I have a threaded application with timing blocks around all the sections of code I believe are performance intensive. These sections are long enough that the timing code can remain active even in final production runs. The application has a bunch of worker threads pulling work from a task pool.
In my current custom pool implementation, I compute two extra times for each worker thread: (1) the amount of idle time spent waiting on an empty pool for a job to arrive and (2) the amount of "missing" time not accounted for by either idling or a timing block (indicating a performance intensive region that I've missed).
I'm now considering switching my code over to Intel Thread Building Blocks. Is it possible to measure idle time when using TBB tasks? The sum of idle time and missing time can easily be computed by subtracting each timed section from the total wall clock interval, but getting at each separately requires special support (or perhaps a special idle task with low priority that spins trying to preempt itself?).
Note: I also asked this question at http://software.intel.com/en-us/forums/showthread.php?t=107203, and will synchronize any answers.
Related
I'm trying to understand why having too many threads can reduce CPU usage due to the increased overhead of context switching. An explanation that sounded plausible to me is that increasing # of threads also increases the frequency of context switches, meaning we end up spending more time context switching and less time doing useful work. Is this correct? Do individual time slices get compressed (with more context switches in between) as we have more threads to schedule?
Generally no. The primary mechanism for lower overhead is that if the scheduler picks the same thread to run on a core for two timeslices in a row, there is no context-switch overhead of stale caches and an FP save/restore.
A "tickless" kernel might set a timer farther in the future if there aren't any other tasks to schedule, instead of the traditional design of having a timer interrupt every 1 or 10 milliseconds where it always calls a scheduler function. (And if there aren't any waiting tasks, it can trivially decide to keep running this one.)
My main thread creates 8 worker threads (on a machine with a 4 core, 8 thread CPU), and then waits for them to complete with pthread_join(). The threads all exit successfully, and the pthread_join() successfully completes. However, I log the times that the threads exit and the time that pthread_join() completes for the last thread; the threads all exit essentially simultaneously (not surprising -- they are servicing a queue of work to be done), and the pthread_join() sometimes takes quite a long time to complete -- I have seen times in excess of 15 minutes after the last worker thread has exited!
More information: The worker threads are all set at the highest allowable round-robin scheduling priority (SCHED_RR); I have tried setting the main thread (waiting on the pthread_join()s) to the same thing and have also tried setting it to the highest SCHED_FIFO priority (where so far I have only seen it take as long as 27 seconds to complete; more testing is needed). My test is very CPU and memory intensive and takes about 90 -- 100 minutes to complete; during that time it is generally using all 8 threads at close to 100% capacity, and fairly quickly gets to where it is using about 90% of the 256 GB of RAM. This is running on a Linux (Fedora) OS at run level 3 (so no graphics or Window Manager -- essentially just a terminal -- because at the usual run level 5, a process using that much memory gets killed by the system).
An earlier version that took closer to 4 hours to complete (I have since made some performance improvements...) and in which I did not bother explicitly setting the priority of the main thread once took over an hour and 20 minutes for the pthread_join() to complete. I mention it because I don't really think that the main thread priority should be much of an issue -- there is essentially nothing else happening on the machine, it is not even on the network.
As I mentioned, all the threads complete with EXIT_SUCCESS. And in lighter weight tests, where the processing is over in seconds, I see no such delay. And so I am left suspecting that this is a scheduler issue. I know very little about the scheduler, but informally the impression I have is that here is this thread that has been waiting on a pthread_join() for well over an hour; perhaps the scheduler eventually shuffles it off to a queue of "very unlikely to require any processing time" tasks, and only checks it rarely.
Okay, eventually it completes. But ultimately, to get my work done, I have to run about 1000 of these, and some are likely to take a great deal longer than the 90 minutes or so that the case I have been testing takes. So I have to worry that the pthread_join() in those cases might delay even longer, and with 1000 iterations, those delays are going to add up to real time...
Thanks in advance for any suggestions.
In response to Nate's excellent questions and suggestions:
I have used top to spy on the process when it is in this state; all I can report is that it is using minimal CPU (maybe an occasional 2%, compared to the usual 700 - 800% that top reports for 8 threads running flat out, modulo some contention for locked resources). I am aware that top has all kinds of options I haven't investigated, and will look into how to run it to display information about the state of the main thread. (I see: I can use the -H option, and look in the S column... will do.) It is definitely not a matter of all the memory being swapped out -- my code is very careful to stay below the limit of physical memory, and does some disk I/O of its own to save and restore information that can't fit in memory. As a result little to no virtual memory is in use at any time.
