I am writing a multi-thread application in Linux.
There is no RT patch in kernel, yet I use threads with priorities.
On checking the time it takes to execute printf , I measure different values every time I measure, although it is done in the highest priority thread :
if(clock_gettime(CLOCK_MONOTONIC, &start))
{ /* handle error */
}
for(int i=0; i< 1000; i++)
printf("hello world");
if(clock_gettime(CLOCK_MONOTONIC, &end))
{
/* handle error */
}
elapsedSeconds = TimeSpecToSeconds(&end) - TimeSpecToSeconds(&start);
Why does printf change timing and in non deterministic way , i.e. each
How should printf be used with RT threads ?
Can it be used inside RT thread or should it be totally avoided ?
Is write to disk should be treated in the same way as printf ? Should it be used only in separate low priority thread ?
printf under the hood triggers a non-realtime (even blocking) mechanism of the buffered IO.
It's not only non-deterministic, but opens the possibility of a priority inversion.
You should be very careful using it from a real time thread (I would say totally avoid it.
Normally, in a latency bound code you would use a wait-free binary audit into a chain of (pre-allocated or memory mapped) ring buffers and flush them using a background lower priority thread (or even a separate process).
Related
Ok, let me try to restate the 2 questions:
Does OS actively preempt a thread as soon as it starts blocking, and never return to the thread until blocking is done? I feel that the OS has the information about disk IO and network IO so it should have enough information to do so.
If the OS can eliminate the CPU idle time by switching to another thread, do we really need asynchronous programming?
Therefore, even though the thread is blocking, the CPU is not blocking
but running other threads.
Yes, that's correct.
If my understanding above is correct, what is the value of
asynchronous programming?
It's useful if you want your program to make progress on multiple tasks simultaneously.
Of course you could get the same effect by explicitly spawning multiple threads yourself, and having each of them work on a separate task (including any blocking calls), and that would work as well, but then you have to explicitly manage all those threads which can be a bit of a pain to get right. (inter-thread synchronization/communication can be tricky, and in particular the case where you want to cancel an operation is difficult to implement well if one or more of your threads is blocked inside a blocking I/O call and thus can't be easily persuaded to exit quickly -- then your other threads may have to wait a long time, possibly forever, before they can join() that thread and terminate safely)
The re-invigoration of asynchronous programming has little to do with OS/kernel architecture to date. OSes have blocked the way you describe since the 1960s; although the reason has changed somewhat.
In early systems, efficiency was measured by useful CPU cycles; so switching tasks when one was blocked was a natural act.
Modern systems architecture is frequently addressing how to keep the CPUs occupied; for example if there are 800 CPUs but 20 tasks; then 780 CPUs have nothing to do.
As a concrete example, a program to sum all the bytes of a file might look a bit like:
for (count=0; (c = getchar()) != EOF; count += c) {}
A multi-threaded version, for performance increase might look like:
for (n=0; n < NCPU; n++) {
if (threadfork() == 0) {
offset_t o = n* (file_size / NCPU);
offset_t l = (file_size / NCPU);
for (count = 0; l-- && pread(fd, &c, 1, o) == 1; count += c) {}
threadexit(count);
}
}
for (n=0; n < NCPU; n++) {
threadwait(&temp);
total += temp;
}
return total;
which is a bit grim, both because it is complex, and probably has inconsistent speed-ups.
In comparison the function:
int GetSum(char *data, int len) {
int count = 0;
while (len--) {
count += *data++;
}
return count;
}
I could construct a sort of dispatcher which, when a lump of file data became available in ram, invoked GetSum() on it, queuing its return value for later accumulation. This dispatcher could invest in familiarity with optimal i/o patterns etc.. since it may be applicable to many problems; and the programmer has a considerably simpler job to do.
Even without that sort of native support; I could mmap(2) the file, then dispatch many threads to just touch a page, then invoke GetSum on that page. This would effectively emulate an asynchronous model in a plain old unix-y framework.
Of course nothing is quite that easy; even a progress bar in a dispatch-oriented asynchronous model is dubious at best (not that the 1950s- based sequential ones were anything to write home about ). Communicating errors is also cumbersome; and because you use asynch to direct maximum cpu resources at yourself, you need to minimize synchronization operations (duh, async :).
Async has a lot of possibilities; but it really needs languages with defacto async support, not as an aspirational nod from the latest du jour standard of some rickety language.
