|
|
Related Documentation:
Digital Signal Processing Solutions - TMS320C27x
TI Announces New DSP Architecture That Renders MCUs Obsolete
TI Simplifies Programming with Extremely Efficient C Compiler, Speeds Up Debug with 100-MHz Real-Time Emulation
TMS320C27x Architecture Feature Summary
TI's Groundbreaking TMS320C27x Architecture Balances the Best Features of DSPs and MCUs
New TMS320C27x Architecture Unites the Best Features of DSPs and Microcontrollers for Real-Time Embedded Systems
TI's TMS320C27x Architecture Redefines DSP for the Storage World, Enabling Future-Ready System-on-a-Chip Development for Mass Storage Electronics
What People Are Saying About TI's TMS320C27x
News Releases
Publications:
Details on Signal Processing
Integration Magazine
Mixed Signal Showcase
Trade Shows
Search the News Archives
|
|
TMS320C27x Architecture Feature Summary
DSP Features
- 16-bit, fixed-point architecture
- 100 MIPS performance in 3V, 0.25-micron version
- 10ns instruction cycle time
- 32-bit on-chip data bus fetches two words in a single cycle
- 32-bit accumulator is individually addressable in words or bytes
- 16x16-bit multiplier
- Single-cycle multiply and accumulate (MAC)
- Flexible memory integration options: flash, static random access memory (RAM), read only memory (ROM)
- Complete library of logic and analog peripherals support customizable DSP (cDSP) designs and standard products
- Static CMOS design, efficient memory usage, power-down (idle) mode all minimize power consumption
- Built-in, non-intrusive, real-time emulation using JTAG scan port
- Enhanced Real-time Data Exchange (RTDX) technology for greater visibility into application performance
High-Performance Microcontroller Features
- Separate program and data buses (modified Harvard architecture) configurable in Von Neumann mode (combined bus) support both DSP- and MCU-style operation
- 16- or 32-bit instructions with total address space equal to 4M of 16-bit words for both program and data
- Register-to-register operations
- Single-cycle, single-word, memory-to-memory operations reduce code size and speed performance
- Code-efficient addressing modes minimize program memory requirements
- Dedicated stack pointer
- Fast interrupt response handles multiple operations efficiently
- 80ns latency
- 160ns for full context switch
- MCU-like instruction set mnemonics
- Format: INSTR dest[,src1[,src2]]
# # #
Trademarks:
cDSP and RTDX are trademarks of Texas Instruments Incorporated.
Windows is a trademark of Microsoft Corporation.
|