Accurate Virtual Memory Research via an Imitation-based Operating System Simulation Methodology
- 30 March 2025
- Published by Association for Computing Machinery (ACM)
- p. 1400-1421
Open AccessVirtuoso: Enabling Fast and Accurate Virtual Memory Research via an Imitation-based Operating System Simulation Methodology
- Open Access
- 22 April 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-8
CoFaaS: Automatic Transformation-based Consolidation of Serverless Functions
The attractive property of decoupling deployment decisions from application development has led to fast adoption of the Function-as-a-Service (FaaS) cloud computing model. FaaS applications are typically highly modular with each function having a specific purpose and well-defined interface as this enables code reuse, simplifies maintenance, improves testability, and provides language independence. To enable these attractive features, FaaS applications often use Remote Procedure Call (RPC) interfaces for inter-function communication --- which comes at the cost of orders of magnitude higher latency than native function calls. Prior works that alleviate this overhead are unattractive because they either require non-standard APIs, only support a single language, or rely on specialized languages or runtimes. Our key insight is that we can exploit the well-defined RPC interfaces to perform code transformations that alleviate inter-function communication overhead. We hence propose CoFaaS which leverages this insight to consolidate FaaS functions on top of a WebAssembly runtime and thereby avoid accessing the network layer when functions are deployed on the same compute node. Our evaluation shows that this strategy is highly effective and reports that CoFaaS reduces inter-function communication latency by up to 100× and application-level request round-trip time by up to 6×.
- 22 April 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Open AccessCoFaaS: Automatic Transformation-based Consolidation of Serverless Functions
The attractive property of decoupling deployment decisions from application development has led to fast adoption of the Function-as-a-Service (FaaS) cloud computing model. FaaS applications are typically highly modular with each function having a specific purpose and well-defined interface as this enables code reuse, simplifies maintenance, improves testability, and provides language independence. To enable these attractive features, FaaS applications often use Remote Procedure Call (RPC) interfaces for inter-function communication --- which comes at the cost of orders of magnitude higher latency than native function calls. Prior works that alleviate this overhead are unattractive because they either require non-standard APIs, only support a single language, or rely on specialized languages or runtimes. Our key insight is that we can exploit the well-defined RPC interfaces to perform code transformations that alleviate inter-function communication overhead. We hence propose CoFaaS which leverages this insight to consolidate FaaS functions on top of a WebAssembly runtime and thereby avoid accessing the network layer when functions are deployed on the same compute node. Our evaluation shows that this strategy is highly effective and reports that CoFaaS reduces inter-function communication latency by up to 100× and application-level request round-trip time by up to 6×.
- Open Access
- 28 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 1196-1212
Utopia: Fast and Efficient Address Translation via Hybrid Restrictive & Flexible Virtual-to-Physical Address Mappings
Conventional virtual memory (VM) frameworks enable a virtual address to flexibly map to any physical address. This flexibility necessitates large data structures to store virtual-to-physical mappings, which leads to high address translation latency and large translation-induced interference in the memory hierarchy, especially in data-intensive workloads. On the other hand, restricting the address mapping so that a virtual address can only map to a specific set of physical addresses can significantly reduce address translation overheads by making use of compact and efficient translation structures. However, restricting the address mapping flexibility across the entire main memory severely limits data sharing across different processes and increases data accesses to the swap space of the storage device even in the presence of free memory. We propose Utopia, a new hybrid virtual-to-physical address mapping scheme that allows both flexible and restrictive hash-based address mapping schemes to harmoniously co-exist in the system. The key idea of Utopia is to manage physical memory using two types of physical memory segments: restrictive segments and flexible segments. A restrictive segment uses a restrictive, hash-based address mapping scheme that maps virtual addresses to only a specific set of physical addresses and enables faster address translation using compact translation structures. A flexible segment employs the conventional fully-flexible address mapping scheme. By mapping data to a restrictive segment, Utopia enables faster address translation with lower translation-induced interference. At the same time, Utopia retains the ability to use the flexible address mapping to (i) support conventional VM features such as data sharing and (ii) avoid storing data in the swap space of the storage device when program data does not fit inside a restrictive segment. Our evaluation using 11 diverse data-intensive workloads shows that Utopia improves performance by 24% in a single-core system over the baseline conventional four-level radix-tree page table design, whereas the best prior state-of-the-art contiguity-aware translation scheme improves performance by 13%. Utopia provides 95% of the performance benefits of an ideal address translation scheme where every translation request hits in the first-level TLB. All of Utopia’s benefits come at a modest cost of 0.64% area overhead and 0.72% power overhead compared to a modern high-end CPU. The source code of Utopia is freely available at https://github.com/CMU-SAFARI/Utopia.
