I'm reading an IMU on the arduino board with a s-function block in simulink by double or single data types though I just need 2 decimals precision as ("xyz.ab").I want to improve the performance with changing data types and wonder that;
is there a way to decrease the precision to 2 decimals in s-function block or by adding/using any other conversion blocks/codes in the simulink aside from using fixed-point tool?
For true fixed point transfer, fixed-point toolbox is the most general answer, as stated in Phil's comment.
However, to avoid toolbox use, you could also devise your own fix-point integer format and add a block that takes a floating point input and convert it into an integer format (and vice versa on the output).
E.g. If you know the range is 327.68 < var < 327.67 you could just define your float as an int16 divided by 10. In a matlab function block you would then just say
y=int16(u*100.0);
to convert the input to the S-function.
On the output it would be a reversal
y=double(u)/100.0;
(Eml/matlab function code can be avoided by using multiply, divide and convert blocks.)
However, be mindful of the bits available and that the scaling (*,/) operations is done on the floating point rather than the integer.
2^(nrOfBits-1)-1 shows you what range you can represent including signeage. For unsigned types uint8/16/32 the range is 2^(nrOfBits)-1. Then you use the scaling to fit the representable bit into your used floating point range. The scaled range divided by 2^nrOfBits will tell you what the resolution will be (how large are the steps).
You will need to scale the variables correspondingly on the Arduino side as well when you go to an integer interface of this type. (I'm assuming you have access to that code - if not it'd be hard to use any other interface than what is already provided)
Note that the intXX(doubleVar*scale) will always truncate the values to integer. If you need proper rounding you should also include the round function, e.g.:
int16(round(doubleVar*scale));
You don't need to use a base 10 scale, any scaling and offsets can be used, but it's easier to make out numbers manually if you keep to base 10 (i.e. 0.1 10.0 100.0 1000.0 etc.).
As a final note, if the Arduino code interface is floating point (single/double) and can't be changed to integer type; you will not get any speedup from rounding decimals since the full floating point is what will be is transferred anyway. Even if you do manage to reduce the data a bit using integers I suspect this might not give a huge speedup unless you transfer large amounts of data. The interface code will have a comparatively large overhead anyway.
Good luck with your project!
Related
I want to convert data(double precision,15 decimal points) to data of another type(quadruple precision,34 decimal points). So, I used vpa function like this:
data = sin(2*pi*frequency*time);
quad_data = vpa(data,34);
But, the type of the result is sym, not double. And when I checked each cell of the sym type data, 1x1 sym was created in each cell. I tried to use fft function using 'quad_data', but it didn't work. Is there any solution that I can change the decimal point of double type from 15 to 34?
The only numeric floating point types that MATLAB currently supports is double, single, and half. Extended precision types can be achieved via the Symbolix Toolbox (e.g., vpa) or 3rd party code (e.g., John D'Errico's FEX submission High Precision Floating HPF class). But even then, only a subset of floating point functions will typically be supported. If the function you are trying to use doesn't support the variable type, then you would have to supply your own function.
Also, you are not building vpa objects properly in the first place. Typically you would convert the operands to vpa first and then do arithmetic on them. Doing the arithmetic in double precision first as you are doing with data, and then converting to extended precision vpa, just adds garbage to the values. E.g., set the digits first and then use vpa('pi') to get the full extended precision version of pi as a vpa variable.
There is a commercial 3rd-party toolbox for this purpose, called the Multiprecision Computing Toolbox for MATLAB.
This tool implements many of the mathematical operations you would expect from double inputs, and according to benchmarks on the website, it's much faster than vpa.
Disclosure: I am not affiliated with the creators of this tool in any way, however I can say that we had a good experience with this tool for one of our lab's projects.
The other suggestion I can give is doing the high-precision arithmetic in an another language\environment to which MATLAB provides interfaces (e.g., C, python, java), and which should have the quad data type implemented.
I am working on a deep learning problem where I am trying to predict time-to-failure on laboratory earthquake data from an observed seismic time series. The target is a single integer number (the time until the next earthquake) ranging, say, from 1 to 10.
I could design the last layer to return a single float and use, say, mean-square-error(MSE), as a loss to make that returned float close to the desired integer. Or, I could think of each integer possibility as a "class" and use a cross-entropy(CE) loss to optimize.
