Merge pull request #2627 from nedsels/improve_binary_heap

Update BinaryHeap to FourAryHeap

Author: vaughnbetz

.gitignore

@@ -153,4 +153,4 @@ tags
 .idea
 cmake-build-debug
 cmake-build-release
-/.metadata/
+/.metadata/

doc/src/api/vprinternals/index.rst

@@ -10,3 +10,4 @@ VPR INTERNALS
    vpr_ui
    draw_files
    vpr_noc
+   vpr_router

doc/src/api/vprinternals/router_heap.rst

@@ -0,0 +1,33 @@
+==============
+Router Heap
+==============
+
+t_heap
+----------
+.. doxygenstruct:: t_heap
+   :project: vpr
+   :members:
+
+HeapInterface
+----------
+.. doxygenclass:: HeapInterface
+   :project: vpr
+   :members:
+
+HeapStorage
+----------
+.. doxygenclass:: HeapStorage
+   :project: vpr
+   :members:
+
+KAryHeap
+----------
+.. doxygenclass:: KAryHeap
+   :project: vpr
+   :members:
+
+FourAryHeap
+----------
+.. doxygenclass:: FourAryHeap
+   :project: vpr
+   :members:

doc/src/api/vprinternals/vpr_router.rst

@@ -0,0 +1,10 @@
+.. _router:
+
+=======
+VPR Router
+=======
+
+.. toctree::
+   :maxdepth: 1
+
+   router_heap

utils/route_diag/src/main.cpp

@@ -103,9 +103,9 @@ static void do_one_route(const Netlist<>& net_list,
                                                   segment_inf,
                                                   is_flat);
 
-    ConnectionRouter<BinaryHeap> router(
-            device_ctx.grid,
-            *router_lookahead,
+    ConnectionRouter<FourAryHeap> router(
+        device_ctx.grid,
+        *router_lookahead,
             device_ctx.rr_graph.rr_nodes(),
             &device_ctx.rr_graph,
             device_ctx.rr_rc_data,

vpr/src/base/ShowSetup.cpp

@@ -1061,6 +1061,8 @@ struct ParseRouterHeap {
         ConvertedValue<e_heap_type> conv_value;
         if (str == "binary")
             conv_value.set_value(e_heap_type::BINARY_HEAP);
+        else if (str == "four_ary")
+            conv_value.set_value(e_heap_type::FOUR_ARY_HEAP);
         else if (str == "bucket")
             conv_value.set_value(e_heap_type::BUCKET_HEAP_APPROXIMATION);
         else {
@@ -1075,6 +1077,8 @@ struct ParseRouterHeap {
         ConvertedValue<std::string> conv_value;
         if (val == e_heap_type::BINARY_HEAP)
             conv_value.set_value("binary");
+        else if (val == e_heap_type::FOUR_ARY_HEAP)
+            conv_value.set_value("four_ary");
         else {
             VTR_ASSERT(val == e_heap_type::BUCKET_HEAP_APPROXIMATION);
             conv_value.set_value("bucket");
@@ -1083,7 +1087,7 @@ struct ParseRouterHeap {
     }
 
     std::vector<std::string> default_choices() {
-        return {"binary", "bucket"};
+        return {"binary", "four_ary", "bucket"};
     }
 };
 
@@ -2648,11 +2652,12 @@ argparse::ArgumentParser create_arg_parser(const std::string& prog_name, t_optio
         .help(
             "Controls what type of heap to use for timing driven router.\n"
             " * binary: A binary heap is used.\n"
+            " * four_ary: A four_ary heap is used.\n"
             " * bucket: A bucket heap approximation is used. The bucket heap\n"
             " *         is faster because it is only a heap approximation.\n"
             " *         Testing has shown the approximation results in\n"
-            " *         similiar QoR with less CPU work.\n")
-        .default_value("binary")
+            " *         similar QoR with less CPU work.\n")
+        .default_value("four_ary")
         .show_in(argparse::ShowIn::HELP_ONLY);
 
     route_timing_grp.add_argument(args.router_first_iteration_timing_report_file, "--router_first_iter_timing_report")

vpr/src/base/read_options.cpp

@@ -46,7 +46,7 @@
 #include "route_common.h"
 #include "route_tree.h"
 #include "read_route.h"
-#include "binary_heap.h"
+#include "four_ary_heap.h"
 
 #include "old_traceback.h"
 
@@ -129,7 +129,7 @@ bool read_route(const char* route_file, const t_router_opts& router_opts, bool v
     fp.close();
 