I want to calculate the context switch time and I am thinking to use mutex and conditional variables to signal between 2 threads so that only one thread runs at a time. I can use CLOCK_MONOTONIC to measure the entire execution time and CLOCK_THREAD_CPUTIME_ID to measure how long each thread runs.
Then the context switch time is the (total_time - thread_1_time - thread_2_time).
To get a more accurate result, I can just loop over it and take the average.
Is this a correct way to approximate the context switch time? I cant think of anything that might go wrong but I am getting answers that are under 1 nanosecond..
I forgot to mention that the more time I loop it over and take the average, the smaller results I get.
Edit
here is a snippet of the code that I have
typedef struct
{
struct timespec start;
struct timespec end;
}thread_time;
...
// each thread function looks similar like this
void* thread_1_func(void* time)
{
thread_time* thread_time = (thread_time*) time;
clock_gettime(CLOCK_THREAD_CPUTIME_ID, &(thread_time->start));
for(x = 0; x < loop; ++x)
{
//where it switches to another thread
}
clock_gettime(CLOCK_THREAD_CPUTIME_ID, &(thread_time->end));
return NULL;
};
void* thread_2_func(void* time)
{
//similar as above
}
int main()
{
...
pthread_t thread_1;
pthread_t thread_2;
thread_time thread_1_time;
thread_time thread_2_time;
struct timespec start, end;
// stamps the start time
clock_gettime(CLOCK_MONOTONIC, &start);
// create two threads with the time structs as the arguments
pthread_create(&thread_1, NULL, &thread_1_func, (void*) &thread_1_time);
pthread_create(&thread_2, NULL, &thread_2_func, (void*) &thread_2_time);
// waits for the two threads to terminate
pthread_join(thread_1, NULL);
pthread_join(thread_2, NULL);
// stamps the end time
clock_gettime(CLOCK_MONOTONIC, &end);
// then I calculate the difference between between total execution time and the total execution time of two different threads..
}
First of all, using CLOCK_THREAD_CPUTIME_ID is probably very wrong; this clock will give the time spent in that thread, in user mode. However the context switch does not happen in user mode, You'd want to use another clock. Also, on multiprocessing systems the clocks can give different values from processor to another! Thus I suggest you use CLOCK_REALTIME or CLOCK_MONOTONIC instead. However be warned that even if you read either of these twice in rapid succession, the timestamps usually will tens of nanoseconds apart already.
As for context switches - tthere are many kinds of context switches. The fastest approach is to switch from one thread to another entirely in software. This just means that you push the old registers on stack, set task switched flag so that SSE/FP registers will be lazily saved, save stack pointer, load new stack pointer and return from that function - since the other thread had done the same, the return from that function happens in another thread.
This thread to thread switch is quite fast, its overhead is about the same as for any system call. Switching from one process to another is much slower: this is because the user-space page tables must be flushed and switched by setting the CR0 register; this causes misses in TLB, which maps virtual addresses to physical ones.
However the <1 ns context switch/system call overhead does not really seem plausible - it is very probable that there is either hyperthreading or 2 CPU cores here, so I suggest that you set the CPU affinity on that process so that Linux only ever runs it on say the first CPU core:
#include <sched.h>
cpu_set_t mask;
CPU_ZERO(&mask);
CPU_SET(0, &mask);
result = sched_setaffinity(0, sizeof(mask), &mask);
Then you should be pretty sure that the time you're measuring comes from a real context switch. Also, to measure the time for switching floating point / SSE stacks (this happens lazily), you should have some floating point variables and do calculations on them prior to context switch, then add say .1 to some volatile floating point variable after the context switch to see if it has an effect on the switching time.
This is not straight forward but as usual someone has already done a lot of work on this. (I'm not including the source here because I cannot see any License mentioned)
https://github.com/tsuna/contextswitch/blob/master/timetctxsw.c
If you copy that file to a linux machine as (context_switch_time.c) you can compile and run it using this
gcc -D_GNU_SOURCE -Wall -O3 -std=c11 -lpthread context_switch_time.c
./a.out
I got the following result on a small VM
2000000 thread context switches in 2178645536ns (1089.3ns/ctxsw)
This question has come up before... for Linux you can find some material here.
Write a C program to measure time spent in context switch in Linux OS
Note, while the user was running the test in the above link they were also hammering the machine with games and compiling which is why the context switches were taking a long time. Some more info here...
how can you measure the time spent in a context switch under java platform
Let's say I want to programmatically pin the current process to a single CPU, but I don't care which CPU that is.
