In Dymola, I'm able to do something like:
when time > 100 then
assert(false,"Simulation taking too long");
end when;
to stop simulations based on the time variable itself.
However, what i'd like to do is stop the simulation based on elapsed CPU time. Dymola has a way to output the CPU time and it shows up in the results as CPUtime, but I don't know how to access the variable. In other words, this is what i'd like to do, but the CPUtime variable isn't in scope:
when CPUtime > 100 then
assert(false,"Simulation taking too long");
end when;
Any suggestions, either how to access CPUtime, or other workarounds to kill simulations based on cpu time?
As already noted:
You can set this in Dymola 2022 in the simulation setup, or alternatively by setting Advanced.Simulation.MaxRunTime.
It's wall-clock time, which means that if you have a parallel simulation it will stop after 10s has passed and not when the cores together have spent 10s, and if you for some weird reason have a long sleep-call in the model it will still end.
(This was already noted in the comment - thanks Priyanka. However, stackoverflow for some reason warns that answers in comments may be lost.)
Related
This is a very basic question. I can't simulate a PWM file, in system time, from its FPGA VI file.
Details
For a NI cRIO-9067 + LabVIEW 2016 + Windows 8 system, under FPGA Interface Mode, I have the Test VI No.1.vi NI LabVIEW file and the corresponding FPGA Desktop Execution Node block file Test VI No.1 DEN.vi as suggested in the Getting Started information [1] [2].
In both files, the Low Pulse and High Pulse Numeric Controls are filled with the 1000 value. The Loop Timer block is set as "mSec" Counter Unit and "32 Bit" Size of Internal Counter.
The compiled FPGA version of the first file executes a square wave changing each 1 second, as expected, after 7 minutes of local compilation.
Under Simulation (Simulated I/O) as Execution Mode, and for reproducing approximatedly and by trial and error the square wave timing every 1 second, I need to put the value 1750 in the Clock Ticks field, from the FPGA 40MHz Onboard Clock reference clock, shown in the block options.
I dont understand this block, and why i should not put any close divisor of 40,000,000 at the Clock Ticks field, or simply, the value 1. Basically i dont understand how to "time" these FPGA simulations.
The desktop execution node is designed for time based simulation you are definately on the right track.
What you are setting at the top is the number of cycles that are executed each time you call the node. In your case you have 1750 ticks so around 43.75us of simulated time per iteration.
To simulate in real time you need to make sure that you execute the same amount of simulated time as the simulation loop takes to run. In your case, you have no timing in your simulation loop so why 1750 works for you is because that is probably how long that loop takes to execute.
If you put a loop timer in of 1ms and set the clock ticks to 40,000 (1ms simulated time) then I think you will find that it also works.
In some cases it may be beneficial to execute faster than real time so you would just have to account for that in your maths. For example if you set the clock ticks to 40 (1us simulated time) then you can count the number of iterations and multiply by 1us to get the actual clock time.
How can I measure the time used by Matlab in the execution of some commands? I am looking for something like linux time command, which returns the actual CPU used (user + sys), instead of the total time transcurred, which can vary depending on system usage by other processes.
Also, the time should be returned in milliseconds.
You can use cputime:
CPUTIME CPU time in seconds.
CPUTIME returns the CPU time in seconds that has been used
by the MATLAB process since MATLAB started.
For example:
t=cputime; your_operation; cputime-t
returns the cpu time used to run your_operation.
The return value may overflow the internal representation
and wrap around.
There's also timeit.
And you can also use tic/ toc (see Daniel's answer).
See some useful comments on measuring CPU time. According to the above link, tic / toc and timeit give more accurate estimations than cputime.
For a simple clock time solution, use tic and toc
profile provides a detailed report code per line, you can choose between cpu time and real time.
I have noticed that the first time I run a script, it takes considerably more time than the second and third time1. The "warm-up" is mentioned in this question without an explanation.
Why does the code run faster after it is "warmed up"?
I don't clear all between calls2, but the input parameters change for every function call. Does anyone know why this is?
1. I have my license locally, so it's not a problem related to license checking.
2. Actually, the behavior doesn't change if I clear all.
One reason why it would run faster after the first time is that many things are initialized once, and their results are cached and reused the next time. For example in the M-side, variables can be defined as persistent in functions that can be locked. This can also occur on the MEX-side of things.