I don't like my theory about the scheduler either... It's just the best I have been able to come up with so far...
As far as how I am determining when things happen: The exiting code does:
time_t now;
time(&now);
printf("Thread exiting, %s", ctime(&now));
pthread_exit(EXIT_SUCCESS);
and then the main thread does:
for (int i = 0; i < WORKER_THREADS; i++)
{
pthread_join(threads[i], NULL);
}
time(&now);
printf("Last worker thread has exited, %s", ctime(&now));
I like the idea of printing something each time pthread_join() returns, to see if we're waiting for the first thread to complete, the last thread to complete, or one in the middle, and will make that change.
A couple of other potentially relevant facts that have occurred to me since my original posting: I am using the GMP (GNU Multiprecision Arithmetic) library, which I can't really imagine matters; and I am also using a 3rd party (open source) library to create "canonical graphs," and that library, in order to be used in a multithreaded environment, does use some thread_local storage. I will have to dig into the particulars; still, it doesn't seem like cleaning that up should take any appreciable amount of time, especially without also using an appreciable amount of CPU.
This question is not a duplicate of any question related to why multithreading is not faster on single-core, read the rest to figure out what I actually want to know
As far as I know, multithreading is only faster on a CPU with multiple cores, since each thread can run in parallel. However, as my understanding of how preemption and multithreading on single-core works, it should also be faster. The image below can describe what I mean better. Consider that our app is a simple loop that takes exactly 4 seconds to execute. In this example, the time slice is constant, but, I don't think it makes any difference because, in the end, all threads with the same priority will get equal time by the scheduler. The first timeline is single-threaded, but the second one has 4 threads. The cycle also means when the preemption ends and the scheduler goes back to the queue of threads from start. I/O has also been removed since that just adds complexity and even if it changes the results, let's assume I'm talking about some code that does not require any sort of I/O.
The red threads are threads related to my process, and others (black) are the ones for other processes and apps
There are a couple of questions here:
Why isn't it faster? What's wrong with my timeline?
What's that cycle point called?
Since the time slice is not fixed, does that means the Cycle time is fixed, or the time slice gets calculated and the cycle will be as much time required to spend the calculated time slice on each thread?
Is the slice time based on time or instruction? I mean, is it like 0.1 sec for each thread or like 10 instructions for each thread?
The CPU utilization is based on CPU time, so why isn't it always on 100% because when a thread's time reaches, it moves to the next thread, and if a thread stops on I/O, it does not wait but executes the next one, so the CPU always tries to find a thread to execute and minimalize the time spent IDLE. Is the time for I/O so significant that more than 50% of CPU time is spent doing nothing because all threads are waiting for something, mostly I/O and the CPU time is elapsed waiting for a thread to become in a ready state?
Note: This timeline is simplified, the time spent on I/O, thread creation, etc. is not calculated and it's assumed that other threads do not finish before the end of the timeline and have the same priority/nice value as our process
I am observing strange effects with the CPU percentage as shown in e.g. top or htop on Linux (Ubuntu 16.04) for one special application. The application uses many threads (around 1000). Each thread has one computational task. About half of these tasks need to be computed once per "trigger" - the trigger is an external event received exactly every 100ms. The other threads are mostly sleeping (waiting for user interaction) and hence do not play a big role here. So to summarise: many threads are waking up basically simultaneously within a short period of time, doing there (relatively short) computation and going back to sleep again.
Since the machine running this application has 8 virtual CPUs (4 cores each 2 threads, it's an i7-3612QE), only 8 threads can really wake up at a time, so many threads will have to wait. Also some of these tasks have interdependencies, so they anyway have to wait, but I think as an approximation one can think of this application as a bunch of threads going to the runnable state at the same time every 100ms and each doing only a short computation (way below 1ms of CPU time each).
Now coming to the strange effect: If I look at the CPU percentage in "top", it shows something like 250%. As far as I know, top looks on the CPU time (user + system) the kernel accounts for this process, so 250% would mean the process uses 3 virtual CPUs on average. So far so good. Now, if I use taskset to force the entire process to use only a single virtual CPU, the CPU percentage drops to 80%. The application has internal accounting which tells me that still all data is being processed. So the application is doing the same amount of work, but it seemingly uses less CPU resources. How can that be? Can I really trust the kernel CPU time accounting, or is this an artefact of the measurement?
