Skip to content
STIMSMITH

Verification of a RISC-V system with multiple cores

Paper
First seen 6/18/2026
Last seen 6/18/2026
Evidence 4 chunks

NEIGHBORHOOD

No graph connections found for this entity yet. It may appear in future ingestion runs.

explore full graph →

RELATIONSHIPS

34 connections
Spike (ISS) uses → 100% 3e
A modified version of Spike ISS is used as the reference model for co-simulation.
AXI interface uses → 100% 2e
AXI interfaces are used in the CPU-Subsystem for memory and peripheral connections.
Core-UVM environment introduces → 100% 2e
The paper introduces a standalone UVM environment at the core-level called Core-UVM.
CPU-Subsystem UVM environment introduces → 100% 2e
The paper introduces a CPU-Subsystem UVM environment at the system level.
Regression Testing uses → 100% 2e
Regression testing is used to detect buggy updates in the design.
Co-simulation uses → 100% 2e
The verification approach employs co-simulation with the Spike ISS to check design correctness.
Reference Model uses → 100% 2e
The paper relies on Spike as a reference model for validating RISC-V designs.
UVM uses → 100% 2e
The paper presents a verification strategy based on the Universal Verification Methodology (UVM).
random binary generation uses → 100% 2e
Random binary generation is used via riscv-dv for test generation.
Directed Tests uses → 100% 2e
Directed tests are used alongside random tests in the verification strategy.
DMA interface verification uses → 100% 2e
UVM directed tests are used for verifying the DMA interface.
riscv-dv uses → 100% 2e
riscv-dv is used to generate random binaries for testing.
interrupt generation uses → 100% 2e
Random interrupt generation is employed to test interrupt handling in the design.
riscv-tests uses → 100% 2e
riscv-tests is used as a source of directed tests.
vector memory instructions uses → 100% 1e
Vector memory instructions are generated with changing SEW and vector length values.
atomic instructions stress tests uses → 100% 1e
Atomic instruction stress tests are included to focus on randomized atomic instructions.
register randomization uses → 100% 1e
Register randomization for key registers is added as a feature to riscv-dv.
JTAG interface verification uses → 100% 1e
UVM directed tests are used for verifying the JTAG interface.
exception and trap handlers uses → 90% 1e
Issues in exception and trap handlers were fixed in riscv-dv for accurate test generation.
Oscar Palomar authored by → 100% 1e
Oscar Palomar is listed as first author of the paper.
Barcelona Zettascale Lab mentions → 100% 1e
The paper is developed as part of the Barcelona Zettascale Lab project.
Roberto Ignacio Genovese authored by → 100% 1e
Roberto Ignacio Genovese is listed as a co-author of the paper.
Iván Díaz Ortega authored by → 100% 1e
Iván Díaz Ortega is listed as a co-author of the paper.
Albert Aguilera authored by → 100% 1e
Albert Aguilera is listed as a co-author of the paper.
Abdul Rauf authored by → 100% 1e
Abdul Rauf is listed as a co-author of the paper, affiliated with ARM.
Barcelona Supercomputing Center authored by → 100% 1e
The majority of authors are affiliated with Barcelona Supercomputing Center.
Functional Coverage uses → 90% 1e
Functional coverage is a key challenge addressed in the paper.
SVA assertions uses → 100% 1e
SVA assertions are implemented in selected module interfaces.
Scoreboard uses → 100% 1e
A scoreboard is used to detect mismatches and signal test failures.
SystemVerilog DPI uses → 100% 1e
SystemVerilog DPI calls are implemented to enable interaction between Spike and the UVM environment.
AXI Master VIP uses → 100% 1e
AXI Master VIP is used to verify the DMA bridge in the subsystem.
bootrom preloading uses → 100% 1e
Spike is modified to support bootrom preloading from an .elf file for flexible environment usage.
RISC-V compliance tests uses → 100% 1e
RISC-V compliance tests are used as directed tests in the verification flow.
RVV 1.0.0 vector instructions uses → 100% 1e
Full support for RVV 1.0.0 vector instructions is added to riscv-dv for testing.