Are there any arguments in favour of either of these options?
Also, what if the target is a float number ranging from 1 to 10? I could also turn this into a class/CE problem.
So far, I have tried the CE option (which works at some level) and am thinking of trying the mse option but wanted to step back and think before proceeding. Such thoughts would include reasoning as to why one approach might outperform the other.
I am working with pytorch version 1.0.1 and Python 3.7.
Thanks for any guidances.
I decided to just implement a float head with an L1Loss in Pytorch and I created a simple but effective synthetic data set to test the implementation. The data set created images into which a number of small squares were randomly drawn. The training label was simply the number of squares divided by 10, a float number with one decimal digit.
The net trained very quickly and to a high degree of precision - the test samples were correct to the one decimal digit.
As to the original question, the runs I made definitely favoured the float over class.
My take on this is that the implementation in classes had a basic imprecision in the assignment of the classes and, perhaps more importantly, the class implementation has no concept of a "metric". That is, the float implementation, even if it misses the exact match, will try to generate an output label "close" to the input label while the class implementation has no concept of "close".
One warning with Pytorch. If you are fitting to one float, be sure to encase it in a length 1 vector in the data generator. Pytorch cannot handle a "naked" float number (even though it does become a vector when batches are done). But it doesn't complain. This cost me a bunch of time.
I am building a neural network running on an FPGA, and the last piece of the puzzle is running a sigmoid function in hardware. This is either:
1/(1 + e^-x)
or
(atan(x) + 1) / 2
Unfortunately, x here is a float value (a real value in SystemVerilog).
Are there any tips on how to implement either of these functions in SystemVerilog?
This is really confusing to me since both of these functions are complex, and I don't even know where to begin implementing them due to the added complexity of being float values.
One simpler way for this is to create a memory/array for this function. However that option can be highly inefficient.
x should be the input address for the memory and the value at that location can be the output of the function.
Suppose value of your function is as follow. (This is just an example)
x = 0 => f(0) = 1
x = 1 => f(0) = 2
x = 2 => f(0) = 3
x = 3 => f(0) = 4
So you can create an array for this, which stored the output values.
int a[4] = `{1, 2, 3, 4};
I just finished this by Vivado HLS, which allows you to write circuits in C.
Here is my C code.
#include math.h
void exp(float a[10],b[10])
{
int i;
for(i=0;i<10;i++)
{
b[i] = exp(a[i]);
}
}
But there is a question that it is impossible to create a unsized matrix. Maybe there is another way that I don't know.
As you seem to realize, type real is not synthesizable. you need to operate on the type integer mantissa and type integer exponent separately and combine them when you are done, having tracked the sign. Once you take care of (e^-x), the rest should be straight-forward.
try this page for a quick explanation: https://www.geeksforgeeks.org/floating-point-representation-digital-logic/
and search on "floating point digital design" for more explanations/examples.
Do you really need a floating number for this? Is fixed point sufficient?
Considering (atan(x) + 1) / 2, quite likely the only useful values of x are those where the exponent is fairly small. (if the exponent is large, your answer is pi/2).
atan of a fixed point number can be calculated in HW fairly easily; there are cordic methods (see https://zipcpu.com/dsp/2017/08/30/cordic.html) and direct methods; see for example https://dspguru.com/dsp/tricks/fixed-point-atan2-with-self-normalization/
FPGA design flows in which hardware (FPGA) is targeted generally do not support floating point numbers in the FPGA fabric. Fixed point with limited precision is more commonly used.
A limited precision fixed point approach:
Use Matlab to create an array of samples for your math function such that the largest value is +/- .99999. For 8 bit precision (actually 7 with sign bit), multiply those numbers by 128, round at the decimal point and drop the fractional part. Write those numbers to a text file in 2s complement hex format. In SystemVerilog you can implement a ROM using that text file. Use $readmemh() to read these numbers into a memory style variable (one that has both a packed and unpacked dimension). Link to a tutorial:
https://projectf.io/posts/initialize-memory-in-verilog/.
Now you have a ROM with limited precision samples of your function
Section 21.4 Loading memory array data from a file in the SystemVerilog specification provides the definition for $readmh(). Here is that doc:
https://ieeexplore.ieee.org/document/8299595
If you need floating point one possibility is to use a processor soft core with a floating point unit implemented in FPGA fabric, and run software on that core. The core interface to the rest of the FPGA fabric over a physical bus such as axi4 steaming. See:
https://www.xilinx.com/products/design-tools/microblaze.html to get started.
