blis/frame/thread/bli_thrcomm_pthreads.c

/*

   BLIS
   An object-based framework for developing high-performance BLAS-like
   libraries.

   Copyright (C) 2014, The University of Texas at Austin
   Copyright (C) 2018 - 2019, Advanced Micro Devices, Inc.

   Redistribution and use in source and binary forms, with or without
   modification, are permitted provided that the following conditions are
   met:
    - Redistributions of source code must retain the above copyright
      notice, this list of conditions and the following disclaimer.
    - Redistributions in binary form must reproduce the above copyright
      notice, this list of conditions and the following disclaimer in the
      documentation and/or other materials provided with the distribution.
    - Neither the name(s) of the copyright holder(s) nor the names of its
      contributors may be used to endorse or promote products derived
      from this software without specific prior written permission.

   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
   HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

*/

#include "blis.h"

#ifdef BLIS_ENABLE_PTHREADS

thrcomm_t* bli_thrcomm_create( rntm_t* rntm, dim_t n_threads )
{
	#ifdef BLIS_ENABLE_MEM_TRACING
	printf( "bli_thrcomm_create(): " );
	#endif

	thrcomm_t* comm = bli_sba_acquire( rntm, sizeof(thrcomm_t) );

	bli_thrcomm_init( n_threads, comm );

	return comm;
}

void bli_thrcomm_free( rntm_t* rntm, thrcomm_t* comm )
{
	if ( comm == NULL ) return;

	bli_thrcomm_cleanup( comm );

	#ifdef BLIS_ENABLE_MEM_TRACING
	printf( "bli_thrcomm_free(): " );
	#endif

	bli_sba_release( rntm, comm );
}

#ifdef BLIS_USE_PTHREAD_BARRIER

void bli_thrcomm_init( dim_t n_threads, thrcomm_t* comm )
{
	if ( comm == NULL ) return;
	comm->sent_object = NULL;
	comm->n_threads = n_threads;
	bli_pthread_barrier_init( &comm->barrier, NULL, n_threads );
}

void bli_thrcomm_cleanup( thrcomm_t* comm )
{
	if ( comm == NULL ) return;
	bli_pthread_barrier_destroy( &comm->barrier );
}

void bli_thrcomm_barrier( dim_t t_id, thrcomm_t* comm )
{
	bli_pthread_barrier_wait( &comm->barrier );
}

#else

void bli_thrcomm_init( dim_t n_threads, thrcomm_t* comm )
{
	if ( comm == NULL ) return;
	comm->sent_object = NULL;
	comm->n_threads = n_threads;
	comm->barrier_sense = 0;
	comm->barrier_threads_arrived = 0;

//#ifdef BLIS_USE_PTHREAD_MUTEX
//	bli_pthread_mutex_init( &comm->mutex, NULL );
//#endif
}

void bli_thrcomm_cleanup( thrcomm_t* comm )
{
//#ifdef BLIS_USE_PTHREAD_MUTEX
//	if ( comm == NULL ) return;
//	bli_pthread_mutex_destroy( &comm->mutex );
//#endif
}

void bli_thrcomm_barrier( dim_t t_id, thrcomm_t* comm )
{
#if 0
	if ( comm == NULL || comm->n_threads == 1 ) return;
	bool_t my_sense = comm->sense;
	dim_t my_threads_arrived;

#ifdef BLIS_USE_PTHREAD_MUTEX
	bli_pthread_mutex_lock( &comm->mutex );
	my_threads_arrived = ++(comm->threads_arrived);
	bli_pthread_mutex_unlock( &comm->mutex );
#else
	my_threads_arrived = __sync_add_and_fetch(&(comm->threads_arrived), 1);
#endif

	if ( my_threads_arrived == comm->n_threads )
	{
		comm->threads_arrived = 0;
		comm->sense = !comm->sense;
	}
	else
	{
		volatile bool_t* listener = &comm->sense;
		while( *listener == my_sense ) {}
	}
#endif
	bli_thrcomm_barrier_atomic( t_id, comm );
}