One easy way to use sched_setaffinity with a fixed CPU number, probably 0 since there should always be a "CPU 0"1.
However, this approach fails if the affinity of the process has been set to a subset of the existing CPUs, not including the one you picked, e.g., by launching it with taskset.
So I want to pick "any CPU" to pin to, but only out of the CPUs that the current affinity mask allows. Here's one approach:
cpu_set_t cpu_set;
if (sched_getaffinity(0, sizeof(cpu_set), &cpu_set)) {
err("failed while getting existing cpu affinity");
}
for (int cpu = 0; cpu < CPU_SETSIZE; cpu++) {
if (CPU_ISSET(cpu, &cpu_set)) {
CPU_ZERO(cpu_set);
CPU_SET(cpu, &cpu_set);
}
}
int result = sched_setaffinity(0, sizeof(cpu_set), &cpu_set);
Basically we get the current affinity mask, then loop over every possible CPU looking for the first one that is allowed, then pass a mask with only this CPU set to sched_setaffinity.
However, if the current affinity mask has changed between the get and set calls the set call will fail. Any way around this race condition?
1 Although CPU zero won't always be online.
You could use getcpu() to discover the cpu that your process is running within, and use the result to set affinity to that cpu:
unsigned mycpu=0;
if( -1 == getcpu(&mycpu,NULL,NULL) ) {
// handle error
}
Presumably any CPU affinity rules that are in place would be honored by the scheduler, thus the getcpu() call would return a CPU that the process is allowed to run on.
There's still the potential that the affinity set might change, but that seems like a very unlikely case, and the allowed CPUs might be affected at some point in the future, outside the control of the process in question.
I suppose you could detect the error in the sched_setaffinity() call, and retry the process until the setaffinity call works...
Considering that the affinity mask of the process can change at any moment, you can iteratively try to pin the process to the current CPU and stop when it is successful.
cpu_set_t cpu_set;
int cpu = 0;
int result = -1;
while (result<0){
cpu = sched_getcpu();
if (cpu>0){
CPU_ZERO(&cpu_set);
CPU_SET(cpu, &cpu_set);
result = sched_setaffinity(0, sizeof(cpu_set), &cpu_set);
}
}
I just wrote my first OpenMP program that parallelizes a simple for loop. I ran the code on my dual core machine and saw some speed up when going from 1 thread to 2 threads. However, I ran the same code on a school linux server and saw no speed-up. After trying different things, I finally realized that removing some useless printf statements caused the code to have significant speed-up. Below is the main part of the code that I parallelized:
#pragma omp parallel for private(i)
for(i = 2; i <= n; i++)
{
printf("useless statement");
prime[i-2] = is_prime(i);
}
I guess that the implementation of printf has significant overhead that OpenMP must be duplicating with each thread. What causes this overhead and why can OpenMP not overcome it?
Speculating, but maybe the stdout is guarded by a lock?
In general, printf is an expensive operation because it interacts with other resources (such as files, the console and such).
My empirical experience is that printf is very slow on a Windows console, comparably much faster on Linux console but fastest still if redirected to a file or /dev/null.
I've found that printf-debugging can seriously impact the performance of my apps, and I use it sparingly.
Try running your application redirected to a file or to /dev/null to see if this has any appreciable impact; this will help narrow down where the problem lays.
Of course, if the printfs are useless, why are they in the loop at all?
To expand a bit on #Will's answer ...
I don't know whether stdout is guarded by a lock, but I'm pretty sure that writing to it is serialised at some point in the software stack. With the printf statements included OP is probably timing the execution of a lot of serial writes to stdout, not the parallelised execution of the loop.
I suggest OP modifies the printf statement to include i, see what happens.
As for the apparent speed-up on the dual-core machine -- was it statistically significant ?
You have here a parallel for loop, but the scheduling is unspecified.
#pragma omp parallel for private(i)
for(i = 2; i <= n; i++)
There are some scheduling types defined in OpenMP 3.0 standard. They can be changed by setting OMP_SCHEDULE environment variable to type[,chunk] where
type is one of static, dynamic, guided, or auto
chunk is an optional positive integer that specifies the chunk size
Another way of changing schedule kind is calling openmp function omp_set_schedule
The is_prime function can be rather fast. /I suggest/
prime[i-2] = is_prime(i);
So, the problem can came from wrong scheduling mode, when a little number is executed before barrier from scheduling.
