Embedded Software Development and Simulation of Processors

If you are developing software for a processor where you do not have access to the hardware or you need a better embedded software development environment – then you probably need to have a look at using a simulator to develop your software on.

The key component in a simulator used for developing embedded software is the CPU model.

This site provides information on the industry’s most comprehensive library of extremely fast and efficient CPU Models of advanced processor cores that work in a variety of simulation environments. The whole focus of these models is to enable you to develop embedded software in a more efficient way, with less bugs, and in less time.

Fast CPU Models

These Fast Processor/CPU Models can be used in C, C++, or SystemC TLM based platforms which you can develop or you can use existing platform models (virtual platforms) available from several sources (e.g. OVP, Imperas). Readily available virtual platform models range from simple bare metal models through to full development board models such as the MIPS Malta or ARM Versatile Express.

All the models have been developed in C using OVP technology and for SystemC TLM have been tested to run with all major SystemC simulators: Cadence, Synopsys, Mentor, Carbon, Accellera/OSCI. The models have also been tested with emulators from Synopsys ZeBu, Cadence Palladium, and Aldec. The models run on both Windows and Linux host platforms. Native OVP simulators (use C platforms) are available from Imperas and OVP.

On this site you will see the scope and variety of the Fast CPU Models available and how easy they are to download and use in C or C++ simulations.

Several companies have downloaded these models and use them within their own internal simulation environments. There are specific APIs to easily allow simulator integration and encapsulation. Cadence working with Imperas is one example.

Largest CPU Model Library in the Industry

Each Fast Processor Model is written in C using the Open Virtual Platforms (OVP) standard public APIs. They include a dedicated native C++ SystemC TLM2 interface provided as source to enable understanding and easy usage. Not only is the specific SystemC TLM2 interface provided as source (click to preview an example), also, the full model is available as source. The models do require a simulator that implements the OVP APIs – such as OVPsim available from OVP, or commercial simulators from companies such as Imperas Software.

There is documentation that explains about the models in general (click to preview) and for each model there is a specific document (click to preview the document for the ARMv8 Cortex-A57MPx4 model) that describes what is available in the model, for example its ports, nets, registers, modes, exceptions, and other configuration/parameter options. On the OVP website there is lot of information about each model (for example click to browse the available information on the ARMv8 Cortex-A57MPx4).

An overview document (click to preview) explains, with the use of examples, how the models are configured and used in SystemC TLM2 platforms.

In a C or C++/SystemC TLM2 environment, the models are used directly, with no inefficient co-simulation. It is very simple to create homogeneous or heterogenous platforms of advanced processor core models. To see examples of platforms ranging from one to twenty-four cores and for platforms that boot full operating systems like Linux and Android, including SMP, visit the the examples and platforms available from the OVP platforms download area or video area.

Many models can be instanced in one platform, virtual platform or virtual system prototype – it is easy to build multi-core multi-processor platforms.

Faster Models means BUGS ARE FOUND SOONER

The models run fast, hundreds of millions of instructions per second (MIPS):

OVP Fast CPU Models run fast

If you need maximum available simulation speed from the Fast CPU Models, then you need to find our more about QuantumLeap from Imperas. This uses the parallel resources of the host PC to accelerate your simulations.

QuantumLeap from Imperas uses host resources to accelerate simulation throughput

For more information on QuantumLeap parallel simulation acceleration using host resources and to find out how to develop your embedded software at the fastest speeds in the industry, browse the Imperas information.

Fastest Simulation of ARM and MIPS cores

If you want to see a video – click here for the fastest ARM model simulation, or for the Imagination MIPS use of QuantumLeap click here.

Industry Standard Debug and IDE

Each model supports standard debugging interfaces and can be connected using RSP to GDB, either standalone or within an Eclipse IDE environment. The models also connect to the advanced multi-core debugger available as part of the Imperas Advanced Multicore Software Development Kit product.

Eclipse GDB Debug

Easy to use – watch the video

To see a short video of a Fast CPU Model running in a simple platform – and see it booting to the Linux prompt in under 10 seconds, click the image:

Nios booting Linux Video

If you want to see other videos, OVP has a collection to view here.