#endif


// A data structure to assist in passing operands to additional threads.
typedef struct thread_data
{
	l3int_t    func;
	opid_t     family;
	pack_t     schema_a;
	pack_t     schema_b;
	obj_t*     alpha;
	obj_t*     a;
	obj_t*     b;
	obj_t*     beta;
	obj_t*     c;
	cntx_t*    cntx;
	rntm_t*    rntm;
	cntl_t*    cntl;
	dim_t      tid;
	thrcomm_t* gl_comm;
	array_t*   array;
} thread_data_t;

// Entry point for additional threads
void* bli_l3_thread_entry( void* data_void )
{
	thread_data_t* data     = data_void;

	l3int_t        func     = data->func;
	opid_t         family   = data->family;
	pack_t         schema_a = data->schema_a;
	pack_t         schema_b = data->schema_b;
	obj_t*         alpha    = data->alpha;
	obj_t*         a        = data->a;
	obj_t*         b        = data->b;
	obj_t*         beta     = data->beta;
	obj_t*         c        = data->c;
	cntx_t*        cntx     = data->cntx;
	rntm_t*        rntm     = data->rntm;
	cntl_t*        cntl     = data->cntl;
	dim_t          tid      = data->tid;
	array_t*       array    = data->array;
	thrcomm_t*     gl_comm  = data->gl_comm;

	// Create a thread-local copy of the master thread's rntm_t. This is
	// necessary since we want each thread to be able to track its own
	// small block pool_t as it executes down the function stack.
	rntm_t           rntm_l = *rntm;
	rntm_t* restrict rntm_p = &rntm_l;

	// Use the thread id to access the appropriate pool_t* within the
	// array_t, and use it to set the sba_pool field within the rntm_t.
	// If the pool_t* element within the array_t is NULL, it will first
	// be allocated/initialized.
	bli_sba_rntm_set_pool( tid, array, rntm_p );

	obj_t          a_t, b_t, c_t;
	cntl_t*        cntl_use;
	thrinfo_t*     thread;

	// Alias thread-local copies of A, B, and C. These will be the objects
	// we pass down the algorithmic function stack. Making thread-local
	// alaises is highly recommended in case a thread needs to change any
	// of the properties of an object without affecting other threads'
	// objects.
	bli_obj_alias_to( a, &a_t );
	bli_obj_alias_to( b, &b_t );
	bli_obj_alias_to( c, &c_t );

	// Create a default control tree for the operation, if needed.
	bli_l3_cntl_create_if( family, schema_a, schema_b,
	                       &a_t, &b_t, &c_t, rntm_p, cntl, &cntl_use );

	// Create the root node of the current thread's thrinfo_t structure.
	bli_l3_thrinfo_create_root( tid, gl_comm, rntm_p, cntl_use, &thread );

	func
	(
	  alpha,
	  &a_t,
	  &b_t,
	  beta,
	  &c_t,
	  cntx,
	  rntm_p,
	  cntl_use,
	  thread
	);

	// Free the thread's local control tree.
	bli_l3_cntl_free( rntm_p, cntl_use, thread );

	// Free the current thread's thrinfo_t structure.
	bli_l3_thrinfo_free( rntm_p, thread );

	return NULL;
}

void bli_l3_thread_decorator
     (
       l3int_t     func,
       opid_t      family,
       obj_t*      alpha,
       obj_t*      a,
       obj_t*      b,
       obj_t*      beta,
       obj_t*      c,
       cntx_t*     cntx,
       rntm_t*     rntm,
       cntl_t*     cntl
     )
{
	// This is part of a hack to support mixed domain in bli_gemm_front().
	// Sometimes we need to specify a non-standard schema for A and B, and
	// we decided to transmit them via the schema field in the obj_t's
	// rather than pass them in as function parameters. Once the values
	// have been read, we immediately reset them back to their expected
	// values for unpacked objects.
	pack_t schema_a = bli_obj_pack_schema( a );
	pack_t schema_b = bli_obj_pack_schema( b );
	bli_obj_set_pack_schema( BLIS_NOT_PACKED, a );
	bli_obj_set_pack_schema( BLIS_NOT_PACKED, b );

	// Query the total number of threads from the context.
	const dim_t n_threads = bli_rntm_num_threads( rntm );

	// NOTE: The sba was initialized in bli_init().

