The MIPS M5150/M5100 processor cores are ideal for IoT, wearable and other embedded and real time applications.

The M51xx cores are superset extensions of the MIPS microAptiv family, featuring:

  • A five stage pipeline architecture delivering class-leading performance
  • The microMIPS code size reduction ISA
  • The MIPS DSP r2 module
  • Fast interrupt handling
  • Advanced debug/profiling capabilities
  • Comprehensive power management

The MIPS M51xx processors are fully OmniShield ready, providing the highest level of security of any embedded microcontroller architecture.

The M51xx cores implement the latest MIPS Release 5 architecture (MIPS32r5), incorporating hardware virtualization for increased security, IP protection and reliability for a wide range of applications including: industrial controllers, Internet of Things (IoT), wearables, wireless communications, automotive and storage.

Virtualization support is provided for both the M5100 microcontroller-class and M5150 embedded processor versions of the M51xx family, a feature not available in competing alternatives.

Additional security is provided by the ‘anti-tamper’ feature in these cores, which includes countermeasures that provide resistance to unwanted access to the processor.

An optional IEEE 754 Floating Point Unit provides high-performance support of both single and double precision instructions.

The M51xx family includes the following processor cores:

MIPS M5100

The MIPS M5100 features an integrated SRAM controller and a real time execution unit, optimized for low cost, low power microcontroller applications.

mips-m5100

MIPS M5150

The MIPS M5150 incorporates a high performance L1 cache controller and virtual memory management support for high performance embedded system applications and rich operating systems.

mips-m5150

MIPS Series5 Warrior M-class cores offer high performance, scalable and trusted solutions for a wide range of embedded applications.

The Warrior M-class CPUs deliver class-leading performance and breakthrough hardware virtualization in the M5100 microcontroller class (MCU) and M5150 embedded processor cores.

The M51xx cores enable a broad range of systems to benefit from high levels of security, reliability and flexibility in a low-power, compact form factor.

The M51xx cores are enhancements of the MIPS microAptiv family designed with the same five-stage pipeline architecture, utilizing the high-performance digital signal processing (DSP) capabilities along with the microMIPS ISA for code size reduction

Support is provided by a comprehensive set of hardware and software development tools from Imagination as well as a growing ecosystem of partners.

Several hypervisors for the M-class cores are available and under development from Imagination and leading third party developers.

This includes open source hypervisors such as KVM, the Kernel-based Virtual Machine, as well as others, including microkernel-type hypervisor such as SELTECH FEXEROX or the open source prplHypervisor.

Documentation

Applications

  • Industrial control and automation
  • Internet of Things (IoT), Machine to Machine (M2M)
  • Wearables
  • Home appliances, digital consumer products
  • Automotive
  • Cloud computing
  • Network communications
  • Storage

Benefits

  • Standard architecture, proven in millions of SoC designs
  • High-performance, area- and energy- efficient architecture: performance requirement achieved at lower frequency and smaller size than the competition
  • Hardware virtualization – supports multiple software environments running independently, securely, efficiently and in complete isolation to each other
  • Available in microcontroller and embedded processor versions for use in a wide range of operating environments
  • Combined MCU and DSP technology for cost-effective signal processing
  • Flexibility and scalability – single design to cover a broad range of applications
  • FPU to accelerate real-time control in industrial, automotive and digital consumer applications
  • Broad software and ecosystem support, and mature toolchain
  • Available as synthesizable IP for implementation in any process node, with standard cells and memories

Architecture

  • MIPS32 Release 5 Architecture
  • microMIPS ISA
    • Enhanced code compression ISA of combined 16- and 32-bit instructions
    • Supports all existing MIPS32 instructions; adds new 16- and 32-bit instructions
    • Enables up to 30% code size reduction relative to 32-bit only code

