$ time foo
real 0m0.003s
user 0m0.000s
sys 0m0.004s
$
What do real, user and sys mean in the output of time? Which one is meaningful when benchmarking my app?
Real, User and Sys process time statistics
One of these things is not like the other. Real refers to actual elapsed time; User and Sys refer to CPU time used only by the process.
Real is wall clock time - time from start to finish of the call. This is all elapsed time including time slices used by other processes and time the process spends blocked (for example if it is waiting for I/O to complete).
User is the amount of CPU time spent in user-mode code (outside the kernel) within the process. This is only actual CPU time used in executing the process. Other processes and time the process spends blocked do not count towards this figure.
Sys is the amount of CPU time spent in the kernel within the process. This means executing CPU time spent in system calls within the kernel, as opposed to library code, which is still running in user-space. Like 'user', this is only CPU time used by the process. See below for a brief description of kernel mode (also known as 'supervisor' mode) and the system call mechanism.
User+Sys will tell you how much actual CPU time your process used. Note that this is across all CPUs, so if the process has multiple threads (and this process is running on a computer with more than one processor) it could potentially exceed the wall clock time reported by Real (which usually occurs). Note that in the output these figures include the User and Sys time of all child processes (and their descendants) as well when they could have been collected, e.g. by wait(2) or waitpid(2), although the underlying system calls return the statistics for the process and its children separately.
Origins of the statistics reported by time (1)
The statistics reported by time are gathered from various system calls. 'User' and 'Sys' come from wait (2) (POSIX) or times (2) (POSIX), depending on the particular system. 'Real' is calculated from a start and end time gathered from the gettimeofday (2) call. Depending on the version of the system, various other statistics such as the number of context switches may also be gathered by time.
On a multi-processor machine, a multi-threaded process or a process forking children could have an elapsed time smaller than the total CPU time - as different threads or processes may run in parallel. Also, the time statistics reported come from different origins, so times recorded for very short running tasks may be subject to rounding errors, as the example given by the original poster shows.
A brief primer on Kernel vs. User mode
On Unix, or any protected-memory operating system, 'Kernel' or 'Supervisor' mode refers to a privileged mode that the CPU can operate in. Certain privileged actions that could affect security or stability can only be done when the CPU is operating in this mode; these actions are not available to application code. An example of such an action might be manipulation of the MMU to gain access to the address space of another process. Normally, user-mode code cannot do this (with good reason), although it can request shared memory from the kernel, which could be read or written by more than one process. In this case, the shared memory is explicitly requested from the kernel through a secure mechanism and both processes have to explicitly attach to it in order to use it.
The privileged mode is usually referred to as 'kernel' mode because the kernel is executed by the CPU running in this mode. In order to switch to kernel mode you have to issue a specific instruction (often called a trap) that switches the CPU to running in kernel mode and runs code from a specific location held in a jump table. For security reasons, you cannot switch to kernel mode and execute arbitrary code - the traps are managed through a table of addresses that cannot be written to unless the CPU is running in supervisor mode. You trap with an explicit trap number and the address is looked up in the jump table; the kernel has a finite number of controlled entry points.
The 'system' calls in the C library (particularly those described in Section 2 of the man pages) have a user-mode component, which is what you actually call from your C program. Behind the scenes, they may issue one or more system calls to the kernel to do specific services such as I/O, but they still also have code running in user-mode. It is also quite possible to directly issue a trap to kernel mode from any user space code if desired, although you may need to write a snippet of assembly language to set up the registers correctly for the call.
More about 'sys'
There are things that your code cannot do from user mode - things like allocating memory or accessing hardware (HDD, network, etc.). These are under the supervision of the kernel, and it alone can do them. Some operations like malloc orfread/fwrite will invoke these kernel functions and that then will count as 'sys' time. Unfortunately it's not as simple as "every call to malloc will be counted in 'sys' time". The call to malloc will do some processing of its own (still counted in 'user' time) and then somewhere along the way it may call the function in kernel (counted in 'sys' time). After returning from the kernel call, there will be some more time in 'user' and then malloc will return to your code. As for when the switch happens, and how much of it is spent in kernel mode... you cannot say. It depends on the implementation of the library. Also, other seemingly innocent functions might also use malloc and the like in the background, which will again have some time in 'sys' then.