     /*Correctly set up the clb opins*/
-    BinaryHeap small_heap;
+    FourAryHeap small_heap;
     small_heap.init_heap(device_ctx.grid);
     if (!flat_router) {
         reserve_locally_used_opins(&small_heap, router_opts.initial_pres_fac,

vpr/src/base/read_route.cpp

@@ -1,57 +1,23 @@
 #include "binary_heap.h"
-#include "rr_graph_fwd.h"
 #include "vtr_log.h"
 
-static size_t parent(size_t i) { return i >> 1; }
 // child indices of a heap
-static size_t left(size_t i) { return i << 1; }
-static size_t right(size_t i) { return (i << 1) + 1; }
-
-BinaryHeap::BinaryHeap()
-    : heap_()
-    , heap_size_(0)
-    , heap_tail_(0)
-    , max_index_(std::numeric_limits<size_t>::max())
-    , prune_limit_(std::numeric_limits<size_t>::max()) {}
-
-BinaryHeap::~BinaryHeap() {
-    free_all_memory();
-}
+static inline size_t left(size_t i) { return i << 1; }
+static inline size_t right(size_t i) { return (i << 1) + 1; }
 
-t_heap* BinaryHeap::alloc() {
-    return storage_.alloc();
-}
-void BinaryHeap::free(t_heap* hptr) {
-    storage_.free(hptr);
-}
+inline size_t BinaryHeap::parent(size_t i) const { return i >> 1; }
 
-void BinaryHeap::init_heap(const DeviceGrid& grid) {
-    size_t target_heap_size = (grid.width() - 1) * (grid.height() - 1);
-    if (heap_.empty() || heap_size_ < target_heap_size) {
-        if (!heap_.empty()) {
-            // coverity[offset_free : Intentional]
-            heap_.clear();
-        }
-        heap_size_ = (grid.width() - 1) * (grid.height() - 1);
-        heap_.resize(heap_size_ + 1); /* heap_size_ + 1 because heap stores from [1..heap_size] */
+bool BinaryHeap::is_valid() const {
+    if (heap_.empty()) {
+        return false;
     }
-    heap_tail_ = 1;
-}
-
-void BinaryHeap::add_to_heap(t_heap* hptr) {
-    expand_heap_if_full();
-    // start with undefined hole
-    ++heap_tail_;
-    sift_up(heap_tail_ - 1, hptr);
 
-    // If we have pruned, rebuild the heap now.
-    if (check_prune_limit()) {
-        build_heap();
+    for (size_t i = 1; i <= heap_tail_ >> 1; ++i) {
+        if (left(i) < heap_tail_ && heap_[left(i)].cost < heap_[i].cost) return false;
+        if (right(i) < heap_tail_ && heap_[right(i)].cost < heap_[i].cost) return false;
     }
-}
 
-bool BinaryHeap::is_empty_heap() const {
-    return (bool)(heap_tail_ == 1);
+    return true;
 }
 
 t_heap* BinaryHeap::get_heap_head() {
@@ -68,173 +34,44 @@ t_heap* BinaryHeap::get_heap_head() {
             return (nullptr);
         }
 
-        cheapest = heap_[1];
+        cheapest = heap_[1].elem_ptr;
 
         hole = 1;
         child = 2;
+
         --heap_tail_;
+
         while (child < heap_tail_) {
-            if (heap_[child + 1]->cost < heap_[child]->cost)
+            if (heap_[child + 1].cost < heap_[child].cost)
                 ++child; // become right child
+
             heap_[hole] = heap_[child];
             hole = child;
             child = left(child);
         }
-        sift_up(hole, heap_[heap_tail_]);
 
+        sift_up(hole, heap_[heap_tail_]);
     } while (!cheapest->index.is_valid()); /* Get another one if invalid entry. */
 