More Information

At the top of this page are several menu picks that list the different families and enable access to the model specific information. The listed items on the right provide news related information.

Getting Started

To explore how easy it is to use these Fast CPU Models, look at the OVP starting page.

If you are looking for products to use to develop embedded software visit the Imperas Software website.

Thank you for your interest – the Open Virtual Platforms Fast Processor Models team. To contact us please visit Imperas or OVP.

Currently available Fast Processor Model Families.

FamilyModel Variant
POWER Models    POWER Models aliases mpc82x UISA m476 m470 m460 m440 (aliases)
Renesas Models    Renesas Models aliases V850 V850E1 V850E1F V850ES V850E2 V850E2M V850E2R RH850G3M m16c r8c RL78-S1 RL78-S2 RL78-S3 (aliases)
RISC-V Models    RISC-V Models aliases RV32I RV32G RV64I RV64G (aliases)
MIPS Models    MIPS Models aliases ISA M14K M14KcTLB M14KcFMM 4KEc 4KEm 4KEp M4K 4Kc 4Km 4Kp 24Kc 24Kf 24KEc 24KEf 34Kc 34Kf 34Kn 74Kc 74Kf 1004Kc 1004Kf 1074Kc 1074Kf microAptivC microAptivP microAptivCF interAptiv interAptivUP proAptiv 5Kf 5Kc 5KEf 5KEc M5100 M5150 M6200 M6250 MIPS32R6 P5600 P6600 I6400 MIPS64R6 (aliases)
ARM Models    ARM Models aliases ARMv4T ARMv4xM ARMv4 ARMv4TxM ARMv5xM ARMv5 ARMv5TxM ARMv5T ARMv5TExP ARMv5TE ARMv5TEJ ARMv6 ARMv6K ARMv6T2 ARMv6KZ ARMv7 ARM7TDMI ARM7EJ-S ARM720T ARM920T ARM922T ARM926EJ-S ARM940T ARM946E ARM966E ARM968E-S ARM1020E ARM1022E ARM1026EJ-S ARM1136J-S ARM1156T2-S ARM1176JZ-S Cortex-R4 Cortex-R4F Cortex-A5UP Cortex-A5MPx1 Cortex-A5MPx2 Cortex-A5MPx3 Cortex-A5MPx4 Cortex-A8 Cortex-A9UP Cortex-A9MPx1 Cortex-A9MPx2 Cortex-A9MPx3 Cortex-A9MPx4 Cortex-A7UP Cortex-A7MPx1 Cortex-A7MPx2 Cortex-A7MPx3 Cortex-A7MPx4 Cortex-A15UP Cortex-A15MPx1 Cortex-A15MPx2 Cortex-A15MPx3 Cortex-A15MPx4 Cortex-A17MPx1 Cortex-A17MPx2 Cortex-A17MPx3 Cortex-A17MPx4 AArch32 AArch64 Cortex-A53MPx1 Cortex-A53MPx2 Cortex-A53MPx3 Cortex-A53MPx4 Cortex-A57MPx1 Cortex-A57MPx2 Cortex-A57MPx3 Cortex-A57MPx4 Cortex-A72MPx1 Cortex-A72MPx2 Cortex-A72MPx3 Cortex-A72MPx4 MultiCluster ARMv6-M ARMv7-M Cortex-M0 Cortex-M0plus Cortex-M1 Cortex-M3 Cortex-M4 Cortex-M4F (aliases)
Other Models    Other Models aliases Synopsys ARC_600 Synopsys ARC_605 Synopsys ARC_700 Synopsys ARC_0x21 Synopsys ARC_0x22 Synopsys ARC_0x31 Synopsys ARC_0x32 openCores_generic Xilinx MicroBlaze_V7_00 Xilinx MicroBlaze_V7_10 Xilinx MicroBlaze_V7_20 Xilinx MicroBlaze_V7_30 Xilinx MicroBlaze_V8_00 Xilinx MicroBlaze_V8_10 Xilinx MicroBlaze_V8_20 Xilinx MicroBlaze_V9_50 Xilinx MicroBlaze_V10_00 Xilinx MicroBlaze_ISA Altera Nios II_Nios_II_F Altera Nios II_Nios_II_S Altera Nios II_Nios_II_E (aliases)