Scribd
Upload
44 views3 pages

2012EEE2188 DSP Experiment 5

This document discusses three different methods for implementing an FIR filter on a C54x DSP: 1. Using parallel multiply-accumulate (MAC) instructions with an asymmetric filter where the number of coefficients is different than the number of input samples. 2. Using parallel MAC with a symmetric filter where the number of coefficients equals the number of input samples. 3. Using double-precision parallel instructions with an asymmetric filter to perform the multiplication and addition in parallel.

Uploaded by

Abhiroop Verma
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
44 views3 pages

2012EEE2188 DSP Experiment 5

This document discusses three different methods for implementing an FIR filter on a C54x DSP: 1. Using parallel multiply-accumulate (MAC) instructions with an asymmetric filter where the number of coefficients is different than the number of input samples. 2. Using parallel MAC with a symmetric filter where the number of coefficients equals the number of input samples. 3. Using double-precision parallel instructions with an asymmetric filter to perform the multiplication and addition in parallel.

Uploaded by

Abhiroop Verma
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

FIR Filter Implementation-2

A. Parallel MAC - Asymmetric N .set 4;h M .set 4;x .mmregs .def start start BCLR C54CM BCLR AR0LC ;Linear Mode AMOV #01000h,XAR0 ;Starting Address for Input X AMOV #01000h+1,XAR2 ;XXXXXXXX BCLR AR1LC ;Linear Mode AMOV #010020h,XAR1 ;Starting Address for Output BSET CDPLC ;Circular Addressing BSET SATD ;To Saturate Accumulator BSET FRCT ;To shift Accumulator left by one bit AMOV #000100h,XCDP ;Starting Address for coefficient H MOV #0h,BSAC ;Offset for circular Addressing of CDP MOV #N,BKC ;Size of Block for Circular Addressing MOV #(M+N-1)/2,BRC0 ;Initializing for local block repeat BR0+1 times MOV #N+1,T0 ;Offset to move pointer of AR0 back. RPTBLOCAL LOOP MOV #0,AC0 MOV #0,AC1 RPT #N-2 MAC *AR0-,*CDP+,AC0 :: MAC *AR2-,*CDP+,AC1 MAC *(AR0+T0),*CDP+,AC0 :: MAC *(AR2+T0),*CDP+,AC1 MOV HI(AC0),*AR1+ ;Storing Output LOOP MOV HI(AC1),*AR1+ ;Storing Output IDLE B. Using Parallel MAC - Symmetric N M .set 4 ;no filter coefficients .set 4 ;size of input .mmregs .def start

start BCLR C54CM BCLR AR0LC ;linear mode

; ; ;

AMOV #010000H,XAR0 ;Input Start Addr BCLR AR2LC ; Set Linear Mode AMOV #01000AH,XAR2 ; Input Start Addr BSET CDPLC AMOV #000100H,XCDP BCLR AR1LC ; Set Linear Mode AMOV #010020h,XAR1 ; Output Start Addr BCLR AR3LC ; Set Linear Mode AMOV #01002Ah,XAR3 BCLR AR3LC ; Set Linear Mode AMOV #01002Ah,XAR4 ; Output Start Addr

BSET CDPLC ; Circular Adressing BSET SATD ; Saturate Accumalator BSET FRCT ; Left Shift Acc by 1 bit ; AMOV #000100h,XCDP ;Statr Addr of Coefficients MOV #0h,BSAC ;Offset for Circular addressing of CDP MOV #N,BKC ;Size of Block for Circular Address ;MOV #0h,BSAC ;MOV #N,BKC ; MOV #0h,AC0 ; MOV #0h,AC1 ; MOV #(M/2)-1,BRC0 ; Initialization of Local Block repeat ; MOV #M-1,BRC2 MOV #N,T0 ;Offset to move pointer Of AR0 back RPTBLOCAL LOOP MPYM *AR0-,*CDP,AC0 ; MPYM *AR2-,*CDP+,AC1 ; ;MOV #N-1,T1 ;RPTBLOCAL LOOP ;MOV #0h,AC0 ;Repeats BRC0+1 times RPT #(N-3) MAC *AR0-,*CDP+,AC0 :: MAC *AR2-,*CDP+,AC1 ; MAC *(AR0+T0),*CDP+,AC0 :: MAC *(AR2+T0),*CDP+,AC1 ; LOOP MOV HI(AC0),*AR1+ :: MOV HI(AC1),*AR4+ ;Store at Oupput loc. IDLE

C. Using dbl-pair-Asymmetric N M .set 4;h .set 4;x .mmregs .def start

start

BCLR C54CM BCLR AR0LC ;Linear Mode AMOV #01000h,XAR0 ;Starting Address for Input X AMOV #01000h+1,XAR2 BCLR AR1LC ;Linear Mode AMOV #010020h,XAR1 ;Starting Address for Output BSET CDPLC ;Circular Addressing BSET SATD ;To Saturate Accumulator BSET FRCT ;To shift Accumulator left by one bit AMOV #000100h,XCDP ;Starting Address for coefficient H MOV #0h,BSAC ;Offset for circular Addressing of CDP MOV #N,BKC ;Size of Block for Circular Addressing MOV #(M+N-1)/2,BRC0 ;Initializing for local block repeat BR0+1 times MOV #N+1,T0 ;Offset to move pointer of AR0 back. RPTBLOCAL LOOP MPY *AR0-,*CDP+,AC0 :: MPY *AR2-,*CDP+,AC1 || RPT #N-3 MAC *AR0-,*CDP+,AC0 :: MAC *AR2-,*CDP+,AC1 MAC *(AR0+T0),*CDP+,AC0 :: MAC *(AR2+T0),*CDP+,AC1

; MOV HI(AC0),*AR1+ ;Storing Output LOOP MOV PAIR(HI(AC0)),DBL(*AR1+) ;Storing Output IDLE

You might also like