In addition many dependencies are loaded after the first time and remain so in memory to be re-used. This include M-functions, OOP classes, Java classes, MEX-functions, and so on. This applies to both builtin and user-defined ones.
For example issue the following command before and after running your script for the first run, then compare:
[M,X,C] = inmem('-completenames')
Note that clear all does not necessarily clear all of the above, not to mention locked functions...
Finally let us not forget the role of the accelerator. Instead of interpreting the M-code every time a function is invoked, it gets compiled into machine code instructions during runtime. JIT compilation occurs only for the first invocation, so ideally the efficiency of running object code the following times will overcome the overhead of re-interpreting the program every time it runs.
Matlab is interpreted. If you don't warm up the code, you will be losing a lot of time due to interpretation instead of the actual algorithm. This can skew results of timings significantly.
Running the code at least once will enable Matlab to actually compile appropriate code segments.
Besides Matlab-specific reasons like JIT-compilation, modern CPUs have large caches, branch predictors, and so on. Warming these up is an issue for benchmarking even in assembly language.
Also, more importantly, modern CPUs usually idle at low clock speed, and only jump to full speed after several milliseconds of sustained load.
Intel's Turbo feature gets even more funky: when power and thermal limits allow, the CPU can run faster than its sustainable max frequency. So the first ~20 seconds to 1 minute of your benchmark may run faster than the rest of it, if you aren't careful to control for these factors.
Another issue not mensioned by Amro and Marc is memory (pre)allocation.
If your script does not pre-allocate its memory it's first run would be very slow due to memory allocation. Once it completed its first iteration all memory is allocated, so consecutive invokations of the script would be more efficient.
An illustrative example
for ii = 1:1000
vec(ii) = ii; %// vec grows inside loop the first time this code is executed only
end
I am trying to run a real-time simulation in Simulink using Real-time Workshop. The target is grt(I have tried rtwin, but my simulation refuses to compile for it). I need the simulation to run in real-time so that one second in simulation lasts one second of real time. Grt ignores realtime and finishes the simulation in shortest time possible. Is there any way to synchronise it?
I have tried http://www.mathworks.com/matlabcentral/fileexchange/3175 but could not get it to work(does not compile).
Thank you for any suggestions.
Looks like it is impossible. I was able to slow down the execution by using Sleep(time in ms) function from WinApi and clock function from time.h, which looked quite good for low sample rates. However, when I increased the sample rate the Sleep function was sleeping for too long, which resulted in errors, with one second in simulation lasting more than one real world second.
The idea was to say that one period of iteration should last, let's say 200ms. Then time how long it takes for one iteration of code to execute using the clock function. Then call Sleep(200 - u), where u is the length of the iteration. The problem is that Sleep function sleeps the process and wakes it up when it wants to, not when you tell it to in the argument.
I know this is not a solution, but post this so that if anyone faces the same problem as me they won't try this dead-end solution. I had to rewrite the simulation for rtwin and now it works fine.
Another idea would be to somehow use interrupts, but I guess it would be quite complicated and not worth the trouble.
we are using LabVIEW Real-Time with the PXI-8110 Controller.
I am facing the following problem:
I have a loop with 500µs period time (time-loop) and no other task. I write the time each loop iteration into ram and then save the data afterwords.
It is necessary that the period is exact, but I see that it is 500µs with +/- 25 µs.
The clock for the timed loop is 1 MHz.
How is it possible to have 500µs - 25µs. I would understand if I get 500µs + xx µs when my compution is to heavy. But till now I just do an addition nothing more.
So does anyone have a clue what is going wrong?
I thought it would be possible to have resolution of 1µs as NI advertise (if the computation isn't so heavy).
Thanks.
You may need to check which thread the code is working in. An easier way to work is to use the Timed Loop as this will try and correct for overruns. Also pre-allocate the array that you are storing the data into and then replace array subset which each new value. You should see a massive improvement with this way.
If you display that value and are running in development mode you will see jitter +- time as you are reporting everything back to the host. Build the executable and again jitter will shrink.