This document compares the arithmetic and digital signal processing capabilities of the Analog Devices ADSP-2101 and the AT&T DSP16A. It details differences in their arithmetic architectures, including the ADSP-2101's ability to fetch instructions and data in a single cycle versus the DSP16A's multi-cycle requirements. The text also covers the distinct features of their Arithmetic Logic Units (ALU), Multiplier/Accumulators (MAC), and barrel shifters.
| ALU Operands | 16-bit data operands |
|---|---|
| MAC Overflow Capacity | 256 MAC overflows |
| ALU Result Register Width | 36 bits |
| DSP16A MAC Accumulator Width | 36 bits |
| DSP16A MAC Overflow Capacity | 16 times |
Yes, the ADSP-2101 ALU supports 32-bit arithmetic using addition with carry and subtraction with borrow instructions.
No, the DSP16A has no provisions for block floating point.