| ??? 08/15/07 19:47 Read: times |
#143318 - as it should Responding to: ???'s previous message |
running on an 89C51RD2 at 20MHz and it is taking a second or two to get results from a simple exponential equation.
as it should the '51 is not, and was never intended to be, a 'number cruncher'. I probably can get by with 4 significant figures and The reason I am using floating point is that the range of my input data is about 10 orders of magnitude If you have to stay with the '51 (which, probably, is not a good idea but may be necessary for I/O reasons) I suggest you switch to a SILabs f12x/f13x that will be (6*5) 30 times faster. I do not know if Keil has included using the MAC in some of these, but that may even speed it up more. Erik 6 clocks to 1 clock 100MHz vs 20MHz |
| Topic | Author | Date |
| Log and Exp routines | 01/01/70 00:00 | |
| Precision | 01/01/70 00:00 | |
| Log and Exp | 01/01/70 00:00 | |
| Tried FP51 | 01/01/70 00:00 | |
| Log and Exp | 01/01/70 00:00 | |
| Shouldn't take 1second | 01/01/70 00:00 | |
| What exactly are you doing? | 01/01/70 00:00 | |
| long not sufficient ? | 01/01/70 00:00 | |
| as it should | 01/01/70 00:00 | |
| Mutually Exclusive? | 01/01/70 00:00 | |
| Coprocessor? http://www.micromegacorp.com | 01/01/70 00:00 | |
| Good idea - here's some more | 01/01/70 00:00 | |
| does that make sense? | 01/01/70 00:00 | |
| What you say ain't what you do. | 01/01/70 00:00 | |
| Why so touchy?! | 01/01/70 00:00 | |
what "hidden agenda" | 01/01/70 00:00 |