00

SGen

Tool ☆ WATCH ⎘ SHARE

First seen 6/16/2026

Last seen 6/16/2026

Evidence 9 chunks

01

NEIGHBORHOOD

28 nodes · 43 edges

graph · SGen · depth=1

02

RELATIONSHIPS

29 connections

reg_helper class uses → 100% 2e evidence

SGen uses the reg_helper class to randomize register usage.

reg_init_seq uses → 100% 2e evidence

SGen uses reg_init_seq to generate preamble code that initializes registers.

bare_sequence class uses → 100% 2e evidence

SGen uses the bare_sequence class for creating sequences on-the-fly.

lambda functions uses → 100% 2e evidence

SGen uses lambda functions extensively for randomization and user-defined overrides.

simd_factory uses → 100% 2e evidence

SGen uses the simd_factory to create SIMD instruction instances.

SIMD instructions uses → 100% 2e evidence

SGen can generate SIMD instructions using the simd factory.

Cavium ← uses 100% 2e evidence

Cavium uses SGen for verifying its ARM server cores.

sequence-based stimulus generation implements → 100% 2e evidence

SGen implements the sequence-based approach to stimulus generation.

exerciser run evaluates → 100% 2e evidence

SGen performance is evaluated through exerciser runs.

UVM ← compares with 90% 2e evidence

SGen is positioned as an alternative to UVM for micro-processor verification stimulus generation.

RTL verification evaluates → 100% 2e evidence

SGen is used to verify RTL designs and catches RTL bugs.

PPIGen ← compares with 100% 2e evidence

SGen is compared to PPIGen in terms of failure detection during exerciser runs.

weighted set uses → 100% 2e evidence

SGen uses a weighted set class to enable random picking from a set of objects.

Directed Test Generation uses → 100% 2e evidence

SGen piggybacks onto PPIGen's directed mode to generate tests.

random stimulus generation uses → 100% 2e evidence

SGen supports generating random exerciser tests.

processor instruction stream generation uses → 100% 2e evidence

SGen generates instruction streams as its primary output.

wset class uses → 100% 2e evidence

SGen uses the wset class for weighted random selection.

C++11 uses → 100% 2e evidence

SGen is developed using C++11 and leverages its new features.

SystemVerilog ← compares with 85% 1e evidence

SGen uses C++ instead of SystemVerilog for stimulus generation.

assembly generator implements → 100% 1e evidence

SGen is implemented as a sequence-based assembly generator.

factory pattern uses → 100% 1e evidence

SGen uses factories to create instances of registered types, similar to UVM.

polymorphic function wrapper uses → 100% 1e evidence

SGen uses the polymorphic function wrapper to store references to lambda functions.

regular expressions uses → 80% 1e evidence

SGen can use regular expressions available in C++11 standard library.

auto specifier uses → 90% 1e evidence

SGen code uses the auto specifier as shown in working code examples.

PPIGen depends on → 100% 1e evidence

SGen produces assembly and config files that are fed through PPIGen's directed flow.

rand_intf class uses → 100% 1e evidence

SGen provides the rand_intf class as a random interface for sequences.

arm_factory uses → 100% 1e evidence

SGen uses the arm_factory to instantiate ARM instruction instances.

factory pattern implements → 100% 1e evidence

SGen implements a hierarchy of factories mirroring the class hierarchy.

Micro-processor verification using a C++11 sequence-based stimulus engine ← introduces 100% 1e evidence

The paper introduces and presents SGen, a sequence-based assembly generator.