Texas Instruments TMS320C6000 Programmer's Manual page 193

Software Pipelining
Example 6–33. Assembly Code for Floating-Point Dot Product (Software Pipelined
With Smallest Code Size)
B .S2
|| MVK .S1
B .S2
|| ZERO .L1
|| ZERO .L2
B .S2
|| ZERO .L1
|| ZERO .L2
B .S2
|| ZERO .L1
|| ZERO .L2
B .S2
|| ZERO .L1
|| ZERO .L2
LOOP:
LDDW .D1
|| LDDW .D2
|| MPYSP .M1X
|| MPYSP .M2X
|| ADDSP .L1
|| ADDSP .L2
||[A1] B .S2
||[A1] SUB .S1
; Branch occurs here
ADDSP .L1X
ADDSP .L2X
ADDSP .L1X
ADDSP .L2X
NOP
ADDSP .L1X
NOP
ADDSP .L2X
NOP
ADDSP .L1X
NOP
6-56
LOOP ; branch to loop
53,A1 ; set up loop counter
LOOP ;* branch to loop
A7 ; zero out mpysp input
B7 ; zero out mpysp input
LOOP ;** branch to loop
A8 ; zero out sum0 accumulator
B8 ; zero out sum0 accumulator
LOOP ;*** branch to loop
A5 ; zero out addsp input
B5 ; zero out addsp input
LOOP ;**** branch to loop
A6 ; zero out mpysp input
B6 ; zero out mpysp input
A4++,A7:A6 ;********* load ai & ai + 1 from memory
B4++,B7:B6 ;********* load bi & bi + 1 from memory
A6,B6,A5 ;**** pi = a0
A7,B7,B5 ;**** pi1 = a1
A5,A8,A8 ; sum0 += (ai
B5,B8,B8 ; sum1 += (ai+1
LOOP ;***** branch to loop
A1,1,A1 ;****** decrement loop counter
A8,B8,A0 ; sum(0) = sum0(0) + sum1(0)
A8,B8,B0 ; sum(1) = sum0(1) + sum1(1)
A8,B8,A0 ; sum(2) = sum0(2) + sum1(2)
A8,B8,B0 ; sum(3) = sum0(3) + sum1(3)
; wait for B0
A0,B0,A5 ; sum(01) = sum(0) + sum(1)
; wait for next B0
A0,B0,B5 ; sum(23) = sum(2) + sum(3)
3
A5,B5,A4 ; sum = sum(01) + sum(23)
3 ;
b0
b1
bi)
bi+1)