Looptest output for Pentium Pro/egcs/linux:
Hardware: Pentium Pro quad SMP, 200 MHz, (n2001.lbl.gov), single CPU results
Theoretical peak: 400 Mflops/cpu
OS: Linux 2.1.77 (SMP kernel)
C++ compiler: egcs-980406
Flags: -O
In-cache:
Mflops/s Description
89.758 for, indirection, unit stride
72.661 for, indirection, unit stride, no +=
95.367 for, indirection, unit stride, backwards loops
125.07 for, unroll=4, unit stride, constants loaded into temps
121.1 for, unroll=4, unit stride, constants loaded into temps,
no +=
101.73 for, unroll=4, unit stride, constants loaded into temps,
CSE for index offsets
103.1 for, unroll=4, unit stride, constants loaded into temps, backwards
119.21 for, unroll=8, unit stride, constants loaded into temps
80.309 for, indirection, unit stride, constants into temps
79.473 for, indirection, non-unit stride
88.714 for, indirection, non-unit stride, constants loaded into temps
40.367 while, pointer increment, unit stride
91.92 while, pointer increment, unit stride,
constants loaded into temps
34.367 while, pointer increment, non-unit stride
100.39 while, pointer increment, unroll=4, non-unit stride,
constants loaded into temps
79.473 for, unroll=4, unit stride, constants loaded into temps, prefetching
72.661 interlaced, for, indirection, unit stride
Out of cache:
Mflops/s Description
13.585 for, indirection, unit stride
13.712 for, indirection, unit stride, no +=
15.419 for, indirection, unit stride, backwards loops
20.868 for, unroll=4, unit stride, constants loaded into temps
20.777 for, unroll=4, unit stride, constants loaded into temps,
no +=
20.71 for, unroll=4, unit stride, constants loaded into temps,
CSE for index offsets
20.777 for, unroll=4, unit stride, constants loaded into temps, backwards
20.937 for, unroll=8, unit stride, constants loaded into temps
13.653 for, indirection, unit stride, constants into temps
13.683 for, indirection, non-unit stride
13.712 for, indirection, non-unit stride, constants loaded into temps
13.673 while, pointer increment, unit stride
13.702 while, pointer increment, unit stride,
constants loaded into temps
13.423 while, pointer increment, non-unit stride
20.755 while, pointer increment, unroll=4, non-unit stride,
constants loaded into temps
19.868 for, unroll=4, unit stride, constants loaded into temps, prefetching
12.775 interlaced, for, indirection, unit stride
--------------------- blitz-dev list --------------------------------
* To subscribe/unsubscribe: mail to majordomo@oonumerics.org, with
"subscribe blitz-dev" or "unsubscribe blitz-dev" in the body of the message
* Blitz++ web page: http://oonumerics.org/blitz/
This archive was generated by hypermail 2b29 : Wed Feb 20 2002 - 04:30:04 EST