Пример #1
0
    def __init__(self, *args, **kwds):
        TemporaryManager.__init__(self, *args, **kwds)

        self._virtual_allocations = {} # id -> VirtualAllocation
        self._real_sizes = SortedCollection(key=lambda x: x.size) # RealSize objects, sorted by size
        self._virtual_to_real = {} # id -> VirtualMapping
        self._real_allocations = {} # real_id -> RealAllocation
        self._real_id_counter = 0
Пример #2
0
class ZeroOffsetManager(TemporaryManager):
    """
    Tries to assign several allocation requests to a single real allocation,
    if dependencies allow that.
    All virtual allocations start from the beginning of real allocations.
    """

    VirtualAllocation = collections.namedtuple('VirtualAllocation',
                                               ['size', 'dependencies'])
    RealAllocation = collections.namedtuple('RealAllocation',
                                            ['buffer', 'virtual_ids'])
    RealSize = collections.namedtuple('RealSize', ['size', 'real_id'])
    VirtualMapping = collections.namedtuple('VirtualMapping',
                                            ['real_id', 'sub_region'])

    def __init__(self, *args, **kwds):
        TemporaryManager.__init__(self, *args, **kwds)

        self._virtual_allocations = {}  # id -> VirtualAllocation
        self._real_sizes = SortedCollection(
            key=lambda x: x.size)  # RealSize objects, sorted by size
        self._virtual_to_real = {}  # id -> VirtualMapping
        self._real_allocations = {}  # real_id -> RealAllocation
        self._real_id_counter = 0

    def _allocate(self, new_id, size, dependencies, pack):

        # Dependencies should be bidirectional.
        # So if some new allocation says it depends on earlier ones,
        # we need to update their dependency lists.
        dep_set = set(dependencies)
        for dep in dependencies:
            if dep in self._virtual_allocations:
                self._virtual_allocations[dep].dependencies.add(new_id)
            else:
                dep_set.remove(dep)

        # Save virtual allocation parameters
        self._virtual_allocations[new_id] = self.VirtualAllocation(
            size, dep_set)

        if pack:
            # If pack is requested, we can just do full re-pack right away.
            self._pack()
        else:
            # If not, find a real allocation using the greedy algorithm.
            self._fast_add(new_id, size, dep_set)

    def _fast_add(self, new_id, size, dep_set):
        """
        Greedy algorithm to find a real allocation for a given virtual allocation.
        """

        # Find the smallest real allocation which can hold the requested virtual allocation.
        try:
            idx_start = self._real_sizes.argfind_ge(size)
        except ValueError:
            idx_start = len(self._real_sizes)

        # Check all real allocations with suitable sizes, starting from the smallest one.
        # Use the first real allocation which does not contain ``new_id``'s dependencies.
        for idx in range(idx_start, len(self._real_sizes)):
            real_id = self._real_sizes[idx].real_id
            buf = self._real_allocations[real_id].buffer
            virtual_ids = self._real_allocations[real_id].virtual_ids
            if virtual_ids.isdisjoint(dep_set):
                virtual_ids.add(new_id)
                break
        else:
            # If no suitable real allocation is found, create a new one.
            buf = self._thr.allocate(size)
            real_id = self._real_id_counter
            self._real_id_counter += 1

            self._real_allocations[real_id] = self.RealAllocation(
                buf, set([new_id]))
            self._real_sizes.insert(self.RealSize(size, real_id))

        # Here it would be more appropriate to use buffer.get_sub_region(0, size),
        # but OpenCL does not allow several overlapping subregions to be used in a single kernel
        # for both read and write, which ruins the whole idea.
        # So we are passing full buffers and hope that overlapping Array class takes care of sizes.
        self._virtual_to_real[new_id] = self.VirtualMapping(
            real_id, self._real_allocations[real_id].buffer)

    def _get_buffer(self, id_):
        return self._virtual_to_real[id_].sub_region

    def _free(self, id_, pack=False):
        # Remove the allocation from the dependency lists of its dependencies
        dep_set = self._virtual_allocations[id_].dependencies
        for dep in dep_set:
            self._virtual_allocations[dep].dependencies.remove(id_)

        vtr = self._virtual_to_real[id_]