- 28 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 1196-1212
Open AccessUtopia: Fast and Efficient Address Translation via Hybrid Restrictive & Flexible Virtual-to-Physical Address Mappings
Conventional virtual memory (VM) frameworks enable a virtual address to flexibly map to any physical address. This flexibility necessitates large data structures to store virtual-to-physical mappings, which leads to high address translation latency and large translation-induced interference in the memory hierarchy, especially in data-intensive workloads. On the other hand, restricting the address mapping so that a virtual address can only map to a specific set of physical addresses can significantly reduce address translation overheads by making use of compact and efficient translation structures. However, restricting the address mapping flexibility across the entire main memory severely limits data sharing across different processes and increases data accesses to the swap space of the storage device even in the presence of free memory. We propose Utopia, a new hybrid virtual-to-physical address mapping scheme that allows both flexible and restrictive hash-based address mapping schemes to harmoniously co-exist in the system. The key idea of Utopia is to manage physical memory using two types of physical memory segments: restrictive segments and flexible segments. A restrictive segment uses a restrictive, hash-based address mapping scheme that maps virtual addresses to only a specific set of physical addresses and enables faster address translation using compact translation structures. A flexible segment employs the conventional fully-flexible address mapping scheme. By mapping data to a restrictive segment, Utopia enables faster address translation with lower translation-induced interference. At the same time, Utopia retains the ability to use the flexible address mapping to (i) support conventional VM features such as data sharing and (ii) avoid storing data in the swap space of the storage device when program data does not fit inside a restrictive segment. Our evaluation using 11 diverse data-intensive workloads shows that Utopia improves performance by 24% in a single-core system over the baseline conventional four-level radix-tree page table design, whereas the best prior state-of-the-art contiguity-aware translation scheme improves performance by 13%. Utopia provides 95% of the performance benefits of an ideal address translation scheme where every translation request hits in the first-level TLB. All of Utopia’s benefits come at a modest cost of 0.64% area overhead and 0.72% power overhead compared to a modern high-end CPU. The source code of Utopia is freely available at https://github.com/CMU-SAFARI/Utopia.
- Open Access
- 28 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 1178-1195
Victima: Drastically Increasing Address Translation Reach by Leveraging Underutilized Cache Resources
Address translation is a performance bottleneck in data-intensive workloads due to large datasets and irregular access patterns that lead to frequent high-latency page table walks (PTWs). PTWs can be reduced by using (i) large hardware TLBs or (ii) large software-managed TLBs. Unfortunately, both solutions have significant drawbacks: increased access latency, power and area (for hardware TLBs), and costly memory accesses, the need for large contiguous memory blocks, and complex OS modifications (for software-managed TLBs). We present Victima, a new software-transparent mechanism that drastically increases the translation reach of the processor by leveraging the underutilized resources of the cache hierarchy. The key idea of Victima is to repurpose L2 cache blocks to store clusters of TLB entries, thereby providing an additional low-latency and high-capacity component that backs up the last-level TLB and thus reduces PTWs. Victima has two main components. First, a PTW cost predictor (PTW-CP) identifies costly-to-translate addresses based on the frequency and cost of the PTWs they lead to. Leveraging the PTW-CP, Victima uses the valuable cache space only for TLB entries that correspond to costly-to-translate pages, reducing the impact on cached application data. Second, a TLB-aware cache replacement policy prioritizes keeping TLB entries in the cache hierarchy by considering (i) the translation pressure (e.g., last-level TLB miss rate) and (ii) the reuse characteristics of the TLB entries. Our evaluation results show that in native (virtualized) execution environments Victima improves average end-to-end application performance by 7.4% (28.7%) over the baseline four-level radix-tree-based page table design and by 6.2% (20.1%) over a state-of-the-art software-managed TLB, across 11 diverse data-intensive workloads. Victima delivers similar performance as a system that employs an optimistic 128K-entry L2 TLB, while avoiding the associated area and power overheads. Victima (i) is effective in both native and virtualized environments, (ii) is completely transparent to application and system software, (iii) unlike large software-managed TLBs, does not require contiguous physical allocations, (iv) is compatible with modern large page mechanisms and (iv) incurs very small area and power overheads of and , respectively, on a modern high-end CPU. The source code of Victima is freely available at https://github.com/CMU-SAFARI/Victima.
