starpu-max/doc/doxygen/chapters/390_faq.doxy

/*
 * This file is part of the StarPU Handbook.
 * Copyright (C) 2009--2011  Universit@'e de Bordeaux
 * Copyright (C) 2010, 2011, 2012, 2013, 2014, 2016  CNRS
 * Copyright (C) 2011, 2012 INRIA
 * See the file version.doxy for copying conditions.
 */

/*! \page FrequentlyAskedQuestions Frequently Asked Questions

\section HowToInitializeAComputationLibraryOnceForEachWorker How To Initialize A Computation Library Once For Each Worker?

Some libraries need to be initialized once for each concurrent instance that
may run on the machine. For instance, a C++ computation class which is not
thread-safe by itself, but for which several instanciated objects of that class
can be used concurrently. This can be used in StarPU by initializing one such
object per worker. For instance, the <c>libstarpufft</c> example does the following to
be able to use FFTW on CPUs.

Some global array stores the instanciated objects:

\code{.c}
fftw_plan plan_cpu[STARPU_NMAXWORKERS];
\endcode

At initialisation time of libstarpu, the objects are initialized:

\code{.c}
int workerid;
for (workerid = 0; workerid < starpu_worker_get_count(); workerid++) {
    switch (starpu_worker_get_type(workerid)) {
        case STARPU_CPU_WORKER:
            plan_cpu[workerid] = fftw_plan(...);
            break;
    }
}
\endcode

And in the codelet body, they are used:

\code{.c}
static void fft(void *descr[], void *_args)
{
    int workerid = starpu_worker_get_id();
    fftw_plan plan = plan_cpu[workerid];
    ...

    fftw_execute(plan, ...);
}
\endcode

This however is not sufficient for FFT on CUDA: initialization has
to be done from the workers themselves.  This can be done thanks to
starpu_execute_on_each_worker().  For instance <c>libstarpufft</c> does the following.

\code{.c}
static void fft_plan_gpu(void *args)
{
    plan plan = args;
    int n2 = plan->n2[0];
    int workerid = starpu_worker_get_id();

    cufftPlan1d(&plan->plans[workerid].plan_cuda, n, _CUFFT_C2C, 1);
    cufftSetStream(plan->plans[workerid].plan_cuda, starpu_cuda_get_local_stream());
}
void starpufft_plan(void)
{
    starpu_execute_on_each_worker(fft_plan_gpu, plan, STARPU_CUDA);
}
\endcode

\section UsingTheDriverAPI Using The Driver API

\ref API_Running_Drivers

\code{.c}
int ret;
struct starpu_driver = {
    .type = STARPU_CUDA_WORKER,
    .id.cuda_id = 0
};
ret = starpu_driver_init(&d);
if (ret != 0)
    error();
while (some_condition) {
    ret = starpu_driver_run_once(&d);
    if (ret != 0)
        error();
}
ret = starpu_driver_deinit(&d);
if (ret != 0)
    error();
\endcode

To add a new kind of device to the structure starpu_driver, one needs to:
<ol>
<li> Add a member to the union starpu_driver::id
</li>
<li> Modify the internal function <c>_starpu_launch_drivers()</c> to
make sure the driver is not always launched.
</li>
<li> Modify the function starpu_driver_run() so that it can handle
another kind of architecture.
</li>
<li> Write the new function <c>_starpu_run_foobar()</c> in the
corresponding driver.
</li>
</ol>

\section On-GPURendering On-GPU Rendering

Graphical-oriented applications need to draw the result of their computations,
typically on the very GPU where these happened. Technologies such as OpenGL/CUDA
interoperability permit to let CUDA directly work on the OpenGL buffers, making
them thus immediately ready for drawing, by mapping OpenGL buffer, textures or
renderbuffer objects into CUDA.  CUDA however imposes some technical
constraints: peer memcpy has to be disabled, and the thread that runs OpenGL has
to be the one that runs CUDA computations for that GPU.