     return (cheapest);
 }
 
-void BinaryHeap::empty_heap() {
-    for (size_t i = 1; i < heap_tail_; i++)
-        free(heap_[i]);
-
-    heap_tail_ = 1;
-}
-
-size_t BinaryHeap::size() const { return heap_tail_ - 1; } // heap[0] is not valid element
-
 // make a heap rooted at index hole by **sifting down** in O(lgn) time
 void BinaryHeap::sift_down(size_t hole) {
-    t_heap* head{heap_[hole]};
+    heap_elem head{heap_[hole]};
     size_t child{left(hole)};
+
     while (child < heap_tail_) {
-        if (child + 1 < heap_tail_ && heap_[child + 1]->cost < heap_[child]->cost)
+        if (child + 1 < heap_tail_ && heap_[child + 1].cost < heap_[child].cost)
             ++child;
-        if (heap_[child]->cost < head->cost) {
+
+        if (heap_[child].cost < head.cost) {
             heap_[hole] = heap_[child];
             hole = child;
             child = left(child);
         } else
             break;
     }
-    heap_[hole] = head;
-}
-
-// runs in O(n) time by sifting down; the least work is done on the most elements: 1 swap for bottom layer, 2 swap for 2nd, ... lgn swap for top
-// 1*(n/2) + 2*(n/4) + 3*(n/8) + ... + lgn*1 = 2n (sum of i/2^i)
-void BinaryHeap::build_heap() {
-    // second half of heap are leaves
-    for (size_t i = heap_tail_ >> 1; i != 0; --i)
-        sift_down(i);
-}
-
-void BinaryHeap::set_prune_limit(size_t max_index, size_t prune_limit) {
-    if (prune_limit != std::numeric_limits<size_t>::max()) {
-        VTR_ASSERT(max_index < prune_limit);
-    }
-    max_index_ = max_index;
-    prune_limit_ = prune_limit;
-}
-
-// O(lgn) sifting up to maintain heap property after insertion (should sift down when building heap)
-void BinaryHeap::sift_up(size_t leaf, t_heap* const node) {
-    while ((leaf > 1) && (node->cost < heap_[parent(leaf)]->cost)) {
-        // sift hole up
-        heap_[leaf] = heap_[parent(leaf)];
-        leaf = parent(leaf);
-    }
-    heap_[leaf] = node;
-}
-
-//expands heap by "realloc"
-void BinaryHeap::expand_heap_if_full() {
-    if (heap_tail_ > heap_size_) { /* Heap is full */
-        heap_size_ *= 2;
-        heap_.resize(heap_size_ + 1);
-    }
-}
 
-// adds an element to the back of heap and expand if necessary, but does not maintain heap property
-void BinaryHeap::push_back(t_heap* const hptr) {
-    expand_heap_if_full();
-    heap_[heap_tail_] = hptr;
-    ++heap_tail_;
-
-    check_prune_limit();
-}
-
-bool BinaryHeap::is_valid() const {
-    if (heap_.empty()) {
-        return false;
-    }
-
-    for (size_t i = 1; i <= heap_tail_ >> 1; ++i) {
-        if (left(i) < heap_tail_ && heap_[left(i)]->cost < heap_[i]->cost) return false;
-        if (right(i) < heap_tail_ && heap_[right(i)]->cost < heap_[i]->cost) return false;
-    }
-    return true;
-}
-
-void BinaryHeap::free_all_memory() {
-    if (!heap_.empty()) {
-        empty_heap();
-
-        // coverity[offset_free : Intentional]
-        heap_.clear();
-    }
-
-    //  heap_ = nullptr; /* Defensive coding:  crash hard if I use these. */
-
-    storage_.free_all_memory();
-}
-
-bool BinaryHeap::check_prune_limit() {
-    if (heap_tail_ > prune_limit_) {
-        prune_heap();
-        return true;
-    }
-
-    return false;
-}
-
-void BinaryHeap::prune_heap() {
-    VTR_ASSERT(max_index_ < prune_limit_);
-
-    std::vector<t_heap*> best_heap_item(max_index_, nullptr);
-
-    // Find the cheapest instance of each index and store it.
-    for (size_t i = 1; i < heap_tail_; i++) {
-        if (heap_[i] == nullptr) {
-            continue;
-        }
-
-        if (!heap_[i]->index.is_valid()) {
-            free(heap_[i]);
-            heap_[i] = nullptr;
-            continue;
-        }
-
-        auto idx = size_t(heap_[i]->index);
-
-        VTR_ASSERT(idx < max_index_);
-
-        if (best_heap_item[idx] == nullptr || best_heap_item[idx]->cost > heap_[i]->cost) {
-            best_heap_item[idx] = heap_[i];
-        }
-    }
-
-    // Free unused nodes.
-    for (size_t i = 1; i < heap_tail_; i++) {
-        if (heap_[i] == nullptr) {
-            continue;
-        }
-
-        auto idx = size_t(heap_[i]->index);
-
-        if (best_heap_item[idx] != heap_[i]) {
-            free(heap_[i]);
-            heap_[i] = nullptr;
-        }
-    }
-
-    heap_tail_ = 1;
-
-    for (size_t i = 0; i < max_index_; ++i) {
-        if (best_heap_item[i] != nullptr) {
-            heap_[heap_tail_++] = best_heap_item[i];
-        }
-    }
-}
+    heap_[hole] = head;
+}