the first link I gave is free web acess. You should have entered your question, and just watch what happens.
I bought the app based on a query about satellites, which it promptly found and expounded on.
the first link I gave is free web acess. You should have entered your question, and just watch what happens.
I bought the app based on a query about satellites, which it promptly found and expounded on.
Den mandag den 23. februar 2015 kl. 00.41.51 UTC+1 skrev John Larkin:
up till the last few years of uni I used one of these:
Then I got one of these:
these days I use google for the simple stuff and matlab for the rest, scripting, a proper keyboard and editing just makes things so much easier
-Lasse
the first link I gave is free web acess. You should have entered your question, and just watch what happens.
I bought the app based on a query about satellites, which it promptly found and expounded on.
2's complement fixed point fractional notation, so 0x7FFF FFFF is almost 1 and 0x800 0000 is almost -1. No overflow is possible in my multiply routines and a minimum amount of resolution is lost.
If you want to multiply by a number >1 then it can be left-shifted before or after (or both), so a multiply by 1.8 could be a multiply by two and a fractional multiply by 0x7333 3333
Best regards, Spehro Pefhany
-- "it's the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
Den mandag den 23. februar 2015 kl. 03.40.03 UTC+1 skrev Spehro Pefhany:
that would be a strange fractional notation, normally
0x7FFFFFFF is almost 1 and 0x80000000 exactly -1 it has to be, you have an odd number of numbers to represent; n-positive n-negative and one zero, but x bit can represent an even number so you have to have an extra somewherebut that also means you have risk of overflow in multiplications*, because of -1*-1 = +1
*when you throw away the extra signbit to get back to the same formatQ1.31 * Q1.31 = Q2.62
-Lasse
You're absolutely right- I mispoke. 0x80... is exactly -1.
Best regards, Spehro Pefhany
-- "it's the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
I used to do that a lot, but now I have an embedded system programmer, and it's hard to explain this sort of math, and we use ARMs with hardware FP, so we do most everything in floats now, just clip and scale to ints at the very end.
-- John Larkin Highland Technology, Inc picosecond timing laser drivers and controllers jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Yes.. same here, but unfortunately the FPUs in the ARMs I've used are only single precision and 24 bits of precision won't always cut it.
I think you have to go to an ARM9 or better to get an option of double precision.
Best regards, Spehro Pefhany
-- "it's the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
there's sixteen candidates, none of them work.
If your allowed to reorder the left side
-- umop apisdn
Me, too. But not back in 1979 which was when I used the HP-25.
In 79, I was running a PDP11 timeshare system, with Basic+. You could run a *300* line program!
-- John Larkin Highland Technology, Inc picosecond timing laser drivers and controllers jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
You probably didn't carry it around with you as I did with the HP-25.
That theme again.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Apparently it is on an HP16.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
I used to use a thing I called Zform, which is 64 bits 2's comp, S32.32. It will handle most any real-world engineering units. I wrote a saturating (exception-free) math package in assembly for the 68K. Some operations, like Z-to-int and adds are easy. No normalizations are needed for those, which saves a lot of hassle. The divide was the nuisance.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
It had a room of its own, with a line printer and an operator!
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Fields has never heard of integer arithmetic or rounding. No surprise.
Since you have the app, I would have thought you could tell us if it worked.
Counting binary places is exactly what I did *not* want to do (but was forced to anyway).
By '79 I was running on MVS, VM, and VS. For a programmable calculator, I used APL. ;-) Graphics, too (19" Tektronix storage tube).
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.