Texas Instruments TMS320F2837 D Series Workshop Guide And Lab Manual

Math Accelerators
1 - 14 TMS320F2837xD Microcontroller Workshop - Architecture Overview
Trigonometric Math Unit (TMU)
Trigonometric Math Unit (TMU)
 Supported by natural C and C-intrinsics
 Significant performance impact on algorithms such as:
• Park / Inverse Park • dq0 Transform & Inverse dq0
• Space Vector GEN • FFT Magnitude & Phase Calculations
Adds instructions to FPU for
calculating common
Trigonometric operations
x
y
r
y = r * sin(rad)
x = r * cos(rad)
Operation Instruction Exe Cycles Result Latency FPU Cycles w/o TMU
Z = Y/X
DIVF32 Rz,Ry,Rx
1 5 ~24
Y = sqrt(X)
SQRTF32 Ry,Rx
1 5 ~26
Y = sin(X/2pi)
SINPUF32 Ry,Rx
1 4 ~33
Y = cos(X/2pi)
COSPUF32 Ry,Rx
1 4 ~33
Y = atan(X)/2pi
ATANPUF32 Ry,Rx
1 4 ~53
Instruction To
Support ATAN2
Calculation
QUADF32 Rw,Rz,Ry,Rx
ATANPUF32 Ra,Rz
ADDF32 Rb,Ra,Rw
3 11 ~90
Y = X * 2pi
MPY2PIF32 Ry,Rx
1 2 ~4
Y = X * 1/2pi
DIV2PIF32 Ry,Rx
1 2 ~4
The Trigonometric Math Unit (TMU) is an extension of the FPU and the C28x instruction set, and
it efficiently executes trigonometric and arithmetic operations commonly found in control system
applications. Similar to the FPU, the TMU provides hardware support for IEEE-754 single-
precision floating-point operations that are specifically focused on trigonometric math functions.
Seamless code integration is accomplished by built-in compiler support that automatically
generates TMU instructions where applicable. This dramatically increases the performance of
trigonometric functions, which would otherwise be very cycle intensive. It uses the same pipeline,
memory bus architecture, and FPU registers as the FPU, thereby removing any special
requirements for interrupt context save or restore.

Texas Instruments TMS320F2837 D Series Specifications

General

Brand	Texas Instruments
Model	TMS320F2837 D Series
Category	Microcontrollers
Language	English