Which type of artificial neural network would you suggest to be able to make continuous time-dependant signal predictions? It should predict smallscale steps over very few signals up to very large scale steps with very many signals, possibly with less precision (abstraction by some kind of hierarchy?).
See:
Actually the system should learn and predict simultaneously.
I think a Spiking Neural Network, which is "third generation" and most similar to the neurons in our brain would do best.
It runs in real time, although I don't think it can learn in real-time.
Instead, you can just continually examine and overhaul it, running a couple seconds behind live input so you can adjust its answers, before it becomes good enough to go real-time.
Related
I have written a CUDA program which already gets a speedup compared to a serial version of 40 (2600k vs GTX 780). Now I am thinking about using several streams for running several kernels parallel. Now my questions are: How can I measure the free resources on my GPU (because if I have no free resources on my GPU the use of streams would make no sense, am I right?), and in which case does the use of streams make sense?
If asked I can provide my code of course, but at the moment I think that it is not needed for the question.
Running kernels concurrently will only happen if the resources are available for it. A single kernel call that "uses up" the GPU will prevent other kernels from executing in a meaningful way, as you've already indicated, until that kernel has finished executing.
The key resources to think about initially are SMs, registers, shared memory, and threads. Most of these are also related to occupancy, so studying occupancy (both theoretical, i.e. occupancy calculator, as well as measured) of your existing kernels will give you a good overall view of opportunities for additional benefit through concurrent kernels.
In my opinion, concurrent kernels is only likely to show much overall benefit in your application if you are launching a large number of very small kernels, i.e. kernels that encompass only one or a small number of threadblocks, and which make very limited use of shared memory, registers, and other resources.
The best optimization approach (in my opinion) is analysis-driven optimization. This tends to avoid premature or possibly misguided optimization strategies, such as "I heard about concurrent kernels, I wonder if I can make my code run faster with it?" Analysis driven optimization starts out by asking basic utilization questions, using the profiler to answer those questions, and then focusing your optimization effort at improving metrics, such as memory utilization or compute utilization. Concurrent kernels, or various other techniques are some of the strategies you might use to address the findings from profiling your code.
You can get started with analysis-driven optimization with presentations such as this one.
If you specified no stream, the stream 0 is used. According to wikipedia (you may also find it in the cudaDeviceProp structure), your GTX 780 GPU has 12 streaming multiprocessors which means there could be an improvement if you use multiple streams. The asyncEngineCount property will tell you how many concurrent asynchronous memory copies can run.
The idea of using streams is to use an asyncmemcopy engine (aka DMA engine) to overlap kernel executions and device2host transfers. The number of streams you should use for best performance is hard to guess because it depends on the number of DMA engines you have, the number of SMs and the balance between synchronizations/amount of concurrency. To get an idea you can read this presentation (for instance slides 5,6 explain the idea very well).
Edit: I agree that using a profiler is needed as a first step.
At work we're discussing the design of a new platform and one of the upper management types said it needed to run our current code base (C on Linux) but be real time because it needed to respond in less than a second to various inputs. I pointed out that:
That point doesn't mean it needs to be "real time" just that it needs a faster clock and more streamlining in its interrupt handling
One of the key points to consider is the OS that's being used. They wanted to stick with embedded Linux, I pointed out we need an RTOS. Using Linux will prevent "real time" because of the kernel/user space memory split thus I/O is done via files and sockets which introduce a delay
What we really need to determine is if it needs to be deterministic (needs to respond to input in <200ms 90% of the time for example).
Really in my mind if point 3 is true, then it needs to be a real time system, and then point 2 is the biggest consideration.
I felt confident answering, but then I was thinking about it later... What do others think? Am I on the right track here or am I missing something?
Is there any difference that I'm missing between a "real time" system and one that is just "deterministic"? And besides a RTC and a RTOS, am I missing anything major that is required to execute a true real time system?
Look forward to some great responses!