	// Check out an array_t from the small block allocator. This is done
	// with an internal lock to ensure only one application thread accesses
	// the sba at a time. bli_sba_checkout_array() will also automatically
	// resize the array_t, if necessary.
	array_t* restrict array = bli_sba_checkout_array( n_threads );

	// Access the pool_t* for thread 0 and embed it into the rntm. We do
	// this up-front only so that we have the rntm_t.sba_pool field
	// initialized and ready for the global communicator creation below.
	bli_sba_rntm_set_pool( 0, array, rntm );

	// Set the packing block allocator field of the rntm. This will be
	// inherited by all of the child threads when they make local copies of
	// the rntm below.
	bli_membrk_rntm_set_membrk( rntm );

	// Allocate a global communicator for the root thrinfo_t structures.
	thrcomm_t* restrict gl_comm = bli_thrcomm_create( rntm, n_threads );

	// Allocate an array of pthread objects and auxiliary data structs to pass
	// to the thread entry functions.

	#ifdef BLIS_ENABLE_MEM_TRACING
	printf( "bli_l3_thread_decorator().pth: " );
	#endif
	bli_pthread_t* pthreads = bli_malloc_intl( sizeof( bli_pthread_t ) * n_threads );

	#ifdef BLIS_ENABLE_MEM_TRACING
	printf( "bli_l3_thread_decorator().pth: " );
	#endif
	thread_data_t* datas    = bli_malloc_intl( sizeof( thread_data_t ) * n_threads );

	// NOTE: We must iterate backwards so that the chief thread (thread id 0)
	// can spawn all other threads before proceeding with its own computation.
	for ( dim_t tid = n_threads - 1; 0 <= tid; tid-- )
	{
		// Set up thread data for additional threads (beyond thread 0).
		datas[tid].func     = func;
		datas[tid].family   = family;
		datas[tid].schema_a = schema_a;
		datas[tid].schema_b = schema_b;
		datas[tid].alpha    = alpha;
		datas[tid].a        = a;
		datas[tid].b        = b;
		datas[tid].beta     = beta;
		datas[tid].c        = c;
		datas[tid].cntx     = cntx;
		datas[tid].rntm     = rntm;
		datas[tid].cntl     = cntl;
		datas[tid].tid      = tid;
		datas[tid].gl_comm  = gl_comm;
		datas[tid].array    = array;

		// Spawn additional threads for ids greater than 1.
		if ( tid != 0 )
			bli_pthread_create( &pthreads[tid], NULL, &bli_l3_thread_entry, &datas[tid] );
		else
			bli_l3_thread_entry( ( void* )(&datas[0]) );
	}

	// We shouldn't free the global communicator since it was already freed
	// by the global communicator's chief thread in bli_l3_thrinfo_free()
	// (called from the thread entry function).

	// Thread 0 waits for additional threads to finish.
	for ( dim_t tid = 1; tid < n_threads; tid++ )
	{
		bli_pthread_join( pthreads[tid], NULL );
	}

	// Check the array_t back into the small block allocator. Similar to the
	// check-out, this is done using a lock embedded within the sba to ensure
	// mutual exclusion.
	bli_sba_checkin_array( array );

	#ifdef BLIS_ENABLE_MEM_TRACING
	printf( "bli_l3_thread_decorator().pth: " );
	#endif
	bli_free_intl( pthreads );

	#ifdef BLIS_ENABLE_MEM_TRACING
	printf( "bli_l3_thread_decorator().pth: " );
	#endif
	bli_free_intl( datas );
}

#endif