And the printf have 2 parts inside it /I consider glibc as popular Linux libc implementation/
Parse the format string and put all parameters into buffer
Write buffer to file descriptor (to FILE buffer, as stdout is buffered by glibc by default)
The first part of printf can be done in parallel, but second part is a critical section and it is locked with _IO_flockfile.
What were your timings - was it much slower with the printf's? In some tight loops the printf's might take a large fraction of the total computing time; for example if is_prime() is pretty fast, and therefore the performance is determined more by the number of calls to printf than the number of (parallelized) calls to is_prime().
I have two CPUs on the chip and they have a shared memory. This is not a SMP architecture. Just two CPUs on the chip with shared memory.
There is a Unix-like operating system on the first CPU and there is a Linux operating system on the second CPU.
The first CPU does some job and the result of this job is some data. After first CPU finishes its job it should say to another CPU that job is finished and the second CPU have to process this data.
What is the way to handle interprocessor communication? What algorithm should I use to do that?
Any reference to an article about it would be greatly appreciated.
It all depends on the hardware. If all you have is shared memory, and no other way of communication, then you have to use a polling of some sort.
Are both of your processor running linux ? How do they handle the shared memory ?
A good solution is to use a linked list as a fifo. On this fifo you put data descriptor, like adress and size.
For example, you can have an input and output fifo, and go like this :
Processor A does some calculation
Processor A push the data descriptoron the output fifo
Processor A wait for data descriptor on the input fifo
loop
Processor B wait for data descriptor on the output fifo
Processor B works with data
Processor B push used data descriptor on the input fifo
loop
Of course, the hard part is in the locking. May be you should reformulate your question to emphasize this is not 'standard' SMP.
If you have no atomic test and set bit operation available on the memory, I guess you have to go with a scheme where some zone of memory is write only for one processor, and read only for the other.
Edit : See Hasturkun answer, for a way of passing messages from one processor to the other, using ordered write instead of atomicity to provide serialized access to some predefined data.
Ok. I understand the question.I have worked on this kind of an issue.
Now first thing that you need to understand is the working of the shared memory that exists between the 2 CPUs. Because these shared memory can be accessed in different ways, u need to figure out which one suits u the best.
Most times hardware semaphores will be provided in the shared memory along with the hardware interrupt to notify the message transfer from one processor to the other processor.
So have a look at this first.
A really good method is to just send IP packets back and forth (using sockets). this has the advantage that you can test stuff off-chip - as in, run a test version of one process on a PC, if you have networking.
If both processors are managed by a single os, then you can use any of the standard IPC to communicate with each other as OS takes care of everything. If they are running on different OSes then sockets would be your best bet.
EDIT
Quick unidirectional version:
Ready flag
Done flag
init:
init()
{
ready = 0;
done = 1;
}
writer:
send()
{
while (!done)
sleep();
/* copy data in */
done = 0;
ready = 1;
}
reader:
poll()
{
while (1)
{
if (ready)
{
recv();
}
sleep();
}
}
recv()
{
/* copy data out */
ready = 0;
done = 1;
}
Build a message passing system via the shared mem (which should be coherent, either by being uncached for both processors, or by use of cache flush/invalidate calls).
Your shared memory structure should have (at least) the following fields:
Current owner
Message active (as in, should be read)
Request usage fields
Flow will probably be like this: (assumed send/recv synchronized not to run at same time)
poll()
{
/* you're better off using interrupts instead, if you have them */
while(1)
{
if (current_owner == me)
{
if (active)
{
recv();
}
else if (!request[me] && request[other])
{
request[other] = 0;
current_owner = other;
}
}
sleep();
}
}
recv()
{
/* copy data... */
active = 0;
/* check if we still want it */
if (!request[me] && request[other])
{
request[other] = 0;
current_owner = other;
}
}
send()
{
request[me] = 1;
while (current_owner != me || active)
{
sleep();
}
request[me] = 0;
/* copy data in... */
/* pass to other side */
active = 1;
current_owner = other;
}
How about using the shared mem?
I don't have a good link right now, but if you google for IPC + shared mem I bet you find some good info :)
Are you sure you need to do this? In my experience you're better off letting your compiler & operating system manage how your process uses multiple CPUs.