Note: we ran the tests for different amounts of time on each processor. The number of loops in the code may not correspond to the data on any particular processor.
Intel Pentium II
Endianness: little-endian
Cache structure (obtained from internet):
line length of 32 bytes
four-way set associative
32 kbyte cache divided into:
16 kbyte instruction cache16 kbyte data cache
Cache structure (obtained through testing):
Baseline 1: 60 seconds
Baseline 2: 62 seconds
Cache Stress: 75 seconds
Two-Way Nice: 50 seconds
char: 218 seconds
long: 218 seconds
double: 224 seconds
float: 223 seconds
Baseline 1: 651 seconds
Baseline 2: 677 seconds
Dependency 3: 657 seconds
Dependency 2: 734 seconds
Intel Pentium III
Endianness: little-endian
Cache structure (obtained from internet):
32 kbyte divided into:
16 kbyte instruction cache
16 kbyte data cache
32 byte line
4-way set associative
Cache structure (obtained through testing):
Baseline 1: 8 seconds
Baseline 2: 8 seconds
Cache Stress: 10 seconds
Two-Way Nice: 4 seconds
char: 129 seconds
long: 123 seconds
double: 111 seconds
float: 112 seconds
Baseline 1: 31 seconds
Baseline 2: 31 seconds
Dependency 3: 30 seconds
Dependency 2: 36 seconds
Endianness: big-endian
Cache structure (obtained from internet):
32 kbyte cache divided into:
16 kbyte instruction cachepseudo-two-way set associative with 32 byte blocks
8-instruction wide line
16 kbyte data cache
direct mapping
2 16 byte sub-blocks per line
Cache structure (obtained through testing):
Baseline 1: 35 seconds
Baseline 2: 34 seconds
Cache Stress: 52 seconds
Two-Way Nice: 35 seconds
char: 4719 seconds
long: 4494 seconds
double: 2027 seconds
float: 1572 seconds
Baseline 1: 2438 seconds
Baseline 2: 2452 seconds
Dependency 3: 2688 seconds
Dependency 2: 3140 seconds
Click here to go to the previous part of the lab write-up.