To expand on the accepted answer, I just wanted to provide another reason why real ≠ user + sys.
Keep in mind that real represents actual elapsed time, while user and sys values represent CPU execution time. As a result, on a multicore system, the user and/or sys time (as well as their sum) can actually exceed the real time. For example, on a Java app I'm running for class I get this set of values:
real 1m47.363s
user 2m41.318s
sys 0m4.013s
• real: The actual time spent in running the process from start to finish, as if it was measured by a human with a stopwatch
• user: The cumulative time spent by all the CPUs during the computation
• sys: The cumulative time spent by all the CPUs during system-related tasks such as memory allocation.
Notice that sometimes user + sys might be greater than real, as
multiple processors may work in parallel.
Minimal runnable POSIX C examples
To make things more concrete, I want to exemplify a few extreme cases of time with some minimal C test programs.
All programs can be compiled and run with:
gcc -ggdb3 -o main.out -pthread -std=c99 -pedantic-errors -Wall -Wextra main.c
time ./main.out
and have been tested in Ubuntu 18.10, GCC 8.2.0, glibc 2.28, Linux kernel 4.18, ThinkPad P51 laptop, Intel Core i7-7820HQ CPU (4 cores / 8 threads), 2x Samsung M471A2K43BB1-CRC RAM (2x 16GiB).
sleep syscall
Non-busy sleep as done by the sleep syscall only counts in real, but not for user or sys.
For example, a program that sleeps for a second:
#define _XOPEN_SOURCE 700
#include <stdlib.h>
#include <unistd.h>
int main(void) {
sleep(1);
return EXIT_SUCCESS;
}
GitHub upstream.
outputs something like:
real 0m1.003s
user 0m0.001s
sys 0m0.003s
The same holds for programs blocked on IO becoming available.
For example, the following program waits for the user to enter a character and press enter:
#include <stdio.h>
#include <stdlib.h>
int main(void) {
printf("%c\n", getchar());
return EXIT_SUCCESS;
}
GitHub upstream.
And if you wait for about one second, it outputs just like the sleep example something like:
real 0m1.003s
user 0m0.001s
sys 0m0.003s
For this reason time can help you distinguish between CPU and IO bound programs: What do the terms "CPU bound" and "I/O bound" mean?
Multiple threads
The following example does niters iterations of useless purely CPU-bound work on nthreads threads:
#define _XOPEN_SOURCE 700
#include <assert.h>
#include <inttypes.h>
#include <pthread.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
uint64_t niters;
void* my_thread(void *arg) {
uint64_t *argument, i, result;
argument = (uint64_t *)arg;
result = *argument;
for (i = 0; i < niters; ++i) {
result = (result * result) - (3 * result) + 1;
}
*argument = result;
return NULL;
}
int main(int argc, char **argv) {
size_t nthreads;
pthread_t *threads;
uint64_t rc, i, *thread_args;
/* CLI args. */
if (argc > 1) {
niters = strtoll(argv[1], NULL, 0);
} else {
niters = 1000000000;
}
if (argc > 2) {
nthreads = strtoll(argv[2], NULL, 0);
} else {
nthreads = 1;
}
threads = malloc(nthreads * sizeof(*threads));
thread_args = malloc(nthreads * sizeof(*thread_args));
/* Create all threads */
for (i = 0; i < nthreads; ++i) {
thread_args[i] = i;
rc = pthread_create(
&threads[i],
NULL,
my_thread,
(void*)&thread_args[i]
);
assert(rc == 0);
}
/* Wait for all threads to complete */
for (i = 0; i < nthreads; ++i) {
rc = pthread_join(threads[i], NULL);
assert(rc == 0);
printf("%" PRIu64 " %" PRIu64 "\n", i, thread_args[i]);
}
free(threads);
free(thread_args);
return EXIT_SUCCESS;
}
GitHub upstream + plot code.
Then we plot wall, user and sys as a function of the number of threads for a fixed 10^10 iterations on my 8 hyperthread CPU:
Plot data.