        # Clear virtual allocation data
        del self._virtual_allocations[id_]
        del self._virtual_to_real[id_]

        if pack:
            self._pack()
        else:
            # Fast and non-optimal free.
            # Remove the virtual allocation from the real allocation,
            # and delete the real allocation if its no longer used by other virtual allocations.
            ra = self._real_allocations[vtr.real_id]
            ra.virtual_ids.remove(id_)
            if len(ra.virtual_ids) == 0:
                del self._real_allocations[vtr.real_id]
                self._real_sizes.remove(
                    self.RealSize(ra.buffer.size, vtr.real_id))

    def _pack(self):
        """
        Full memory re-pack.
        In theory, should find the optimal (with the minimal real allocation size) distribution
        of virtual allocations.
        """

        # Need to synchronize, because we are going to change allocation addresses,
        # and we do not want to free the memory some kernel is reading from.
        self._thr.synchronize()

        # Clear all real allocation data.
        self._real_sizes.clear()
        self._real_allocations = {}
        self._real_id_counter = 0

        va = self._virtual_allocations

        # Sort all virtual allocations by size
        virtual_sizes = sorted([(va[id_].size, id_) for id_ in va],
                               key=lambda x: x[0])

        # Application of greedy algorithm for virtual allocations starting from the largest one
        # should give the optimal distribution.
        for size, id_ in reversed(virtual_sizes):
            self._fast_add(id_, size,
                           self._virtual_allocations[id_].dependencies)

    def _statistics(self):

        stats = dict(virtual_size_total=0,
                     virtual_num=0,
                     real_size_total=0,
                     real_num=0,
                     virtual_sizes=[],
                     real_sizes=[])

        for va in self._virtual_allocations.values():
            stats['virtual_size_total'] += va.size
            stats['virtual_num'] += 1
            stats['virtual_sizes'].append(va.size)

        for ra in self._real_allocations.values():
            stats['real_size_total'] += ra.buffer.size
            stats['real_num'] += 1
            stats['real_sizes'].append(ra.buffer.size)

        stats['virtual_sizes'] = sorted(stats['virtual_sizes'])
        stats['real_sizes'] = sorted(stats['real_sizes'])

        return stats
Пример #3
0
class ZeroOffsetManager(TemporaryManager):
    """
    Tries to assign several allocation requests to a single real allocation,
    if dependencies allow that.
    All virtual allocations start from the beginning of real allocations.
    """

    VirtualAllocation = collections.namedtuple('VirtualAllocation', ['size', 'dependencies'])
    RealAllocation = collections.namedtuple('RealAllocation', ['buffer', 'virtual_ids'])
    RealSize = collections.namedtuple('RealSize', ['size', 'real_id'])
    VirtualMapping = collections.namedtuple('VirtualMapping', ['real_id', 'sub_region'])

    def __init__(self, *args, **kwds):
        TemporaryManager.__init__(self, *args, **kwds)

        self._virtual_allocations = {} # id -> VirtualAllocation
        self._real_sizes = SortedCollection(key=lambda x: x.size) # RealSize objects, sorted by size
        self._virtual_to_real = {} # id -> VirtualMapping
        self._real_allocations = {} # real_id -> RealAllocation
        self._real_id_counter = 0

    def _allocate(self, new_id, size, dependencies, pack):

        # Dependencies should be bidirectional.
        # So if some new allocation says it depends on earlier ones,
        # we need to update their dependency lists.
        dep_set = set(dependencies)
        for dep in dependencies:
            if dep in self._virtual_allocations:
                self._virtual_allocations[dep].dependencies.add(new_id)
            else:
                dep_set.remove(dep)

        # Save virtual allocation parameters
        self._virtual_allocations[new_id] = self.VirtualAllocation(size, dep_set)

        if pack:
            # If pack is requested, we can just do full re-pack right away.
            self._pack()
        else:
            # If not, find a real allocation using the greedy algorithm.
            self._fast_add(new_id, size, dep_set)

    def _fast_add(self, new_id, size, dep_set):
        """
        Greedy algorithm to find a real allocation for a given virtual allocation.
        """