The CPU percentage also goes down, if I start other processes which take a lot of CPU, even if the do nothing ("while(true);") and are running at low priority (nice). If I launch 8 of these CPU-eating processes, the application reaches again 80%. With fewer CPU-eaters, I get gradually higher CPU%.
Not sure if this plays a role: I have used the profiler vtune, which tells me my application is actually quite inefficient (only about 1 IPC), mostly because it's memory bound. This does not change if I restrict the process to a single virtual CPU, so I assume the effect is not caused by a huge increase in efficiency when running everything on the same core (which would be strange anyway).
My question was essentially already answered by myself in the last paragraph: The process is memory bound. Hence not the CPU is the limited resource but the memory bandwidth. Allowing such process to run on multiple CPU cores in parallel will mainly have the effect that more CPU cores are waiting for data to arrive from RAM. This is counted as CPU load, since the CPU is executing the thread, but just quite slowly. All my other observations go along with this.
Every process has at least one thread of execution and I read somewhere that modern Operating Systems only schedule Thread and not process.
So if there are two processes running in the system - P1 with 1 thread and P2 with 100 threads, how will OS scheduling algorithm ensure that both P1 and P2 get approximately same amount of CPU time? If OS blindly schedules threads, P2 will get 100 times more CPU time than P1.
Does it also take into account which Process a particular thread belong to? Otherwise, it seems too easy for a process to hog all the CPU by creating more threads.
Does it also take into account which Process a particular thread belong to? Otherwise, it seems too easy for a process to hog all the CPU by creating more threads.
Wrong question. Consider two jobs that are trying to solve the exact same problem by doing the same work and are perfectly identical except for one thing -- one uses dozens of threads, the other uses dozens of processes. Why should the one that uses dozens of processes get more CPU time than the one that uses dozens of threads?
Your notion of fairness is not really a sensible one.
Instead, scheduling is more designed around trying to get as much work done as possible per unit time. The assumption is that everything the computer is doing is useful and it benefits competing tasks to have other tasks competing with them finish as quickly as possible too.
This is actually all you need the vast majority of the time. But occasionally you have special situations where this doesn't work. One is ultra-high-priority tasks like keeping video or audio flowing or keeping a user interface responsive. Another is ultra-low-priority tasks where there's an enormous amount of work you want done and you don't want the system to be slow for a long time while you're working on it. Priorities are used for this, and generally the system allows higher-priority threads to interrupt lower-priority ones to keep responsiveness.
In general, "fair thread scheduling" attempts to give each thread an equal amount of CPU time (regardless of how much CPU time all threads in a process get); and "fair process scheduling" attempts to give each process the same amount of CPU time (e.g. by giving threads belonging to different processes unequal amounts of CPU time). These are mutually exclusive - you can't have both (unless each process has the same number of threads).
Note that it's all a broken joke anyway. For example, if one thread gets 10 ms of time on a CPU that is running slow due to thermal throttling (and/or because another logical CPU in the same core is busy) and another thread gets 10 ms of time on a CPU that is running faster than normal (e.g. due to "turbo-boost" and/or because the other logical CPU in the core is not being used); then these threads have received an equal amount of CPU time but have not received anything that could be considered "fair" (because one thread might be able to get 20 times as much work done than the other).
Note that it's all unwanted anyway. For example, for a good OS threads would be given a priority to indicate how important the work they do is, and you don't want a high priority thread (doing very important work) to get the same "fair share" of CPU time as a low priority thread (doing irrelevant/unimportant work). For cases where two threads have equal priority you might (in theory) want them to get an "equal" amount of CPU time; but in practice this isn't common and threads block and unblock so often that it isn't worth caring about; and in practice it can lead to "two half finished jobs instead of one completed job and one unstarted job" scenarios that increases the average amount of time a job (e.g. request for work) takes to complete.
If the thread is the basic unit of scheduling (a generally safe assumption these days) then the process scheduler is the one to decide who to allocate the CPUs. How (and whether) it takes thread usage into account is entirely system specific. AND the behavior ma depends upon the type of process. For example, in VMS (and adopted in Windoze) realtime processes are treated differently than other types of processes.
In the VMS-type scheduling, a process with more threads gets more CPU by design. Better for an application to use more threads and for it to use more processes.
Keep in mind that a system may impose limits on the number of threads in a process.