To achieve this with StarPU, pass the option
\ref disable-cuda-memcpy-peer "--disable-cuda-memcpy-peer"
to <c>./configure</c> (TODO: make it dynamic), OpenGL/GLUT has to be initialized
first, and the interoperability mode has to
be enabled by using the field
starpu_conf::cuda_opengl_interoperability, and the driver loop has to
be run by the application, by using the field
starpu_conf::not_launched_drivers to prevent StarPU from running it in
a separate thread, and by using starpu_driver_run() to run the loop.
The examples <c>gl_interop</c> and <c>gl_interop_idle</c> show how it
articulates in a simple case, where rendering is done in task
callbacks. The former uses <c>glutMainLoopEvent</c> to make GLUT
progress from the StarPU driver loop, while the latter uses
<c>glutIdleFunc</c> to make StarPU progress from the GLUT main loop.

Then, to use an OpenGL buffer as a CUDA data, StarPU simply needs to be given
the CUDA pointer at registration, for instance:

\code{.c}
/* Get the CUDA worker id */
for (workerid = 0; workerid < starpu_worker_get_count(); workerid++)
        if (starpu_worker_get_type(workerid) == STARPU_CUDA_WORKER)
                break;

/* Build a CUDA pointer pointing at the OpenGL buffer */
cudaGraphicsResourceGetMappedPointer((void**)&output, &num_bytes, resource);

/* And register it to StarPU */
starpu_vector_data_register(&handle, starpu_worker_get_memory_node(workerid),
                            output, num_bytes / sizeof(float4), sizeof(float4));

/* The handle can now be used as usual */
starpu_task_insert(&cl, STARPU_RW, handle, 0);

/* ... */

/* This gets back data into the OpenGL buffer */
starpu_data_unregister(handle);
\endcode

and display it e.g. in the callback function.

\section UsingStarPUWithMKL Using StarPU With MKL 11 (Intel Composer XE 2013)

Some users had issues with MKL 11 and StarPU (versions 1.1rc1 and
1.0.5) on Linux with MKL, using 1 thread for MKL and doing all the
parallelism using StarPU (no multithreaded tasks), setting the
environment variable <c>MKL_NUM_THREADS</c> to <c>1</c>, and using the threaded MKL library,
with <c>iomp5</c>.

Using this configuration, StarPU only uses 1 core, no matter the value of
\ref STARPU_NCPU. The problem is actually a thread pinning issue with MKL.

The solution is to set the environment variable KMP_AFFINITY to <c>disabled</c>
(http://software.intel.com/sites/products/documentation/studio/composer/en-us/2011Update/compiler_c/optaps/common/optaps_openmp_thread_affinity.htm).

\section ThreadBindingOnNetBSD Thread Binding on NetBSD

When using StarPU on a NetBSD machine, if the topology
discovery library <c>hwloc</c> is used, thread binding will fail. To
prevent the problem, you should at least use the version 1.7 of
<c>hwloc</c>, and also issue the following call:

\verbatim
$ sysctl -w security.models.extensions.user_set_cpu_affinity=1
\endverbatim

Or add the following line in the file <c>/etc/sysctl.conf</c>

\verbatim
security.models.extensions.user_set_cpu_affinity=1
\endverbatim


\section PauseResume Interleaving StarPU and non-StarPU code

If your application only partially uses StarPU, and you do not want to
call starpu_init() / starpu_shutdown() at the beginning/end
of each section, StarPU workers will poll for work between the
sections. To avoid this behavior, you can "pause" StarPU with the 
starpu_pause() function. This will prevent the StarPU workers from
accepting new work (tasks that are already in progress will not be
frozen), and stop them from polling for more work.

Note that this does not prevent you from submitting new tasks, but
they won't execute until starpu_resume() is called. Also note
that StarPU must not be paused when you call starpu_shutdown(), and
that this function pair works in a push/pull manner, i.e you need to
match the number of calls to these functions to clear their effect.


One way to use these functions could be:
\code{.c}
starpu_init(NULL);
starpu_pause(); // To submit all the tasks without a single one executing
submit_some_tasks();
starpu_resume(); // The tasks start executing


starpu_task_wait_for_all();
starpu_pause(); // Stop the workers from polling

// Non-StarPU code

starpu_resume();
// ...
starpu_shutdown();
\endcode

\section GPUEatingCores When running with CUDA or OpenCL devices, I am seeing less CPU cores

Yes, this is on purpose.