        # Find the smallest real allocation which can hold the requested virtual allocation.
        try:
            idx_start = self._real_sizes.argfind_ge(size)
        except ValueError:
            idx_start = len(self._real_sizes)

        # Check all real allocations with suitable sizes, starting from the smallest one.
        # Use the first real allocation which does not contain ``new_id``'s dependencies.
        for idx in range(idx_start, len(self._real_sizes)):
            real_id = self._real_sizes[idx].real_id
            buf = self._real_allocations[real_id].buffer
            virtual_ids = self._real_allocations[real_id].virtual_ids
            if virtual_ids.isdisjoint(dep_set):
                virtual_ids.add(new_id)
                break
        else:
            # If no suitable real allocation is found, create a new one.
            buf = self._thr.allocate(size)
            real_id = self._real_id_counter
            self._real_id_counter += 1

            self._real_allocations[real_id] = self.RealAllocation(buf, set([new_id]))
            self._real_sizes.insert(self.RealSize(size, real_id))

        # Here it would be more appropriate to use buffer.get_sub_region(0, size),
        # but OpenCL does not allow several overlapping subregions to be used in a single kernel
        # for both read and write, which ruins the whole idea.
        # So we are passing full buffers and hope that overlapping Array class takes care of sizes.
        self._virtual_to_real[new_id] = self.VirtualMapping(
            real_id, self._real_allocations[real_id].buffer)

    def _get_buffer(self, id_):
        return self._virtual_to_real[id_].sub_region

    def _free(self, id_, pack=False):
        # Remove the allocation from the dependency lists of its dependencies
        dep_set = self._virtual_allocations[id_].dependencies
        for dep in dep_set:
            self._virtual_allocations[dep].dependencies.remove(id_)

        vtr = self._virtual_to_real[id_]

        # Clear virtual allocation data
        del self._virtual_allocations[id_]
        del self._virtual_to_real[id_]

        if pack:
            self._pack()
        else:
            # Fast and non-optimal free.
            # Remove the virtual allocation from the real allocation,
            # and delete the real allocation if its no longer used by other virtual allocations.
            ra = self._real_allocations[vtr.real_id]
            ra.virtual_ids.remove(id_)
            if len(ra.virtual_ids) == 0:
                del self._real_allocations[vtr.real_id]
                self._real_sizes.remove(self.RealSize(ra.buffer.size, vtr.real_id))

    def _pack(self):
        """
        Full memory re-pack.
        In theory, should find the optimal (with the minimal real allocation size) distribution
        of virtual allocations.
        """

        # Need to synchronize, because we are going to change allocation addresses,
        # and we do not want to free the memory some kernel is reading from.
        self._thr.synchronize()

        # Clear all real allocation data.
        self._real_sizes.clear()
        self._real_allocations = {}
        self._real_id_counter = 0

        va = self._virtual_allocations

        # Sort all virtual allocations by size
        virtual_sizes = sorted(
            [(va[id_].size, id_) for id_ in va],
            key=lambda x: x[0])

        # Application of greedy algorithm for virtual allocations starting from the largest one
        # should give the optimal distribution.
        for size, id_ in reversed(virtual_sizes):
            self._fast_add(id_, size, self._virtual_allocations[id_].dependencies)

    def _statistics(self):

        stats = dict(
            virtual_size_total=0,
            virtual_num=0,
            real_size_total=0,
            real_num=0,
            virtual_sizes=[],
            real_sizes=[])

        for va in self._virtual_allocations.values():
            stats['virtual_size_total'] += va.size
            stats['virtual_num'] += 1
            stats['virtual_sizes'].append(va.size)

        for ra in self._real_allocations.values():
            stats['real_size_total'] += ra.buffer.size
            stats['real_num'] += 1
            stats['real_sizes'].append(ra.buffer.size)

        stats['virtual_sizes'] = sorted(stats['virtual_sizes'])
        stats['real_sizes'] = sorted(stats['real_sizes'])

        return stats