EDIT:
Got some good responses so far, looks like there's a little curiosity about my system and requirements so I'll add a few notes for those who are interested:
My company sells units in the 10s of thousands, so I don't want to go over kill on the price
Typically we sell a main processor board and an independent display. There's also an attached network of other CAN devices.
The board (currently) runs the devices and also acts as a webserver sending basic XML docs to the display for end users
The requirements come in here where management wants the display to be updated "quickly" (<1s), however the true constraints IMO come from the devices that can be attached over CAN. These devices are frequently motor controlled devices with requirements including "must respond in less than 200ms".
You need to distinguish between:
Hard realtime: there is an absolute limit on response time that must not be breached (counts as a failure) - e.g. this is appropriate for example when you are controlling robotic motors or medical devices where failure to meet a deadline could be catastrophic
Soft realtime: there is a requirement to respond quickly most of the time (perhaps 99.99%+), but it is acceptable for the time limit to be occasionally breached providing the response on average is very fast. e.g. this is appropriate when performing realtime animation in a computer game - missing a deadline might cause a skipped frame but won't fundamentally ruin the gaming experience
Soft realtime is readily achievable in most systems as long as you have adequate hardware and pay sufficient attention to identifying and optimising the bottlenecks. With some tuning, it's even possible to achieve in systems that have non-deterministic pauses (e.g. the garbage collection in Java).
Hard realtime requires dedicated OS support (to guarantee scheduling) and deterministic algorithms (so that once scheduled, a task is guaranteed to complete within the deadline). Getting this right is hard and requires careful design over the entire hardware/software stack.
It is important to note that most business apps don't require either: in particular I think that targeting a <1sec response time is far away from what most people would consider a "realtime" requirement. Having said that, if a response time is explicitly specified in the requirements then you can regard it as soft realtime with a fairly loose deadline.
From the definition of the real-time tag:
A task is real-time when the timeliness of the activities' completion is a functional requirement and correctness condition, rather than merely a performance metric. A real-time system is one where some (though perhaps not all) of the tasks are real-time tasks.
In other words, if something bad will happen if your system responds too slowly to meet a deadline, the system needs to be real-time and you will need a RTOS.
A real-time system does not need to be deterministic: if the response time randomly varies between 50ms and 150ms but the response time never exceeds 150ms then the system is non-deterministic but it is still real-time.
Maybe you could try to use RTLinux or RTAI if you have sufficient time to experiment with. With this, you can keep the non realtime applications on the linux, but the realtime applications will be moved to the RTOS part. In that case, you will(might) achieve <1second response time.
The advantages are -
Large amount of code can be re-used
You can manually partition realtime and non-realtime tasks and try to achieve the response <1s as you desire.
I think migration time will not be very high, since most of the code will be in linux
Just on a sidenote be careful about the hardware drivers that you might need to run on the realtime part.
The following architecture of RTLinux might help you to understand how this can be possible.
It sounds like you're on the right track with the RTOS. Different RTOSs prioritize different things either robustness or speed or something. You will need to figure out if you need a hard or soft RTOS and based on what you need, how your scheduler is going to be driven. One thing is for sure, there is a serious difference betweeen using a regular OS and a RTOS.
Note: perhaps for the truest real time system you will need hard event based resolution so that you can guarantee that your processes will execute when you expect them too.
RTOS or real-time operating system is designed for embedded applications. In a multitasking system, which handles critical applications operating systems must be
1.deterministic in memory allocation,
2.should allow CPU time to different threads, task, process,
3.kernel must be non-preemptive which means context switch must happen only after the end of task execution. etc
SO normal windows or Linux cannot be used.
example of RTOS in an embedded system: satellites, formula 1 cars, CAR navigation system.
Embedded System: System which is designed to perform a single or few dedicated functions.
The system with RTOS: also can be an embedded system but naturally RTOS will be used in the real-time system which will need to perform many functions.
Real-time System: System which can provide the output in a definite/predicted amount of time. this does not mean the real-time systems are faster.
