The optimized routine

AN2367

6/13

All these results were obtained on a STR910-EVAL best case configuration for speed. The

clock is 96 MHz, the number of wait states is minimum for Flash and for SRAM.

The code is run from Flash.

Using the former table and adding a line for a 342-byte TCP-IP packet gives the following

numbers.

2.5

Portability

Since the code is written in assembler, in principle it can used with any other toolchain.

However it contains a keyword DC32 that is a directive specific to the IAR toolchain.

This directive is used to declare a 32-bit constant.

This can be easily ported by using the Toolchain respective directives.

GNU tool chain

.word

ARM ADS Toolchain

DCD

317

417

544

644

734

834

961

10

61

11

49

12

38

13

65

14

65

15

55

Table 5.

Performance results summary

Number of bytes

Number of cycles

Number of microseconds

32

126

1.38

128

276

2.88

342

1620

6.58

512

852

8.88

1024

1620

16.88

Table 4.

Performance results (no. of bytes not divisible by 16)

Remaining bytes

Additional cycles