Texas Instruments TMS320C6000 Programmer's Manual page 271

Software Pipelining the Outer Loop
Example 6–71. Final Assembly Code for FIR Filter With Redundant Load Elimination and
No Memory Hits With Outer Loop Software-Pipelined (Continued)
ADD
|| MPY
|| MPY
|| LDH
|| LDH
LOOP:
ADD
|| ADD
|| MPY
|| MPY
|| LDH
|| LDH
ADD
|| ADD
|| MPY
|| MPY
|| LDH
|| LDH
ADD
|| ADD
||[B2] B
|| MPY
|| MPY
|| LDH
|| LDH
||[B2] SUB
ADD
|| ADD
|| MPY
|| MPY
|| LDH
|| LDH
; inner loop branch occurs here
ADD
|| ADD
|| MPY
|| MPY
|| SUB
|| SUB
||[A2] B
6-134
.L1 A0,A9,A9
.M2X A5,B8,B8
.M1X B0,A7,A5
.D2 *B1++[2],B0
.D1 *A4++[2],A0
.L2X A1,B9,B9
.L1X B6,A9,A9
.M2 B5,B8,B7
.M1 A5,A7,A7
.D1 *A8++[2],B6
.D2 *B4++[2],A1
.L2 B7,B9,B9
.L1 A5,A9,A9
.M1X B5,A1,A0
.M2X A0,B6,B6
.D1 *A4++[2],A5
.D2 *B1++[2],B5
.L2X A7,B9,B9
.L1X B8,A9,A9
.S1 LOOP
.M2 B0,B6,B7
.M1 A0,A1,A1
.D2 *B4++[2],A7
.D1 *A8++[2],B8
.S2 B2,1,B2
.L2 B7,B9,B9
.L1 A0,A9,A9
.M2X A5,B8,B8
.M1X B0,A7,A5
.D2 *B1++[2],B0
.D1 *A4++[2],A0
.L2X A1,B9,B9
.L1X B6,A9,A9
.M2 B5,B8,B7
.M1 A5,A7,A7
.D1 A4,A3,A4
.D2 B4,B10,B4
.S1 OUTLOOP
; sum0 += x0 * h0
; x3 * h3
; x2 * h2
;** x2 = x[j+i+2]
;** x1 = x[j+i+1]
; sum1 += x1 * h0
; sum0 += x1 * h1
; x0 * h3
; x3 * h2
;** h1 = h[i+1]
;** h0 = h[i]
; sum1 += x2 * h1
; sum0 += x2 * h2
;* x0 * h0
;* x1 * h1
;** x3 = x[j+i+3]
;** x0 = x[j+i+4]
; sum1 += x3 * h2
; sum0 += x3 * h3
;* branch to loop
;* x2 * h1
;* x1 * h0
;** h2 = h[i+2]
;** h3 = h[i+3]
;** decrement loop counter
; sum1 += x0 * h3
;* sum0 += x0 * h0
;* x3 * h3
;* x2 * h2
;*** x2 = x[j+i+2]
;*** x1 = x[j+i+1]
;e sum1 += x1 * h0
;e sum0 += x1 * h1
;e x0 * h3
;e x3 * h2
;o reset x pointer to x[j]
;o reset h pointer to h[0]
;o branch to outer loop