Lesson2_C.asm - Texas Instruments TMS320C6000 Programmer's Manual

Lesson 2: Balancing Resources With Dual-Data Paths
Example 2–9. lesson2_c.asm
;*––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––*
;* SOFTWARE PIPELINE INFORMATION
;*
;* Loop Unroll Multiple
;* Known Minimum Trip Count
;* Known Maximum Trip Count
;* Known Max Trip Count Factor
;* Loop Carried Dependency Bound(^) : 0
;* Unpartitioned Resource Bound
;* Partitioned Resource Bound(*)
;* Resource Partition:
;*
;* .L units
;* .S units
;* .D units
;* .M units
;* .X cross paths
;* .T address paths
;* Long read paths
;* Long write paths
;* Logical
;* Addition ops (.LSD)
;* Bound(.L .S .LS)
;* Bound(.L .S .D .LS .LSD)
;*
;* Searching for software pipeline schedule at ...
;* ii = 3
;* done
;*
;* Epilog not entirely removed
;* Collapsed epilog stages
;*
;* Prolog not entirely removed
;* Collapsed prolog stages
;*
;* Minimum required memory pad : 8 bytes
;*
;* Minimum safe trip count
2-16
A–side
ops (.LS)
Schedule found with 5 iterations in parallel
Notice the following things in the feedback:
A schedule with three cycles (ii=3) : You can tell by looking at the .D units and
.T address paths that this 3–cycle loop comes after the loop has been unrolled
because the resources show a total of six memory accesses evenly balanced
between the A side and B side. Therefore, our new effective loop iteration inter-
val is 3/2 or 1.5 cycles.
A Known Minimum Trip Count of 10 : This is because we specified the count
of the original loop to be greater than or equal to twenty and a multiple of two
and after unrolling, this is cut in half. Also, a new line, Known Maximum Trip
: 2x
: 10
: 1073741823
: 1
: 3
: 3
B–side
0 0
2 1
3* 3*
2 2
1 1
3* 3*
1 1
0 0
1 1 (.L or .S unit)
0 1 (.L or .S or .D unit)
2 1
2 2
: 2
: 3
: 4