From the graph, we see that:
for a CPU intensive single core application, wall and user are about the same
for 2 cores, user is about 2x wall, which means that the user time is counted across all threads.
user basically doubled, and while wall stayed the same.
this continues up to 8 threads, which matches my number of hyperthreads in my computer.
After 8, wall starts to increase as well, because we don't have any extra CPUs to put more work in a given amount of time!
The ratio plateaus at this point.
Note that this graph is only so clear and simple because the work is purely CPU-bound: if it were memory bound, then we would get a fall in performance much earlier with less cores because the memory accesses would be a bottleneck as shown at What do the terms "CPU bound" and "I/O bound" mean?
Quickly checking that wall < user is a simple way to determine that a program is multithreaded, and the closer that ratio is to the number of cores, the more effective the parallelization is, e.g.:
multithreaded linkers: Can gcc use multiple cores when linking?
C++ parallel sort: Are C++17 Parallel Algorithms implemented already?
Sys heavy work with sendfile
The heaviest sys workload I could come up with was to use the sendfile, which does a file copy operation on kernel space: Copy a file in a sane, safe and efficient way
So I imagined that this in-kernel memcpy will be a CPU intensive operation.
First I initialize a large 10GiB random file with:
dd if=/dev/urandom of=sendfile.in.tmp bs=1K count=10M
Then run the code:
#define _GNU_SOURCE
#include <assert.h>
#include <fcntl.h>
#include <stdlib.h>
#include <sys/sendfile.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>
int main(int argc, char **argv) {
char *source_path, *dest_path;
int source, dest;
struct stat stat_source;
if (argc > 1) {
source_path = argv[1];
} else {
source_path = "sendfile.in.tmp";
}
if (argc > 2) {
dest_path = argv[2];
} else {
dest_path = "sendfile.out.tmp";
}
source = open(source_path, O_RDONLY);
assert(source != -1);
dest = open(dest_path, O_WRONLY | O_CREAT | O_TRUNC, S_IRUSR | S_IWUSR);
assert(dest != -1);
assert(fstat(source, &stat_source) != -1);
assert(sendfile(dest, source, 0, stat_source.st_size) != -1);
assert(close(source) != -1);
assert(close(dest) != -1);
return EXIT_SUCCESS;
}
GitHub upstream.
which gives basically mostly system time as expected:
real 0m2.175s
user 0m0.001s
sys 0m1.476s
I was also curious to see if time would distinguish between syscalls of different processes, so I tried:
time ./sendfile.out sendfile.in1.tmp sendfile.out1.tmp &
time ./sendfile.out sendfile.in2.tmp sendfile.out2.tmp &
And the result was:
real 0m3.651s
user 0m0.000s
sys 0m1.516s
real 0m4.948s
user 0m0.000s
sys 0m1.562s
The sys time is about the same for both as for a single process, but the wall time is larger because the processes are competing for disk read access likely.
So it seems that it does in fact account for which process started a given kernel work.
Bash source code
When you do just time <cmd> on Ubuntu, it use the Bash keyword as can be seen from:
type time
which outputs:
time is a shell keyword
So we grep source in the Bash 4.19 source code for the output string:
git grep '"user\b'
which leads us to execute_cmd.c function time_command, which uses:
gettimeofday() and getrusage() if both are available
times() otherwise
all of which are Linux system calls and POSIX functions.
GNU Coreutils source code
If we call it as:
/usr/bin/time
then it uses the GNU Coreutils implementation.
This one is a bit more complex, but the relevant source seems to be at resuse.c and it does:
a non-POSIX BSD wait3 call if that is available
times and gettimeofday otherwise
1: https://i.stack.imgur.com/qAfEe.png**Minimal runnable POSIX C examples**
To make things more concrete, I want to exemplify a few extreme cases of time with some minimal C test programs.
All programs can be compiled and run with:
gcc -ggdb3 -o main.out -pthread -std=c99 -pedantic-errors -Wall -Wextra main.c
time ./main.out
and have been tested in Ubuntu 18.10, GCC 8.2.0, glibc 2.28, Linux kernel 4.18, ThinkPad P51 laptop, Intel Core i7-7820HQ CPU (4 cores / 8 threads), 2x Samsung M471A2K43BB1-CRC RAM (2x 16GiB).
sleep
Non-busy sleep does not count in either user or sys, only real.
