Take advantage of the performance standard for 32-bit mid-range cores in the embedded industry while minimizing design time and reducing product costs. Tailored SoC design methodologies, an Open Core Protocol (OCP) interconnect structure, standard libraries and on-chip memories from industry-leading companies ensure that products based on 24K cores are brought to marketwith speed, ease and efficiency.
Designed to power through graphics, Java and demanding code and with features like an ultrafast multiply, intelligent caches, floating point support and the CorExtend™ capability - which allows users to supercharge application performance by defining and adding their own instructions - the 24K family is the ideal solution for digital and interactive television, set-top boxes, DVD and other performance-driven applications.
MIPS32® 24K® Processor Core Family Software User's Manual v3.11
MIPS32® 24Kc™ Processor Core Datasheet v4.00
MIPS32® 24Kf™ Processor Core Datasheet v4.00
Programming the MIPS32® 24K® Core Family v4.63
The MIPS32® 24KE™ Core Family: High-Performance RISC Cores with DSP Enhancements (PDF)
Working with ScratchPad RAMs for MIPS32® 24K® and 34K® Cores
MIPS32 24K Summary
- With an 8-stage pipeline and a maximum clock frequency exceeding 1400 MHz in 40nm, the 24K family of cores enable SoC designers to reduce product costs and speed time-to-market by giving them the performance headroom to implement more features now and upgrades in the future with software flexibility rather than rigid, fixed hardware.
- Cadence, Synopsys, Magma and other EDA industry leaders help minimize design time and offer a proven path to silicon by co-developing tailored SoC design methodologies. This couples the high-performance, low-power 24K cores with cutting-edge core hardening technologies.
- By standardizing the core interface on OCP (www.ocpip.org), the 24K cores accelerate time-to-market by enabling easy reuse of standard SoC IP. Memory controllers, bus interconnects and other standardized peripherals are now easily integrated through common on-chip interfaces.
- The highly-scalable 24K microarchitecture implements the industry-standard MIPS32 Release 2 architecture, which includes features such as enhanced bit-field manipulation, reduced interrupt latency and enhanced cache control.
- A rich environment of third-party tools and software support the 24K family of cores.
MIPS32 24K Features
32-bit MIPS32® architecture
- 8-stage pipeline
- 32-bit address
- Vectored interrupts and support for external interrupt controller
- GPR shadow registers (optionally, one or three additional shadows can be added to minimize latency for interrupt handlers)
Floating Point Unit (FPU)
- Floating point version of core available
- IEEE std 754 compliant, supporting single and double precision calculations
- Contains 32 64-bit registers for more operations with less load/store overhead
DSP ASE Instructions
- 8-,16-and 32-bit SIMD instructions
- Saturating and fractional math
- Popular DSP operations, such as MAC, dot-product, absolute and complex-multiply
- Key features such as variable bit insert/extract and virtual circular buffers, complex multiply
MIPS16e™ Code Compression
- 16-bit encoding of 32-bit instructions to improve code density
- Special PC-relative instructions for efficient loading of addresses and constants
- SAVE & RESTORE macro instructions for setting up and tearing down stack frames within subroutines
Programmable Cache Size
- Individually configurable instruction and data caches, sizes of 0KB, 8KB, 16KB, 32KB, and 64KB
- 4-way set-associative
- Up to eight outstanding load misses
- Write-back and write-through support
- 32-byte cache line size
Scratchpad SPRAM Support
- Independent cache configuration
- 64b OCP 2.1 memory interface
- Can support arrays up to 1MB
- Separate RAMs for instruction and data
Memory Management Unit (MMU)
- 4-entry instruction, 8-entry data micro-TLBs
- Configurable 16/32/64 dual-entry joint TLB with variable page sizes
- Optional fixed mapping translation (FMT) for applications not requiring address mapping or protection
Bus Interface Unit (BIU)
- Implements the Open Core Protocol (OCP) Release 2.1
- 64-bit read and write data buses to efficiently transfer data between memory and L1 caches
- Supports a variety of CPU-to-system bus clock ratios to give greater flexibility for system implementation
- 4-entry write buffer
Integer Multiply/Divide Unit (MDU)
- Fully pipelined single-cycle repeat rate for 32x32 MAC instructions
- Minimum frequency: 0 MHz
- Power-down (sleep) mode (triggered by the software WAIT instruction)
- Support for software-controlled clock divider
- Fine-grain clock gating
- Performance counters
- Support for single stepping
- Virtual instruction and data address breakpoints
- PC and data address and value tracing with trace compression
General Purpose Coprocessor (COP2) Interface
- 64-bit interface to a user defined coprocessor
MIPS32 24K Specifications
|Process||40nm G TSMC|
||1468 (worst case)|
|Performance||1.6 (DMIPS/MHz) |
|Core Area (mm2)||0.36 (core only, extracted from fully layed out GDSII database)|
Frequency and size depend upon configuration options, synthesis, silicon vendor, process and cell libraries.
* DSP ASE available on 24KE core
24Kc™ & 24KEc Cores
Include a high-performance 32x32 multiply/divide unit and configurable MMU with TLB or fixed mapping.
24Kf™ and 24KEf Cores
Add hardware floating point support that is fully compliant with IEEE 754.
24K® and 24KE Pro Cores
4Kc/Kf Pro and 24Ec/KEf Pro cores feature the CorExtend™ capability.
All 24KE variants (24KEc, 24KEf, 24KEc/KEf Pro) include support for the DSP ASE.