|
|
MAXQ Competitive Performance
Analysis of the ATmega8
Benchmark Study Details
The following table shows the tools used for comparison with the TI benchmark study, plus the raw results of each individual benchmark in both fastest and smallest configurations.
Notes on the benchmark study:
- The FIR filter benchmark would not run correctly (simulator would not terminate) unless "Block Localization Optimization" was disabled.
| Processor |
ATmega8 |
| Compiler |
CrossWorks for AVR Release 1.0 Build 1 |
| Linker |
CrossWorks for AVR Release 1.0 Build 1 |
| |
Smallest Configuration (Screen Shot) |
Fastest Configuration (Screen Shot) |
| PROJECT |
CODE + CONST |
CYCLES |
CODE + CONST |
CYCLES |
| 8-bit Math |
212 |
279 |
212 |
278 |
| 8-bit Matrix |
220 |
7,348 |
250 |
3,763 |
| 8-bit Switch |
202 |
45 |
200 |
45 |
| 16-bit Math |
222 |
348 |
238 |
330 |
| 16-bit Matrix |
294 |
5,251 |
350 |
5,250 |
| 16-bit Switch |
212 |
50 |
240 |
50 |
| 32-bit Math |
380 |
995 |
460 |
885 |
| Floating Point Math |
816 |
1,491 |
936 |
919 |
| FIR Filter |
860 |
73,598 |
896 |
66,249 |
| Matrix Multiplication |
294 |
11,081 |
348 |
9,231 |
| TOTALS |
3,712 |
100,486 |
4,130 |
87,000 |
|
