Embedded Processors
Embedded processor IP: ColdFire and 8051 cores
Put a 32-bit ColdFire core or an 8-bit 8051-compatible core into your SoC. Silvaco licenses this embedded processor IP as synthesizable Verilog, either as a bare core or as a platform already wired to AMBA peripherals.
From 19K gates to over 500 DMIPS
Every ColdFire core in Silvaco’s embedded processor IP runs one instruction set, so the choice
between them comes down to performance and area. The 8051 cores take the control jobs whose
code is already written.
ColdFire V1
Parts 70008 and 70009 · from 19K gates. The smallest, lowest-power ColdFire reaches up to 230 DMIPS at 200 MHz, measured on a typical 90 nm process. The core carries an EMAC, a divider and a cryptographic unit, plus optional trace and real-time debug. Platform 70009 turns it into a configurable microcontroller subsystem on AMBA 3 AHB-Lite.
ColdFire V2
Parts 70019, 70011 and 70010 · from 40K gates. Over 250 DMIPS at 240 MHz, with up to 32 KB of single-cycle local RAM and both AHB master and slave ports. Two platforms add AMBA peripherals. One of them, 70010, adds a 10/100 Fast Ethernet controller and is the same core and peripheral IP that went into NXP’s MCF5208 devices.
ColdFire V4
Parts 70012, 70020, 70021 and 70013. Built for the heaviest code: over 500 DMIPS at 345 MHz. It adds an MMU for a full operating system, and an optional floating-point unit that runs alongside the integer pipeline. Three platforms add AMBA peripherals, and 70013 carries two Fast Ethernet controllers.
8-bit
Parts 70030 and 70029 · M8051EW and M8051W. Both run standard 8051 software but take 2 clocks per machine cycle where the original takes 12. So you get six times the throughput at the same clock, or the same throughput at a sixth of the clock and a fraction of the power. The M8051EW adds on-chip debug over JTAG, with breakpoints and trace.
Find your part number
Replacing a core you already know? Match it by part number.
| PART NUMBER | PRODUCT | FAMILY | WHAT IT IS |
|---|---|---|---|
| 70008 | ColdFire V1 Processor | ColdFire V1 | V1 core carrying an EMAC, divider and cryptographic unit |
| 70009 | ColdFire V1 Processor Platform | ColdFire V1 | V1 core and AMBA peripherals as one subsystem |
| 70019 | ColdFire V2 Processor | ColdFire V2 | Low-gate-count V2 core, over 250 DMIPS |
| 70011 | ColdFire V2SPPC1 | ColdFire V2 | V2 core and AMBA peripherals as one subsystem |
| 70010 | ColdFire V2 SPP5208 Processor Platform | ColdFire V2 | V2 subsystem with a Fast Ethernet controller |
| 70012 | ColdFire V4 Processor | ColdFire V4 | V4 core, over 500 DMIPS |
| 70020 | ColdFire V4 SPP Processor Platform | ColdFire V4 | V4 core and AMBA peripherals as one subsystem |
| 70021 | ColdFire V4 SPPC1 Processor Platform | ColdFire V4 | V4 core and AMBA peripherals as one subsystem |
| 70013 | ColdFire V4 SPPC2 Processor Platform | ColdFire V4 | V4 subsystem with two Fast Ethernet controllers |
| 70030/70029 | M8051EW and M8051W | 8-bit | Low-power 8051-compatible cores, with or without on-chip debug |
Need a core that isn’t listed? Ask an IP specialist.
Four jobs these cores do
Microcontroller subsystems in ASICs
A platform part hands you the core already wired to its bus and peripherals, so integration starts further along.
Legacy design migration and re-spins
Re-spinning a design built on NXP's MCF5208? Platform 70010 is the same ColdFire IP that's inside it.
Low-cost control functions
An 8051-compatible core runs the control code you already have, and drops into idle or power-down when there's nothing to do.
Designs already built on these architectures
HCS08 code has a migration path to ColdFire V1. From there, your code moves up to V2 and V4.
What's behind the architecture
Your processor choice will outlive the chip it first ships in. These cores came from Tier 1
semiconductor companies, and every part has its own brief.
11
part numbers across ColdFire cores, ColdFire platforms and 8051 cores
230 to 500+ DMIPS
from ColdFire V1 up to V4
Over 500 million
devices worldwide built on the ColdFire architecture
Verilog source
with an integration testbench, tests, documentation and simulation and synthesis scripts
Product brief · PDF
Summer 2026
ColdFire V1 Core
The pipelines and BDM debug of the smallest ColdFire, with its deliverables list.
Product brief · PDF
March 2026
ColdFire V2 Core
Local RAM and bus ports on the 250 DMIPS core.
Product brief · PDF
March 2026
ColdFire V4 SPPC2 Processor Platform
Two Fast Ethernet controllers on a 345 MHz V4 platform, with FPU and MMU.
Product brief · PDF
March 2026
M8051EW and M8051W
How the two cores differ on debug, and what each power mode shuts off.
Which core does your SoC need?
Send an IP specialist the workload and your target foundry and node. What comes
back: the core that fits, how licensing and evaluation work, and whether it’s available on
your process. If the core you want isn’t listed, ask about that too.
What SoC teams ask about embedded processor IP
What embedded processors does Silvaco offer?
Two families. The 32-bit ColdFire line spans three generations, V1, V2 and V4. Each comes as a bare core and as platforms with AMBA peripherals, two of them with Fast Ethernet. The 8-bit line is the M8051EW and M8051W, both 8051-compatible. That's eleven part numbers, each licensed as synthesizable Verilog.
What is ColdFire?
ColdFire is a 32-bit RISC architecture with variable-length instructions of 16, 32 or 48 bits, which keeps code compact in memory. Every ColdFire core shares one architecture and one instruction set, so code carries across the family. Over 500 million devices use it, with development tools from NXP, Green Hills, Wind River and GNU.
Are these cores still supported?
Silvaco licenses all eleven parts, and each ships as Verilog source with an integration testbench, tests, documentation, and scripts for simulation and synthesis. The architecture's software ecosystem is still in place: compilers, debuggers, uClinux and RTOS options such as MQX. For support terms on your own program, talk to an IP specialist.
What performance do the cores deliver?
ColdFire V1 reaches up to 230 DMIPS at 200 MHz on a typical 90 nm process, from as few as 19K gates. V2 passes 250 DMIPS at 240 MHz from 40K gates, and V4 passes 500 DMIPS at 345 MHz. On the 8-bit side, the M8051EW and M8051W execute six times faster than a standard 8051 at the same clock.
How do they integrate with AMBA?
The V2 and V4 cores sit on the AMBA 2 AHB system bus through a master port. A slave port lets another AHB master reach the core's local RAM. V1 also uses AMBA 2 AHB, and its platform uses AMBA 3 AHB-Lite. Platforms arrive joined to AMBA peripherals, and Silvaco licenses the AMBA fabric around them too.
Is an 8-bit core still appropriate?
Often, yes, when the job is control and the software already exists for the 8051. The M8051EW and M8051W cut the machine cycle from 12 clocks to 2 and run standard 8051 code unchanged. You can clock them at a sixth of the usual rate for the same work, and idle and power-down modes save more.
Related IP
AMBA IP cores and subsystems
Bus fabrics and peripherals for the rest of the SoC
Other IP
More licensable blocks from Silvaco
Automotive IP
Controllers for CAN, FlexRay and LIN buses