This a general programming question. Let's say I have a thread doing a specific simulation, where speed is quite important. At every iteration I want to extract data from it and write it to a file.
Is it a better practice to hand over the data to a different thread and let the simulation thread focus on his job, or since speed is very important, make the simulation thread do the data recording too without any copying of data. (in my case it is 3-5 deques of integers with a size of 1000-10000)
Firstly it surely depends on how much data we are copying, but what else can it depend on? Can the cost of synchronization and copying be worth? Is it a good practice to create small runnables at each iteration to handle the recording task in case of 50 or more iterations per second?
If you truly want low latency on this stat capturing, and you want it during the simulation itself then two techniques come to mind. They can be used together very effectively. Please note that these two approaches are fairly far from the standard Java trodden path, so measure first and confirm that you need these techniques before abusing them; they can be difficult to implement correctly.
The fastest way to write the data to file during a simulation, without slowing down the simulation is to hand the work off to another thread. However care has to be taken on how the hand off occurs, as a memory barrier in the simulation thread will slow the simulation. Given the writer only cares that the values will come eventually I would consider using the memory barrier that sits behind AtomicLong.lazySet, it requests a thread safe write out to a memory address without blocking for the write to actually become visible to the other thread. Unfortunately direct access to this memory barrier is currently only availble via lazySet or via class sun.misc.Unsafe, which obviously is not part of the public Java API. However that should not be too large of a hurdle as it is on all current JVM implementations and Doug Lea is talking about moving parts of it into the mainstream.
To avoid the slow, blocking file IO that Java uses; make use of a memory mapped file. This lets the OS perform async IO for you on your behalf, and is very efficient. It also supports use of the same memory barrier mentioned above.
For examples of both techniques, I strongly recommend reading the source code to HFT Chronicle by Peter Lawrey. In fact, HFT Chronicle may be just the library for you to use here. It offers a highly efficient and simple to use disk backed queue that can sustain a million or so messages per second.
In my work on a stress-testing HTTP client I stored the stats into an array and, when the array was ready to send to the GUI, I would create a new array for the tester client and hand off the full array to the network layer. This means that you don't need to pay for any copying, just for the allocation of a fresh array (an ultra-fast operation on the JVM, involving hand-coded assembler macros to utilize the best SIMD instructions available for the task).
I would also suggest not throwing yourself head-on into the realms of optimal memory barrier usage; the difference between a plain volatile write and an AtomicReference.lazySet() can only be measurable if your thread does almost nothing else but excercise the memory barrier (at least millions of writes per second). Depending on your target I/O throughput, you may not even need NIO to meet the goal. Better try first with simple, easily maintainable code than dig elbows-deep into highly specialized APIs without a confirmed need for that.
Related
Given we have an application that is heavily polluted with concurrency constructs
multiple techniques are used (different people worked without clear architecture in mind),
multiple questionable locks that are there "just in case", thread safe queues. CPU usage is around 20%.
Now my goal is to optimize it such that it is making better use of caches and generally improve its performance and service time.
I'm considering to pin the parent process to a single core, remove all things that cause membars,
replace all thread safe data structures and replace all locks with some UnsafeReentrantLock
which would simply use normal reference field but take care of exclusive execution
needs...
I expect that we would end up with much more cache friendly application,
since we don't have rapid cache flushes all the time (no membars).
We would have less overhead since we dont need thread safe data structures,
volaties, atomics and replace all sorts of locks with I would assume that service time would improve also,
since we no longer synchronize on multiple thread safe queues...
Is there something that I'm overlooking here?
Maybe blocking operations would have to be paid attention to since they would not show up in that 20% usage?
Use case: a single data structure (hashtable, array, etc) whose members are accessed frequently by multiple threads and modified infrequently by those same threads. How do I maintain performance while guaranteeing thread safety (ie, preventing dirty reads).
Java: Concurrent version of the data structure (concurrent hashmap, Vector, etc).
Python: No need if only threads accessing it, because of GIL. If it's multiple processes that will be reading and updating the data structure, then use threading.Lock. Force the each process's code to acquire the lock before and release the lock after accessing the data structure.
Does that sound reasonable? Will Java's concurrent data structure impose too much penalty to read speed? Is there higher level concurrency mechanism in python?
Instead of reasoning about performance, I highly recommend to measure it for your application. Don't risk thread problems for a performance improvement that you most probably won't ever notice.
So: write thread-safe code without any performance-tricks, use a decent profiler to find the percentage of time spent inside the data structure access, and then decide if that part is worth any improvement.
I bet there will be other bottlenecks, not the shared data structure.
If you like, come back to us with your code and the profiler results.
Disclaimer: I don't know much about the theoretical background of CSP.
Since I read about it, I tend to structure most of my multi-threading "CSP-like", meaning I have threads waiting for jobs on a BlockingQueue.
This works very well and simplified my thinking about threading a lot.
What are the downsides of this approach?