For example, a program that sleeps for a second:
#define _XOPEN_SOURCE 700
#include <stdlib.h>
#include <unistd.h>
int main(void) {
sleep(1);
return EXIT_SUCCESS;
}
GitHub upstream.
outputs something like:
real 0m1.003s
user 0m0.001s
sys 0m0.003s
The same holds for programs blocked on IO becoming available.
For example, the following program waits for the user to enter a character and press enter:
#include <stdio.h>
#include <stdlib.h>
int main(void) {
printf("%c\n", getchar());
return EXIT_SUCCESS;
}
GitHub upstream.
And if you wait for about one second, it outputs just like the sleep example something like:
real 0m1.003s
user 0m0.001s
sys 0m0.003s
For this reason time can help you distinguish between CPU and IO bound programs: What do the terms "CPU bound" and "I/O bound" mean?
multithreaded linkers: Can gcc use multiple cores when linking?
C++ parallel sort: Are C++17 Parallel Algorithms implemented already?
Real shows total turn-around time for a process;
while User shows the execution time for user-defined instructions
and Sys is for time for executing system calls!
Real time includes the waiting time also (the waiting time for I/O etc.)
In very simple terms, I like to think about it like this:
real is the actual amount of time it took to run the command (as if you had timed it with a stopwatch)
user and sys are how much 'work' the CPU had to do to execute the command. This 'work' is expressed in units of time.
Generally speaking:
user is how much work the CPU did to run to run the command's code
sys is how much work the CPU had to do to handle 'system overhead' type tasks (such as allocating memory, file I/O, ect.) in order to support the running command
Since these last two times are counting 'work' done, they don't include time a thread might have spent waiting (such as waiting on another process or for disk I/O to finish).
real, however, is a measure of actual runtime and not 'work', so it does include any time spent waiting (which is why sometimes real > usr+sys).
And finally, sometimes the reverse is true (usr+sys > real) for multi-threaded applications. This also arises because we are comparing 'work-time' to actual time. For example, if 3 processors each run continuously for 10 minutes to execute a command, you will get real = 10m but usr = 30m.
I want to mention some other scenario when the real-time is much much bigger than user + sys. I've created a simple server which respondes after a long time
real 4.784
user 0.01s
sys 0.01s
the issue is that in this scenario the process waits for the response which is not on the user site nor in the system.
Something similar happens when you run the find command. In that case, the time is spent mostly on requesting and getting a response from SSD.
Must mention that at least on my AMD Ryzen CPU, the user is always large than real in multi-threaded program(or single threaded program compiled with -O3).
eg.
real 0m5.815s
user 0m8.213s
sys 0m0.473s
I have a scientific camera, a USB3Vision device, that I wrote a program to captures images for. No problems when I use it on my desktop, it works fine every time. If I take the exact same application to another (2 other actually) computer it acquires one image but hangs the second time that you try to capture another.
The desktop I'm developing on is Fedora 25 and I am having this problem on clean installs of both Fedora 24 and 25. Also, on both computers the permissions of the USB device have been set to 666 using a udev rule so it's not some permissions error.
If I attach gdb to the process that's hung it shows that it's waiting on a blocking call to poll(), I've tried waiting a really long time but it never completes. This is where I'm at a bit of a loss now, I'm not really sure what to poke at to troubleshoot why that call never finishes on computers that aren't the mine. I'm also not really sure how to determine what's different between the USB buses on the different computers, that would probably be really useful too.
Example code to recreate this issue is irrelevant IMO, but here it is because it may be asked for if not included:
#include <arv.h>
#include <stdlib.h>
#include <stdio.h>
int
main (int argc, char **argv)
{
ArvCamera *camera;
ArvBuffer *buffer;
camera = arv_camera_new (argc > 1 ? argv[1] : NULL);
buffer = arv_camera_acquisition (camera, 0);
if (ARV_IS_BUFFER (buffer))
printf ("Image successfully acquired\n");
else
printf ("Failed to acquire a single image\n");
g_clear_object (&camera);
g_clear_object (&buffer);
return EXIT_SUCCESS;
}
This is just one of the test programs that comes with the Aravis library that works with GenICam type cameras.