Since GPU devices are way faster than CPUs, StarPU needs to react quickly when
a task is finished, to feed the GPU with another task (StarPU actually submits
a couple of tasks in advance so as to pipeline this, but filling the pipeline
still has to be happening often enough), and thus it has to dedicate threads for
this, and this is a very CPU-consuming duty. StarPU thus dedicates one CPU core
for driving each GPU by default.

Such dedication is also useful when a codelet is hybrid, i.e. while kernels are
running on the GPU, the codelet can run some computation, which thus be run by
the CPU core instead of driving the GPU.

One can choose to dedicate only one thread for all the CUDA devices by setting
the STARPU_CUDA_THREAD_PER_DEV environment variable to 1. The application
however should use STARPU_CUDA_ASYNC on its CUDA codelets (asynchronous
execution), otherwise the execution of a synchronous CUDA codelet will
monopolize the thread, and other CUDA devices will thus starve while it is
executing.

\section CUDADrivers StarPU does not see my CUDA device

First make sure that CUDA is properly running outside StarPU: build and
run the following program with -lcudart:

\code{.c}
#include <stdio.h>
#include <cuda.h>
#include <cuda_runtime.h>

int main(void) {
	int n, i, version;
	cudaError_t err;

	err = cudaGetDeviceCount(&n);
	if (err)
	{
		fprintf(stderr,"cuda error %d\n", err);
		exit(1);
	}
	cudaDriverGetVersion(&version);
	printf("driver version %d\n", version);
	cudaRuntimeGetVersion(&version);
	printf("runtime version %d\n", version);
	printf("\n");

	for (i = 0; i < n; i++) {
		struct cudaDeviceProp props;
		printf("CUDA%d\n", i);
		err = cudaGetDeviceProperties(&props, i);
		if (err)
		{
			fprintf(stderr,"cuda error %d\n", err);
			continue;
		}
		printf("%s\n", props.name);
		printf("%0.3f GB\n", (float) props.totalGlobalMem / (1<<30));
		printf("%u MP\n", props.multiProcessorCount);
		printf("\n");
	}
	return 0;
}
\endcode

If that program does not find your device, the problem is not at the StarPU
level, but the CUDA drivers, check the documentation of your CUDA
setup.

\section OpenCLDrivers StarPU does not see my OpenCL device

First make sure that OpenCL is properly running outside StarPU: build and
run the following program with -lOpenCL:

\code{.c}
#include <CL/cl.h>
#include <stdio.h>
#include <assert.h>

int main(void) {
    cl_device_id did[16];
    cl_int err;
    cl_platform_id pid, pids[16];
    cl_uint nbplat, nb;
    char buf[128];
    size_t size;
    int i, j;

    err = clGetPlatformIDs(sizeof(pids)/sizeof(pids[0]), pids, &nbplat);
    assert(err == CL_SUCCESS);
    printf("%u platforms\n", nbplat);
    for (j = 0; j < nbplat; j++) {
        pid = pids[j];
        printf("    platform %d\n", j);
        err = clGetPlatformInfo(pid, CL_PLATFORM_VERSION, sizeof(buf)-1, buf, &size);
        assert(err == CL_SUCCESS);
        buf[size] = 0;
        printf("        platform version %s\n", buf);

        err = clGetDeviceIDs(pid, CL_DEVICE_TYPE_ALL, sizeof(did)/sizeof(did[0]), did, &nb);
        assert(err == CL_SUCCESS);
        printf("%d devices\n", nb);
        for (i = 0; i < nb; i++) {
            err = clGetDeviceInfo(did[i], CL_DEVICE_VERSION, sizeof(buf)-1, buf, &size);
            buf[size] = 0;
            printf("    device %d version %s\n", i, buf);
        }
    }

    return 0;
}
\endcode

If that program does not find your device, the problem is not at the StarPU
level, but the OpenCL drivers, check the documentation of your OpenCL
implementation.

*/