Difference between both :
1.normal Embedded systems are not Real-Time System
2. Systems with RTOS are real-time systems.
I am starting to learn OpenMP, running examples (with gcc 4.3) from https://computing.llnl.gov/tutorials/openMP/exercise.html in a cluster. All the examples work fine, but I have some questions:
How do I know in which nodes (or cores of each node) have the different threads been "run"?
Case of nodes, what is the average transfer time in microsecs or nanosecs for sending the info and getting it back?
What are the best tools for debugging OpenMP programs?
Best advices for speeding up real programs?
Typically your OpenMP program does not know, nor does it care, on which cores it is running. If you have a job management system that may provide the information you want in its log files. Failing that, you could probably insert calls to the environment inside your threads and check the value of some environment variable. What that is called and how you do this is platform dependent, I'll leave figuring it out up to you.
How the heck should I (or any other SOer) know ? For an educated guess you'd have to tell us a lot more about your hardware, o/s, run-time system, etc, etc, etc. The best answer to the question is the one you determine from your own measurements. I fear that you may also be mistaken in thinking that information is sent around the computer -- in shared-memory programming variables usually stay in one place (or at least you should think about them staying in one place the reality may be a lot messier but also impossible to discern) and is not sent or received.
Parallel debuggers such as TotalView or DDT are probably the best tools. I haven't yet used Intel's debugger's parallel capabilities but they look promising. I'll leave it to less well-funded programmers than me to recommend FOSS options, but they are out there.
i) Select the fastest parallel algorithm for your problem. This is not necessarily the fastest serial algorithm made parallel.
ii) Test and measure. You can't optimise without data so you have to profile the program and understand where the performance bottlenecks are. Don't believe any advice along the lines that 'X is faster than Y'. Such statements are usually based on very narrow, and often out-dated, cases and have become, in the minds of their promoters, 'truths'. It's almost always possible to find counter-examples. It's YOUR code YOU want to make faster, there's no substitute for YOUR investigations.
iii) Know your compiler inside out. The rate of return (measured in code speed improvements) on the time you spent adjusting compilation options is far higher than the rate of return from modifying the code 'by hand'.
iv) One of the 'truths' that I cling to is that compilers are not terrifically good at optimising for use of the memory hierarchy on current processor architectures. This is one area where code modification may well be worthwhile, but you won't know this until you've profiled your code.
You cannot know, the partition of threads on different cores is handled entirely by the OS. You speaking about nodes, but OpenMP is a multi-thread (and not multi-process) parallelization that allow parallelization for one machine containing several cores. If you need parallelization across different machines you have to use a multi-process system like OpenMPI.
The order of magnitude of communication times are :
huge in case of communications between cores inside the same CPU, it can be considered as instantaneous
~10 GB/s for communications between two CPU across a motherboard
~100-1000 MB/s for network communications between nodes, depending of the hardware
All the theoretical speeds should be specified in your hardware specifications. You should also do little benchmarks to know what you will really have.
For OpenMP, gdb do the job well, even with many threads.
I work in extreme physics simulation on supercomputer, here are our daily aims :
use as less communication as possible between the threads/processes, 99% of the time it is communications that kill performances in parallel jobs
split the tasks optimally, machine load should be as close as possible to 100% all the time
test, tune, re-test, re-tune... . Parallelization is not at all a generic "miracle solution", it generally needs some practical work to be efficient.
I might be mixing apples and oranges in this question since I'm noob in mentioned areas, so please try to understand what I mean.
I read that WPF (and Silverlight, as an extension to it) uses GPU acceleration to render itself on the screen.
If so, what happens when I have GPU-intensive application already running (say, game in windowed mode), and then run WPF/Silverlight application, which needs some GPU time/power? Who gets greater priority? Do they both start "stuttering"?
Thanks!
I think it would be safe to say that you'd have similar issues as a CPU intensive task running...it will starve other tasks of processing time. So yes, you may see lowered performance all around, but it all depends on whatever is involved :).