- 28 October 2023
- Published by Association for Computing Machinery (ACM)
- p. 1178-1195
Open AccessVictima: Drastically Increasing Address Translation Reach by Leveraging Underutilized Cache Resources
Address translation is a performance bottleneck in data-intensive workloads due to large datasets and irregular access patterns that lead to frequent high-latency page table walks (PTWs). PTWs can be reduced by using (i) large hardware TLBs or (ii) large software-managed TLBs. Unfortunately, both solutions have significant drawbacks: increased access latency, power and area (for hardware TLBs), and costly memory accesses, the need for large contiguous memory blocks, and complex OS modifications (for software-managed TLBs). We present Victima, a new software-transparent mechanism that drastically increases the translation reach of the processor by leveraging the underutilized resources of the cache hierarchy. The key idea of Victima is to repurpose L2 cache blocks to store clusters of TLB entries, thereby providing an additional low-latency and high-capacity component that backs up the last-level TLB and thus reduces PTWs. Victima has two main components. First, a PTW cost predictor (PTW-CP) identifies costly-to-translate addresses based on the frequency and cost of the PTWs they lead to. Leveraging the PTW-CP, Victima uses the valuable cache space only for TLB entries that correspond to costly-to-translate pages, reducing the impact on cached application data. Second, a TLB-aware cache replacement policy prioritizes keeping TLB entries in the cache hierarchy by considering (i) the translation pressure (e.g., last-level TLB miss rate) and (ii) the reuse characteristics of the TLB entries. Our evaluation results show that in native (virtualized) execution environments Victima improves average end-to-end application performance by 7.4% (28.7%) over the baseline four-level radix-tree-based page table design and by 6.2% (20.1%) over a state-of-the-art software-managed TLB, across 11 diverse data-intensive workloads. Victima delivers similar performance as a system that employs an optimistic 128K-entry L2 TLB, while avoiding the associated area and power overheads. Victima (i) is effective in both native and virtualized environments, (ii) is completely transparent to application and system software, (iii) unlike large software-managed TLBs, does not require contiguous physical allocations, (iv) is compatible with modern large page mechanisms and (iv) incurs very small area and power overheads of and , respectively, on a modern high-end CPU. The source code of Victima is freely available at https://github.com/CMU-SAFARI/Victima.
- Open Access
- 7 March 2022
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Architecture and Code Optimization
- Vol. 19 (2) , 1-28
Dependence-aware Slice Execution to Boost MLP in Slice-out-of-order Cores
Exploiting memory-level parallelism (MLP) is crucial to hide long memory and last-level cache access latencies. While out-of-order (OoO) cores, and techniques building on them, are effective at exploiting MLP, they deliver poor energy efficiency due to their complex and energy-hungry hardware. This work revisits slice-out-of-order (sOoO) cores as an energy-efficient alternative for MLP exploitation. sOoO cores achieve energy efficiency by constructing and executing slices of MLP-generating instructions out-of-order only with respect to the rest of instructions; the slices and the remaining instructions, by themselves, execute in-order. However, we observe that existing sOoO cores miss significant MLP opportunities due to their dependence-oblivious in-order slice execution, which causes dependent slices to frequently block MLP generation. To boost MLP generation, we introduce Freeway, a sOoO core based on a new dependence-aware slice execution policy that tracks dependent slices and keeps them from blocking subsequent independent slices and MLP extraction. The proposed core incurs minimal area and power overheads, yet approaches the MLP benefits of fully OoO cores. Our evaluation shows that Freeway delivers 12% better performance than the state-of-the-art sOoO core and is within 7% of the MLP limits of full OoO execution.