Can you think of situations where I'm performance-wise better off with a synchronized block?
...or Atomics?
If I have many threads mostly sleeping/waiting, is there some kind of performance impact, except the memory they use? For example during scheduling?
This is one possibly way to designing the architecture of your code to prevent thread issues from even happening, this is however not the only one and sometimes not the best one.
First of all you obviously need to have a series of tasks that can be splitted and put into such a queue, which is not always the case if you for example have to calculate the result of a single yet very straining formula, which just cannot be taken apart to utilize multi-threading.
Then there is the issue if the task at hand is so tiny, that creating the task and adding it into the list is already more expensive than the task itself. Example: You need to set a boolean flag on many objects to true. Splittable, but the operation itself is not complex enough to justify a new Runnable for each boolean.
You can of course come up with solutions to work around this sometimes, for example the second example could be made reasonable for your approach by having each thread set 100 flags per execution, but then this is only a workaround.
You should imagine those ideas for threading as what they are: tools to help you solve your problem. So the concurrent framework and patters using those are all together nothing but a big toolbox, but each time you have a task at hand, you need to select one tool out of that box, because in the end putting in a screw with a hammer is possible, but probably not the best solution.
My recommendation to get more familiar with the tools is, that each time you have a problem that involves threading: go through the tools, select the one you think fits best, then experiment with it until you are satisfied that this specific tool fits the specific task best. Prototyping is - after all - another tool in the box. ;)
What are the downsides of this approach?
Not many. A queue may require more overhead than an uncontended lock - a lock of some sort is required internally by the queue classs to protect it from multiple access. Compared with the advantages of thread-pooling and queued comms in general, some extra overhead does not bother me much.
better off with a synchronized block?
Well, if you absolutely MUST share mutable data between threads :(
is there some kind of performance impact,
Not so anyone would notice. A not-ready thread is, effectively, an extra pointer entry in some container in the kernel, (eg. a queue belonging to a semaphore). Not worth bothering about.
You need synchronized blocks, Atomics, and volatiles whenever two or more threads access mutable data. Keep this to a minimum and it needn't affect your design. There are lots of Java API classes that can handle this for you, such as BlockingQueue.
However, you could get into trouble if the nature of your problem/solution is perverse enough. If your threads try to read/modify the same data at the same time, you'll find that most of your threads are waiting for locks and most of your cores are doing nothing. To improve response time you'll have to let a lot more threads run, perhaps forgetting about the queue and letting them all go.
It becomes a trade off. More threads chew up a lot of CPU time, which is okay if you've got it, and speed response time. Fewer threads use less CPU time for a given amount of work (but what will you do with the savings?) and slow your response time.
Key point: In this case you need a lot more running threads than you have cores to keep all your cores busy.
This sort of programming (multithreaded as opposed to parallel) is difficult and (irreproducible) bug prone, so you want to avoid it if you can before you even start to think about performance. Plus, it only helps noticably if you've got more than 2 free cores. And it's only needed for certain sorts of problems. But you did ask for downsides, and it might pay to know this is out there.
Currently, I'm running on a thread-less model that isn't working simply because I'm running out of memory before I can process the data I'm being handed. I've made all the changes that I can to optimize the code, and it's still just not quite quick enough.
Clearly I should move on to a threaded model. I'm wondering what the simplest, easiest way to do the following is:
The main thread passes some info to the worker
That worker performs some work that I'll refactor out of the main method
The workers will disappear and new ones will be instantiated when needed
I've never worked with java threading and from what I've read up on it seems pretty complicated, even if what I'm looking for seems pretty simple.
If you have multiple independent units of work of equal priority, the best solution is generally some sort of work queue, where a limited number of threads (the number chosen to optimize performance) sit in a while(true) loop dequeuing work units from the queue and executing them.
Generally the optimum number of threads is going to be the number of processors +/- 1, though in some cases a larger number will be optimal if the threads tend to get stalled by disk I/O requests or some such.
But keep in mind that tuning the entire system may be required. Eg, you may need more disk arms, and certainly more RAM may be required.
I'd start by having a read through Java Concurrency as refresher ;)
In particular, I would spend some time getting to know the Executors API as it will do most of what you've described without a lot of the overhead of dealing with to many locks ;)
Distributing the memory consumption to multiple threads will not change overall memory consumption. From what I read out of your question, I would like to step forward and tell you: Increase the heap of the Java engine, this will help. Looks like you have to optimize the Java startup parameters and not your code. If I am wrong, then you will have to buffer the data. To Disk! Not to a thread in the same memory model.
I recently inherited a small Java program that takes information from a large database, does some processing and produces a detailed image regarding the information. The original author wrote the code using a single thread, then later modified it to allow it to use multiple threads.
In the code he defines a constant;
// number of threads
public static final int THREADS = Runtime.getRuntime().availableProcessors();
Which then sets the number of threads that are used to create the image.