Generally speaking, intensive games are usually written to use most/all the computers resources, expecting that this is the only thing a user will probably be doing. Other, nicer applications will lower their need for processing (ie lower FPS) when the window is not in the foreground. Windows itself actually will give more priority to the foreground window (unless configured otherwise), but you never know what other programs do to their own thread priorities.
I am a beginning C programmer (though not a beginning programmer) looking to dive into a project to teach myself C. My project is music-based, and because of this I am curious whether there are any 'best practices' per-se, when it comes to timing functions.
Just to clarify, my project is pretty much an attempt to build some barebones music notation/composition software (remember, emphasis on barebones). I was originally thinking about using OSX as my platform, but I want to do it in C, not Obj-C (though I know it would probably be easier...CoreAudio looked like a pretty powerful tool for this kind of stuff). So even though I don't have to build OSX apps in Obj-C, I will probably end up building this on a linux system (probably debian...).
Thanks everyone, for your great answers.
There are two accurate methods for timing functions:
Single process execution.
Timer event handler / callback
Single Process Execution
Most modern computers execute more than one program simultaneously. Actually, they execute pieces of many programs, swapping them out based on priorities and other metrics to look like more than one program is executing at the same time. This overhead effects timing in programs. Either the program gets delayed in reading the time or the OS gets delayed in setting its own time variables.
The solution in this case is to eliminate as many tasks from running. The ideal environment is for best accuracy is to have your program as the sole program running. Some OSes provide API for superuser applications to block all other programs or kill them.
Timer event handling / callback
Since the OS can't be trusted to execute your program with high precision, most OS's will provide Timer APIs. Many of these APIs include the ability to call one of your functions when the timer expires. This is known as a callback function. Other OS's may send a message or generate an event when the timer expires. These fall under the class of timer handlers. The callback process has less overhead than the handlers and thus is more accurate.
Music Hardware
Although you may have your program send music to the speakers, many computers now have separate processors that play music. This frees up the main processor and provides more continuous notes, rather than sounds separated by silent gaps due to platform overhead of your program send the next sounds to the speaker.
A quality music processor has at least these to functions:
Start Playing
End Music Notification
Start Playing
This is the function where you tell the music processor where your data is and the size of the data. The processor will start playing the music.
End Music Notification
You provide the processor with a pointer to a function that it will call when the music data has been processed. Nice processors will call the function early so there will be no gaps in the sounds while reloading.
All of this is platform dependent and may not be standard across platforms.
Hope this helps.
This is quite a vast area, and, depending on exactly what you want to do, potentially very difficult.
You don't give much away by saying your project is "music based".
Is it a musical score typesetting program?
Is it processing audio?
Is it filtering MIDI data?
Is it sequencing MIDI data?
Is it generating audio from MIDI data
Does it only perform playback?
Does it need to operate in a real time environment?
Your question though hints at real time operation, so in that case...
The general rule when working in a real time environment is don't do anything which may block the real time thread. This includes:
Calling free/malloc/calloc/etc (dynamic memory allocation/deallocation).
File I/O.any
Use of spinlocks/semaphores/mutexes upon threads.
Calls to GUI code.
Calls to printf.
Bearing these considerations in mind for a real time music application, you're going to have to learn how to do multi-threading in C and how to pass data from the UI/GUI thread to the real time thread WITHOUT breaking ANY of the above restrictions.
For an open source real time audio (and MIDI) (routing) server take a look at http://jackaudio.org
gettimeofday() is the best for wall clock time. getrusage() is the best for CPU time, although it may not be portable. clock() is more portable for CPU timing, but it may have integer overflow.
This is pretty system-dependent. What OS are you using?
You can take a look at gettimeofday() for fairly high granularity. It should work ok if you just need to read time once in awhile.
SIGALRM/setitimer can be used to receive an interrupt periodically. Additionally, some systems have higher level libraries for dealing with time.