Hardware Virtualization

  • Create multiple execution environments (Guest) isolated from each other, operating at kernel privilege level
  • Hypervisor/Secure Monitor (Root) manages access rights for each Guest
    • Supports Type 1 and Type II hypervisors
  • Supports up to 7 Guests, each supplied a unique ID; Guest OS runs un-modified
  • 7 new instructions facilitate Root-to-Guest communication
  • Supports multiple Memory Management Unit options for optimum area vs. functionality
    • M5100 – FMT + Root Protection Unit
    • M5150 – Guest TLB + Root TLB
    • M5150 – Guest TLB + Root Protection Unit
  • Allows sharing of resources (memory, DSP, FPU etc.) between Guests

DSP Module r2

  • Dedicated pipeline, operates in parallel with core integer pipeline
  • Implements over 150 instructions, including 70 SIMD and 38 Multiply/MAC instructions
  • Enhanced Multiply & Divide Unit
    • Single cycle throughput multiply and MAC operations
    • Supports 32×32, 16×16, dual 16×16, dual 8×8, dual 8×16
  • Supports up to 4 Accumulators

Floating Point Unit (FPU)

  • Single and double precision IEEE 754 compliant FPU
  • Supports IEEE-754 2008 Nan and ABS instructions
  • Dedicated 7-stage pipeline, operating in parallel with core integer pipeline
  • Most instructions execute with 1 cycle throughput and 4 cycle latency
  • Executes 1:1 Core:FPU clock ratio
  • Supports both MIPS32 and microMIPS instructions

Anti-Tamper

  • Injection of random pipeline stalls
  • Cache/SPRAM address and data scrambling
  • 2 pseudo random number generators for use by the user software and core logic

Memory Controller

  • M5150 – L1 cache controller for Instruction and Data sizes up to 64KB, 4-way set associative
  • M5100 – 32-bit address and data SRAM interface, separate or unified instruction and data interface

Bus Interface Unit

  • AMBA 3 AHB

EJTAG Debug & Trace

  • Secure debug feature – prevents streaming instructions through the EJTAG port
  • Supports enhanced iFlowtrace™ with additional event trace modes
  • Simple/Complex instruction and data breakpoint support – 2I/1D, 4I/2D, 6I/2D, 8I/4D
  • Support for 2 Performance Counters with multiple event type options
  • Instruction and data address sampling: zero overhead, qualified read/write
  • Support for 2-wire cJTAG debug interface

Power Management

  • Incorporates extensive fine-grain clock gating
  • Implements a Power Down mode initiated by a WAIT instruction

Expandability

  • Optional co-processor (COP2) and CorExtend™ / User Defined Instruction (UDI) interfaces

M5100 Core Specifications

Process Node65LP28HPM
Optimization****SpeedAreaSpeedArea
Frequency* (MHz)322100497100
Performance (DMIPS)505157780157
Area** (mm2)0.770.20.230.04
Core Active Power*** (mW/MHz)0.110.050.040.017
Sleep Power (µW/MHz)5.82.31.70.8
Cell Library9T LVt9T SVt12T SVt9T LVt

M5150 Core Specifications

Process Node65LP28HPM
Optimization****SpeedSpeed
Frequency* (MHz)372576
Performance (DMIPS)584904
Area** (mm2)0.890.26
Core Active Power*** (mW/MHz)0.130.07
Sleep Power (µW/MHz)2.71.5
Cell Library9T LVt12T SVt

Notes: Frequency, power consumption and size depend upon configuration options, synthesis, silicon vendor, process

* Production frequency PTSI
65LP:  +/- 5% OCV, 50ps clock jitter
28HPM: +/- 4% OCV, 25ps clock jitter

** Core Floorplanned area
*** Power measured at typical corner, 25C at 1.2V/65LP, 0.9V/28HPM
**** M5100 Speed Optimized – microMIPS + FPU + DSP + 32KB/32KB I/D SRAM, FMT+32RPU + AHB
**** M5100 Area Optimized – microMIPS + DSP +8RPU
Area optimized frequency can be higher/lower  than the target chosen
**** M5150 Speed Optimized – microMIPS + FPU + DSP + 32KB/32KB I$/D$ + 32RTLB/32GTLB + AHB