It is a very different workflow than ordinary FPGA design and uses different tools. C or C++ compiler with math libraries (tan, exp, div, etc) would be used along with the processor core.
Another possibility for fixed point is an FPGA with a hard core processor. Xilinx Zynq is one of them. This is a complex and powerful approach. A free free book provides knowledge on how to use Zynq
http://www.zynqbook.com/.
This workflow is even more complex that soft core approach because the Zynq is a more complex platform (hard processor & FPGA integrated on one chip).
Its pretty hard to implement non-linear functions like that in hardware, and on top of that floating point arithmetic is even more costly. Its definitely better(and recommended) to work with fixed point arithmetic as mentioned in answers before. The number of precision bits in fixed point arithmetic will depend on your result accuracy and the error tolerance.
For hardware implementations, any kind of non-linear function can be approximated as piecewise linear function, and use a ROM based implementation approach as described in previous answers. The number of sample points that you take from the non-linear function determines your accuracy. The more samples you store the better approximation of the function you get. Often in hardware , number of samples you can store can become restricted by the amount of fast/local memory available to you. In this case to optimize the memory resources, you can add a little extra compute resources and perform linear interpolation to calculate the needed values.
We can write a simple Rational Number class using two integers representing A/B with B != 0.
If we want to represent an irrational number class (storing and computing), the first thing came to my mind is to use floating point, which means use IEEE 754 standard (binary fraction). This is because irrational number must be approximated.
Is there another way to write irrational number class other than using binary fraction (whether they conserve memory space or not) ?
I studied jsbeuno's solution using Python: Irrational number representation in any programming language?
He's still using the built-in floating point to store.
This is not homework.
Thank you for your time.
With a cardinality argument, there are much more irrational numbers than rational ones. (and the number of IEEE754 floating point numbers is finite, probably less than 2^64).
You can represent numbers with something else than fractions (e.g. logarithmically).
jsbeuno is storing the number as a base and a radix and using those when doing calcs with other irrational numbers; he's only using the float representation for output.
If you want to get fancier, you can define the base and the radix as rational numbers (with two integers) as described above, or make them themselves irrational numbers.
To make something thoroughly useful, though, you'll end up replicating a symbolic math package.
You can always use symbolic math, where items are stored exactly as they are and calculations are deferred until they can be performed with precision above some threshold.
For example, say you performed two operations on a non-irrational number like 2, one to take the square root and then one to square that. With limited precision, you may get something like:
(√2)²
= 1.414213562²
= 1.999999999
However, storing symbolic math would allow you to store the result of √2 as √2 rather than an approximation of it, then realise that (√x)² is equivalent to x, removing the possibility of error.
Now that obviously involves a more complicated encoding that simple IEEE754 but it's not impossible to achieve.
I tried to assign a very small number to a double value, like so:
double verySmall = 0.000000001;
9 fractional digits. For some reason, when I multiplicate this value by 10, I get something like 0.000000007. I slighly remember there were problems writing big numbers like this in plain text into source code. Do I have to wrap it in some function or a directive in order to feed it correctly to the compiler? Or is it fine to type in such small numbers in text?
The problem is with floating point arithmetic not with writing literals in source code. It is not designed to be exact. The best way around is to not use the built in double - use integers only (if possible) with power of 10 coefficients, sum everything up and display the final useful figure after rounding.
Standard floating point numbers are not stored in a perfect format, they're stored in a format that's fairly compact and fairly easy to perform math on. They are imprecise at surprisingly small precision levels. But fast. More here.
If you're dealing with very small numbers, you'll want to see if Objective-C or Cocoa provides something analagous to the java.math.BigDecimal class in Java. This is precisely for dealing with numbers where precision is more important than speed. If there isn't one, you may need to port it (the source to BigDecimal is available and fairly straightforward).
EDIT: iKenndac points out the NSDecimalNumber class, which is the analogue for java.math.BigDecimal. No port required.
As usual, you need to read stuff like this in order to learn more about how floating-point numbers work on computers. You cannot expect to be able to store any random fraction with perfect results, just as you can't expect to store any random integer. There are bits at the bottom, and their numbers are limited.