I understand his reasoning that the number of threads cannot be greater than the number of available processors, so set it the the amount to get the full potential out of the processor(s). Is this correct? or is there a better way to utilize the full potential of the processor(s)?
EDIT: To give some more clarification, The specific algorithm that is being threaded scales to the resolution of the picture being created, (1 thread per pixel). That is obviously not the best solution though. The work that this algorithm does is what takes all the time, and is wholly mathematical operations, there are no locks or other factors that will cause any given thread to sleep. I just want to maximize the programs CPU utilization to decrease the time to completion.
Threads are fine, but as others have noted, you have to be highly aware of your bottlenecks. Your algorithm sounds like it would be susceptible to cache contention between multiple CPUs - this is particularly nasty because it has the potential to hit the performance of all of your threads (normally you think of using multiple threads to continue processing while waiting for slow or high latency IO operations).
Cache contention is a very important aspect of using multi CPUs to process a highly parallelized algorithm: Make sure that you take your memory utilization into account. If you can construct your data objects so each thread has it's own memory that it is working on, you can greatly reduce cache contention between the CPUs. For example, it may be easier to have a big array of ints and have different threads working on different parts of that array - but in Java, the bounds checks on that array are going to be trying to access the same address in memory, which can cause a given CPU to have to reload data from L2 or L3 cache.
Splitting the data into it's own data structures, and configure those data structures so they are thread local (might even be more optimal to use ThreadLocal - that actually uses constructs in the OS that provide guarantees that the CPU can use to optimize cache.
The best piece of advice I can give you is test, test, test. Don't make assumptions about how CPUs will perform - there is a huge amount of magic going on in CPUs these days, often with counterintuitive results. Note also that the JIT runtime optimization will add an additional layer of complexity here (maybe good, maybe not).
On the one hand, you'd like to think Threads == CPU/Cores makes perfect sense. Why have a thread if there's nothing to run it?
The detail boils down to "what are the threads doing". A thread that's idle waiting for a network packet or a disk block is CPU time wasted.
If your threads are CPU heavy, then a 1:1 correlation makes some sense. If you have a single "read the DB" thread that feeds the other threads, and a single "Dump the data" thread and pulls data from the CPU threads and create output, those two could most likely easily share a CPU while the CPU heavy threads keep churning away.
The real answer, as with all sorts of things, is to measure it. Since the number is configurable (apparently), configure it! Run it with 1:1 threads to CPUs, 2:1, 1.5:1, whatever, and time the results. Fast one wins.
The number that your application needs; no more, and no less.
Obviously, if you're writing an application which contains some parallelisable algorithm, then you can probably start benchmarking to find a good balance in the number of threads, but bear in mind that hundreds of threads won't speed up any operation.
If your algorithm can't be parallelised, then no number of additional threads is going to help.
Yes, that's a perfectly reasonable approach. One thread per processor/core will maximize processing power and minimize context switching. I'd probably leave that as-is unless I found a problem via benchmarking/profiling.
One thing to note is that the JVM does not guarantee availableProcessors() will be constant, so technically, you should check it immediately before spawning your threads. I doubt that this value is likely to change at runtime on typical computers, though.
P.S. As others have pointed out, if your process is not CPU-bound, this approach is unlikely to be optimal. Since you say these threads are being used to generate images, though, I assume you are CPU bound.
number of processors is a good start; but if those threads do a lot of i/o, then might be better with more... or less.
first think of what are the resources available and what do you want to optimise (least time to finish, least impact to other tasks, etc). then do the math.
sometimes it could be better if you dedicate a thread or two to each i/o resource, and the others fight for CPU. the analisys is usually easier on these designs.
The benefit of using threads is to reduce wall-clock execution time of your program by allowing your program to work on a different part of the job while another part is waiting for something to happen (usually I/O). If your program is totally CPU bound adding threads will only slow it down. If it is fully or partially I/O bound, adding threads may help but there's a balance point to be struck between the overhead of adding threads and the additional work that will get accomplished. To make the number of threads equal to the number of processors will yield peak performance if the program is totally, or near-totally CPU-bound.
As with many questions with the word "should" in them, the answer is, "It depends". If you think you can get better performance, adjust the number of threads up or down and benchmark the application's performance. Also take into account any other factors that might influence the decision (if your application is eating 100% of the computer's available horsepower, the performance of other applications will be reduced).
This assumes that the multi-threaded code is written properly etc. If the original developer only had one CPU, he would never have had a chance to experience problems with poorly-written threading code. So you should probably test behaviour as well as performance when adjusting the number of threads.
By the way, you might want to consider allowing the number of threads to be configured at run time instead of compile time to make this whole process easier.
After seeing your edit, it's quite possible that one thread per CPU is as good as it gets. Your application seems quite parallelizable. If you have extra hardware you can use GridGain to grid-enable your app and have it run on multiple machines. That's probably about the only thing, beyond buying faster / more cores, that will speed it up.