- 7 March 2022
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Architecture and Code Optimization
- Vol. 19 (2) , 1-28
Open AccessDependence-aware Slice Execution to Boost MLP in Slice-out-of-order Cores
Exploiting memory-level parallelism (MLP) is crucial to hide long memory and last-level cache access latencies. While out-of-order (OoO) cores, and techniques building on them, are effective at exploiting MLP, they deliver poor energy efficiency due to their complex and energy-hungry hardware. This work revisits slice-out-of-order (sOoO) cores as an energy-efficient alternative for MLP exploitation. sOoO cores achieve energy efficiency by constructing and executing slices of MLP-generating instructions out-of-order only with respect to the rest of instructions; the slices and the remaining instructions, by themselves, execute in-order. However, we observe that existing sOoO cores miss significant MLP opportunities due to their dependence-oblivious in-order slice execution, which causes dependent slices to frequently block MLP generation. To boost MLP generation, we introduce Freeway, a sOoO core based on a new dependence-aware slice execution policy that tracks dependent slices and keeps them from blocking subsequent independent slices and MLP extraction. The proposed core incurs minimal area and power overheads, yet approaches the MLP benefits of fully OoO cores. Our evaluation shows that Freeway delivers 12% better performance than the state-of-the-art sOoO core and is within 7% of the MLP limits of full OoO execution.
- 3 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Computer Architecture Letters
- Vol. 20 (2) , 134-137
BTB-X: A Storage-Effective BTB Organization
Many contemporary applications feature multi-megabyte instruction footprints that overwhelm the capacity of branch target buffers (BTB) and instruction caches (L1-I), causing frequent front-end stalls that inevitably hurt performance. BTB is crucial for performance as it enables the front-end to accurately resolve the upcoming execution path and steer instruction fetch appropriately. Moreover, it also enables highly effective fetch-directed instruction prefetching that can eliminate many L1-I misses. For these reasons, commercial processors allocate vast amounts of storage capacity to BTBs. This letter aims to reduce BTB storage requirements by optimizing the organization of BTB entries. Our key insight is that today's BTBs store the full target address for each branch, yet the vast majority of dynamic branches have short offsets requiring just a handful of bits to encode. Based on this insight, we organize the BTB as an ensemble of smaller BTBs, each storing offsets within a particular range. Doing so enables a dramatic reduction in storage for target addresses. We also compress tags to reduce the tag storage cost. Our final design, called BTB-X, uses an ensemble of five BTBs with compressed tags that enables it to track 2.8x more branches than a conventional BTB with the same storage budget.
- 3 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Computer Architecture Letters
- Vol. 20 (2) , 134-137
BTB-X: A Storage-Effective BTB Organization
Many contemporary applications feature multi-megabyte instruction footprints that overwhelm the capacity of branch target buffers (BTB) and instruction caches (L1-I), causing frequent front-end stalls that inevitably hurt performance. BTB is crucial for performance as it enables the front-end to accurately resolve the upcoming execution path and steer instruction fetch appropriately. Moreover, it also enables highly effective fetch-directed instruction prefetching that can eliminate many L1-I misses. For these reasons, commercial processors allocate vast amounts of storage capacity to BTBs. This letter aims to reduce BTB storage requirements by optimizing the organization of BTB entries. Our key insight is that today's BTBs store the full target address for each branch, yet the vast majority of dynamic branches have short offsets requiring just a handful of bits to encode. Based on this insight, we organize the BTB as an ensemble of smaller BTBs, each storing offsets within a particular range. Doing so enables a dramatic reduction in storage for target addresses. We also compress tags to reduce the tag storage cost. Our final design, called BTB-X, uses an ensemble of five BTBs with compressed tags that enables it to track 2.8x more branches than a conventional BTB with the same storage budget.
