Post Created date
Less intensive checksum..
Yes, I understand about the inline asm, what I don't understand is how that ASM correlates at all to the pseudocode. Basically, I'm not seeing anything that corresponds to the ^...
Monday, 14 September 2020 - 20:14
Less intensive checksum..
I am really confused by that, I can't see how it does anything of the sort, but presumably it does.  
Sunday, 13 September 2020 - 15:58
Less intensive checksum..
Huh. What are the 0x98/0xFE coming from? I suppose that could be the address lookup for the table, maybe? Because that would actually sort of make sens; it looks a bit like 242-...
Saturday, 12 September 2020 - 23:06
Less intensive checksum..
I would absolutely not recommend FNV as a replacement for CRC, it's going to be dramatically slower than a reasonable CRC implementation on AVR, and I wouldn't consider it a very...
Friday, 11 September 2020 - 15:15
Less intensive checksum..
The crc hash function is not especially cheap, although a lot of modern hashes that are cheaper on desktop hardware might well be significantly worse on AVR, where anything over 8...
Friday, 11 September 2020 - 04:00
Bit manipulation in Arrays In AVR Embedded C
And yes, good point, the last value is indeed octal.
Friday, 28 August 2020 - 14:59
Bit manipulation in Arrays In AVR Embedded C
Values are not binary or non-binary. Values are just values.   Constants you put in your code could be binary, decimal, hexadecimal, or whatever. But values aren't.   So...
Thursday, 27 August 2020 - 21:59
Bit manipulation in Arrays In AVR Embedded C
Those are not binary values. Those are decimal values that have only zeroes and ones in them.
Thursday, 27 August 2020 - 14:56
Fast dividing by 5 and by 10 for uint32_t input data for ATMEGA16A
I did confirm, by the way, that "divide by 100 instead of by 10" wins significantly.
Tuesday, 7 April 2020 - 01:12
Fast dividing by 5 and by 10 for uint32_t input data for ATMEGA16A
Hmm. So, idle thought: 2^24 is 16M. The OP's examples appeared to strongly suggest only needing 7-digit values... which means that you possibly don't need to actually compute the...
Sunday, 5 April 2020 - 18:16
Fast division of 32 bits by 5, code in assembly
I think you should use a lookup table for the inner loop to divide an arbitrary 8-bit value by 5. Costs 256 bytes of flash, saves a LOT of cycles.
Tuesday, 31 March 2020 - 04:20
Fast dividing by 5 and by 10 for uint32_t input data for ATMEGA16A
westfw wrote: I'd guess you could compute X/5, where X is an 8-bit value, We want 32bits, not 8bits.  That makes a lookup table impractical. An interesting question is...
Monday, 30 March 2020 - 14:57

Pages