In some purpose it is needed to make JVM think about it runs on machine with Ncores on board instead of real number of cores (e.g. 4 cores instead of 16).
JVM runs under some Linux build, based on Mandriva/Red Hat Linux core.
This question is borderline case because I expect various solutions of this problem. This is not pure linux-administration question, and it isn't pure programmer's question.
So... any ideas?
In order to make Runtime.getRuntime().availableProcessors() return whatever you want, you can override JVM_ActiveProcessorCount function using LD_PRELOAD trick. Here is a tiny program to do this:
#include <stdlib.h>
#include <unistd.h>
int JVM_ActiveProcessorCount(void) {
char* val = getenv("_NUM_CPUS");
return val != NULL ? atoi(val) : sysconf(_SC_NPROCESSORS_ONLN);
}
First, make a shared library of this:
gcc -O3 -fPIC -shared -Wl,-soname,libnumcpus.so -o libnumcpus.so numcpus.c
Then run Java as follows:
$ LD_PRELOAD=/path/to/libnumcpus.so _NUM_CPUS=2 java AvailableProcessors
The following Java program prints the number of processors as seen by the Java VM:
public class AvailableProcessors {
public static void main(String... args) {
System.out.println(Runtime.getRuntime().availableProcessors());
}
}
If I execute this program on my home computer, it prints 4, which is the actual number of cores (including hyper threading). Now let's trick the Java VM into believing there are only two processors:
$ echo '0-1' > /tmp/online
$ mount --bind /tmp/online /sys/devices/system/cpu/online
If I run the above program again, it prints 2 instead of 4.
This trick affects all processes on your system. However, it's possible to restrict the effect only to certain processes. Each process on Linux can have its own namespace of mount points. See for example the section Pre-process namespaces in the man page of mount(2). You can for example use lxc to start new processes with their own mount namespace.
I have developed MPI program which can execute matrix multiplication on different cores in distributed environment and I can demonstrate the execution on different nodes by getting hostname of the node. But when we are running program on single node can I get core id which demonstrate the execution on multiple node sample code is given below
#include"stdio.h"
#include"stdlib.h"
#include"mpi.h"
int main(int argc , char **argv)
{
int size,rank;
int a,b,c;
MPI_Init(&argc, &argv);
MPI_Comm_size(MPI_COMM_WORLD, &size);
MPI_Comm_rank(MPI_COMM_WORLD, &rank);
if(rank==0)
{
for(i=0;i<size;i++)
{
printf("insert a and b");
scanf("%d",&b);
scanf("%d",&c);
MPI_Send(&b,1,MPI_INT,i+1,6,MPI_COMM_WORLD);
MPI_Send(&c,1,MPI_INT,i+1,6,MPI_COMM_WORLD);
}
}
if(rank!=0)
{
MPI_Recv(&b,1,MPI_INT,0,6,MPI_COMM_WORLD,&s);
MPI_Recv(&c,1,MPI_INT,0,6,MPI_COMM_WORLD,&s);
a=b*c;
printf("Mul = %d\n",a);
//Print name of core on which my process is running
}
MPI_Finalize();
return 0;
}
While it is possible to obtain the ID of the logical processor that currently executes the code, this often makes no sense unless you enable MPI process binding, also known as process pinning (in Intel's parlance). Binding (or pinning) restricts the CPU affinity set of each MPI process, i.e. the set of CPUs on which the process is allowed to execute. If the affinity set includes only a single logical CPU, then the process would only execute on that logical CPU. A logical CPU usually corresponds to a hardware thread on CPUs with SMT/hyperthreading or to a CPU core on non-SMT/non-hyperthreaded CPUs. Given affinity sets that include more than one logical CPUs, the scheduler is allowed to migrate the process around in order to keep the CPUs in the set equally busy. The default affinity set usually includes all available logical CPUs, that is the process could be scheduled for execution on any core or hardware thread.
