An Efficient LBlock Lightweight Block Cipher Coprocessor on RISC-V: Combinational Key Schedule Fusion and S-Box-to-LUT Mapping

Resource-constrained Internet-of-Things (IoT) terminals require encryption engines that combine low silicon cost with adequate throughput, a balance that is hard to reach with general-purpose software alone. This paper presents an LBlock lightweight block cipher coprocessor tightly coupled to an open-source RISC-V (Hummingbird E203) core through the NICE custom-instruction interface. LBlock serves here as a compact Feistel-cipher benchmark targeting legacy and low-volume IoT deployments rather than as a substitute for newer standards such as ASCON. We exploit two structural properties of LBlock: first, its 4-bit S-boxes map naturally onto the six-input look-up tables (LUTs) of modern FPGAs, so the entire substitution layer is realized as eight parallel single-LUT-depth tables instead of multi-cycle table lookups. Second, the LBlock key schedule is a one-way feedback-free recurrence, which lets us refactor key expansion from an independent multi-cycle sequential module into a pure combinational function that is fused with the round function and executed in the same clock cycle. The resulting encryption core performs one round per cycle, reducing core-only single-block latency from 226 cycles in the baseline implementation to 34 cycles, while the complete NICE coprocessor operation requires 85 cycles including data movement, instruction issue, computation, and write-back. The design is described in Chisel and integrated as a coprocessor with three custom instructions. On an FPGA-based SoC, the coprocessor produces outputs identical to the LBlock test vectors and accelerates encryption by 121.45× over a software baseline on the same core, while the encryption core reaches 321.7 MHz (643 Mbps) on Artix-7 and up to 472.2 MHz (944 Mbps) on Virtex-7 while occupying only 187 LUTs, as validated across three FPGA families. These results show that matching the algorithmic symmetry of LBlock to the underlying hardware fabric yields a lightweight and low-overhead cryptographic accelerator suitable for RISC-V IoT endpoints.

Authors

Institutions

Publication Details

Journal
Symmetry
Published
2026-07-22
DOI
https://doi.org/10.3390/sym18071239
Primary Topic
Cryptographic Implementations and Security
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

An Efficient LBlock Lightweight Block Cipher Coprocessor on RISC-V: Combinational Key Schedule Fusion and S-Box-to-LUT Mapping

L L Zhang, Z H Wang, Cheng Qian, Yi-Xiang Wang et al.
Symmetry
Cryptographic Implementations and Security
article

An Efficient LBlock Lightweight Block Cipher Coprocessor on RISC-V: Combinational Key Schedule Fusion and S-Box-to-LUT Mapping

L L Zhang, Z H Wang, Cheng Qian, Yi-Xiang Wang, Kaibo Sun, Maosheng He, Runze Zhou, Chaoen Xiao
article en

Abstract

Resource-constrained Internet-of-Things (IoT) terminals require encryption engines that combine low silicon cost with adequate throughput, a balance that is hard to reach with general-purpose software alone. This paper presents an LBlock lightweight block cipher coprocessor tightly coupled to an open-source RISC-V (Hummingbird E203) core through the NICE custom-instruction interface. LBlock serves here as a compact Feistel-cipher benchmark targeting legacy and low-volume IoT deployments rather than as a substitute for newer standards such as ASCON. We exploit two structural properties of LBlock: first, its 4-bit S-boxes map naturally onto the six-input look-up tables (LUTs) of modern FPGAs, so the entire substitution layer is realized as eight parallel single-LUT-depth tables instead of multi-cycle table lookups. Second, the LBlock key schedule is a one-way feedback-free recurrence, which lets us refactor key expansion from an independent multi-cycle sequential module into a pure combinational function that is fused with the round function and executed in the same clock cycle. The resulting encryption core performs one round per cycle, reducing core-only single-block latency from 226 cycles in the baseline implementation to 34 cycles, while the complete NICE coprocessor operation requires 85 cycles including data movement, instruction issue, computation, and write-back. The design is described in Chisel and integrated as a coprocessor with three custom instructions. On an FPGA-based SoC, the coprocessor produces outputs identical to the LBlock test vectors and accelerates encryption by 121.45× over a software baseline on the same core, while the encryption core reaches 321.7 MHz (643 Mbps) on Artix-7 and up to 472.2 MHz (944 Mbps) on Virtex-7 while occupying only 187 LUTs, as validated across three FPGA families. These results show that matching the algorithmic symmetry of LBlock to the underlying hardware fabric yields a lightweight and low-overhead cryptographic accelerator suitable for RISC-V IoT endpoints.

SymmetryVol. 18(7)
Beijing Electronic Science and Technology Institute (CN)
Openalex Percentile: Top 6%
Cryptographic Implementations and Security
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.