- Open Access
- 30 November 2020
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computer Systems
- Vol. 38 (3-4) , 1-30
Shooting Down the Server Front-End Bottleneck
The front-end bottleneck is a well-established problem in server workloads owing to their deep software stacks and large instruction footprints. Despite years of research into effective L1-I and BTB prefetching, state-of-the-art techniques force a trade-off between metadata storage cost and performance. Temporal Stream prefetchers deliver high performance but require a prohibitive amount of metadata to accommodate the temporal history. Meanwhile, BTB-directed prefetchers incur low cost by using the existing in-core branch prediction structures but fall short on performance due to BTB’s inability to capture the massive control flow working set of server applications. This work overcomes the fundamental limitation of BTB-directed prefetchers, which is capturing a large control flow working set within an affordable BTB storage budget. We re-envision the BTB organization to maximize its control flow coverage by observing that an application’s instruction footprint can be mapped as a combination of its unconditional branch working set and, for each unconditional branch, a spatial encoding of the cache blocks around the branch target. Effectively capturing a map of the application’s instruction footprint in the BTB enables highly effective BTB-directed prefetching that outperforms the state-of-the-art prefetchers by up to 10% for equivalent storage budget.
- 30 November 2020
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computer Systems
- Vol. 38 (3-4) , 1-30
Open AccessShooting Down the Server Front-End Bottleneck
The front-end bottleneck is a well-established problem in server workloads owing to their deep software stacks and large instruction footprints. Despite years of research into effective L1-I and BTB prefetching, state-of-the-art techniques force a trade-off between metadata storage cost and performance. Temporal Stream prefetchers deliver high performance but require a prohibitive amount of metadata to accommodate the temporal history. Meanwhile, BTB-directed prefetchers incur low cost by using the existing in-core branch prediction structures but fall short on performance due to BTB’s inability to capture the massive control flow working set of server applications. This work overcomes the fundamental limitation of BTB-directed prefetchers, which is capturing a large control flow working set within an affordable BTB storage budget. We re-envision the BTB organization to maximize its control flow coverage by observing that an application’s instruction footprint can be mapped as a combination of its unconditional branch working set and, for each unconditional branch, a spatial encoding of the cache blocks around the branch target. Effectively capturing a map of the application’s instruction footprint in the BTB enables highly effective BTB-directed prefetching that outperforms the state-of-the-art prefetchers by up to 10% for equivalent storage budget.
- 1 January 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
A Voice Based One Step Solution for Bulk IoT Device Onboarding
Device Onboarding refers to the first time registration of a device into an Internet of Things (IoT) network. The process can be as simple as pairing of two devices or as complicated as connecting thousands of devices. This process varies significantly across different device vendors. Currently, onboarding process can onboard one device at a time and can take around 5-20 minutes for the lay user. As per International Data Corporation (IDC) [1], number of smart devices is growing faster and approximately 80 billion devices will be connected by 2025. The need for onboarding multiple devices quickly will therefore assume great importance. In this paper, we propose a software-based one-step solution that can securely onboard multiple microphone enabled devices with a single voice command (e.g., “Onboard all devices”). The efficacy of the proposed approach is demonstrated by onboarding multiple IoT devices successfully within the voice range with a single voice command.
- 1 January 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
A Voice Based One Step Solution for Bulk IoT Device Onboarding
Device Onboarding refers to the first time registration of a device into an Internet of Things (IoT) network. The process can be as simple as pairing of two devices or as complicated as connecting thousands of devices. This process varies significantly across different device vendors. Currently, onboarding process can onboard one device at a time and can take around 5-20 minutes for the lay user. As per International Data Corporation (IDC) [1], number of smart devices is growing faster and approximately 80 billion devices will be connected by 2025. The need for onboarding multiple devices quickly will therefore assume great importance. In this paper, we propose a software-based one-step solution that can securely onboard multiple microphone enabled devices with a single voice command (e.g., “Onboard all devices”). The efficacy of the proposed approach is demonstrated by onboarding multiple IoT devices successfully within the voice range with a single voice command.