Only when MPI process binding is in place and each process is bound to a single logical CPU it makes sense to actually query the OS for the location of the process. You have to consult your MPI implementation manual on how to enable it. For example, with Open MPI you would do something like:
mpiexec --bind-to-core --bycore -n 120 ...
--bind-to-core tells Open MPI to bind each process to a single CPU core and --bycore tells it to allocate cores consecutively on multisocket machines (that is, first to allocate all cores in the first socket, then in the second socket, etc.) With Intel MPI the binding (called pinning by Intel) is enabled by setting the environment variable I_MPI_PIN to 1. The process placement strategy is controlled by the value of I_MPI_PIN_DOMAIN. To achieve the same as the above shown Open MPI command line, one would to the following with Intel MPI:
mpiexec -n 120 -env I_MPI_PIN 1 -env I_MPI_PIN_DOMAIN "core:compact" ...
To obtain the location of your processes in a platform-independent way, you could use hwloc_get_last_cpu_location() from the hwloc library. It is developed as part of the Open MPI project but can be used as a stand-alone library. It provides an abstract interface to query the system topology and to manipulate the affinity of processes and threads. hwloc supports Linux, Windows and many other OSes.
We can get core id by using the sched_getcpu() function present in utmpx.h header file
sample program from net is shown below
#include <stdio.h>
#include<mpi.h>
#include <utmpx.h>
int main( int argc , char **argv )
{
MPI_Init(&argc, &argv);
printf( "cpu = %d\n", sched_getcpu() );
MPI_Finalize();
return 0;
}
You could use the Linux specific getcpu(2) syscall (or, as Krishna answered, the Linux specific sched_getcpu(3) function wrapping it). Read carefully the man page (that getcpu syscall does not have any libc wrapper!). Notice that it may gives you some obsolete information (because the kernel can -and does- migrate processes or tasks -e.g. threads- at any time from one CPU core to another).
Otherwise, your MPI implementation is probably using threads or processes. You could query them with gettid(2) (which needs some wrapper) for threads, or getpid(2) for processes.
You may want to code:
#include <unistd.h>
#include <asm/unistd.h>
#include <sys/syscall.h>
static inline long my_gettid(void)
{ return syscall(SYS_gettid); }
and perhaps something similar for getcpu....
You could also use proc(5) e.g. query /proc/self/stat (the 39th field gives the processor number)... Perhaps just displaying all of it is the simplest way:
{
char line[128];
FILE *fs = fopen("/proc/self/stat","r");
if (!fs) { perror("/proc/self/stat"); exit(EXIT_FAILURE); };
while (!feof(fs)) {
memset(line, 0, sizeof(line));
fgets(line, sizeof(line), fs);
fputs(line, stdout);
};
fclose(fs);
}
Remember that the Linux kernel (its scheduler) is migrating tasks (i.e. processes or threads) from one CPU core to another one at any time. So querying that is next to useless (the task could have migrated between the moment you query it and the moment you display it).
On Linux/NPTL, threads are created as some kind of process.
I can see some of my process have a weird cmdline:
cat /proc/5590/cmdline
hald-addon-storage: polling /dev/scd0 (every 2 sec)
Do you have an idea how I could do that for each thread of my process? That would be very helpful for debugging.
If you want to do this in a portable way, something that will work across multiple Unix variations, there are very few options available.
What you have to do is that your caller process must call exec with the argv [0] argument pointing to the name that you would like to see in the process output, and the filename pointing to the actual executable.
You can try this behavior from the shell by using:
exec -a "This is my cute name" bash
That will replace the current bash process with one named "This is my cute name".
For doing this in C, you can look at the source code of sendmail or any other piece of software that has been ported extensively and find all the variations that are needed across operating systems to support this.
Some operating systems have a setproctitle(3) API, some others allow you to override the contents of argv [0] and show that result.
argv points to writable strings. Just write stuff to them:
#include <string.h>
#include <unistd.h>
int
main(int argc, char** argv)
{
strcpy(argv[0], "Hello, world!");
sleep(10);
return 0;
}
Bah.. the code is not that nice, the trick is to reuse the environ (here argv_buffer) pointer:
memset (argv_buffer[0] + len, 0, argv_size - len);
argv_buffer[1] = NULL;
Any better idea?
Is that working for different threads?