Merge pull request #1962 from RapidSilicon/rr_nodes

Clean-up DeviceContext (Remove shadowed data structure)

Author: tangxifan

utils/route_diag/src/main.cpp

@@ -101,7 +101,7 @@ static void do_one_route(int source_node, int sink_node,
     ConnectionRouter<BinaryHeap> router(
             device_ctx.grid,
             *router_lookahead,
-            device_ctx.rr_nodes,
+            device_ctx.rr_graph.rr_nodes(),
             &device_ctx.rr_graph,
             device_ctx.rr_rc_data,
             device_ctx.rr_graph.rr_switch(),

vpr/src/base/read_route.cpp

@@ -126,7 +126,7 @@ bool read_route(const char* route_file, const t_router_opts& router_opts, bool v
     recompute_occupancy_from_scratch();
 
     /* Note: This pres_fac is not necessarily correct since it isn't the first routing iteration*/
-    OveruseInfo overuse_info(device_ctx.rr_nodes.size());
+    OveruseInfo overuse_info(device_ctx.rr_graph.num_nodes());
     pathfinder_update_acc_cost_and_overuse_info(router_opts.acc_fac, overuse_info);
 
     reserve_locally_used_opins(&small_heap, router_opts.initial_pres_fac,
@@ -505,21 +505,19 @@ static bool check_rr_graph_connectivity(RRNodeId prev_node, RRNodeId node) {
     if (prev_node == node) return false;
 
     auto& device_ctx = g_vpr_ctx.device();
-    const auto& rr_graph = device_ctx.rr_nodes;
-    /*TODO We need to remove temp_rr_graph once rr_graph is resolved*/
-    const auto& temp_rr_graph = device_ctx.rr_graph;
+    const auto& rr_graph = device_ctx.rr_graph;
     // If it's starting a new sub branch this is ok
-    if (device_ctx.rr_graph.node_type(prev_node) == SINK) return true;
+    if (rr_graph.node_type(prev_node) == SINK) return true;
 
     for (RREdgeId edge : rr_graph.edge_range(prev_node)) {
         //If the sink node is reachable by previous node return true
-        if (rr_graph.edge_sink_node(edge) == node) {
+        if (rr_graph.rr_nodes().edge_sink_node(edge) == node) {
             return true;
         }
 
         // If there are any non-configurable branches return true
-        short edge_switch = rr_graph.edge_switch(edge);
-        if (!(temp_rr_graph.rr_switch_inf(RRSwitchId(edge_switch)).configurable())) return true;
+        short edge_switch = rr_graph.rr_nodes().edge_switch(edge);
+        if (!(rr_graph.rr_switch_inf(RRSwitchId(edge_switch)).configurable())) return true;
     }
 
     // If it's part of a non configurable node list, return true

vpr/src/base/vpr_context.h

@@ -147,7 +147,6 @@ struct DeviceContext : public Context {
     /*
      * Structures to define the routing architecture of the FPGA.
      */
-    t_rr_graph_storage rr_nodes; // autogenerated in build_rr_graph
 
     vtr::vector<RRIndexedDataId, t_rr_indexed_data> rr_indexed_data; // [0 .. num_rr_indexed_data-1]
 
@@ -163,19 +162,12 @@ struct DeviceContext : public Context {
     /* A writeable view of routing resource graph to be the ONLY database 
      * for routing resource graph builder functions.
      */
-    RRGraphBuilder rr_graph_builder{&rr_nodes};
+    RRGraphBuilder rr_graph_builder{};
 
     /* A read-only view of routing resource graph to be the ONLY database 
      * for client functions: GUI, placer, router, timing analyzer etc.
      */
-    RRGraphView rr_graph{rr_nodes,
-                         rr_graph_builder.node_lookup(),
-                         rr_graph_builder.rr_node_metadata(),
-                         rr_graph_builder.rr_edge_metadata(),
-                         rr_indexed_data,
-                         rr_graph_builder.rr_segments(),
-                         rr_graph_builder.rr_switch()};
-
+    RRGraphView rr_graph{rr_graph_builder.rr_nodes(), rr_graph_builder.node_lookup(), rr_graph_builder.rr_node_metadata(), rr_graph_builder.rr_edge_metadata(), rr_indexed_data, rr_graph_builder.rr_segments(), rr_graph_builder.rr_switch()};
     int num_arch_switches;
     t_arch_switch_inf* arch_switch_inf; // [0..(num_arch_switches-1)]
 
@@ -194,7 +186,6 @@ struct DeviceContext : public Context {
      *      a single value
      */
     int virtual_clock_network_root_idx;
-
     /**
      * @brief switch_fanin_remap is only used for printing out switch fanin stats
      *        (the -switch_stats option)

vpr/src/device/rr_graph_builder.cpp

@@ -7,11 +7,9 @@
 
 //#include "globals.h"
 
-RRGraphBuilder::RRGraphBuilder(t_rr_graph_storage* node_storage)
-    : node_storage_(*node_storage) {
-}
+RRGraphBuilder::RRGraphBuilder() {}
 
-t_rr_graph_storage& RRGraphBuilder::node_storage() {
+t_rr_graph_storage& RRGraphBuilder::rr_nodes() {
     return node_storage_;
 }
 
@@ -63,6 +61,7 @@ void RRGraphBuilder::add_node_to_all_locs(RRNodeId node) {
 
 void RRGraphBuilder::clear() {
     node_lookup_.clear();
+    node_storage_.clear();
     rr_node_metadata_.clear();
     rr_edge_metadata_.clear();
     rr_segments_.clear();

vpr/src/device/rr_graph_builder.h

@@ -21,7 +21,7 @@ class RRGraphBuilder {
     /* -- Constructors -- */
   public:
     /* See detailed comments about the data structures in the internal data storage section of this file */
-    RRGraphBuilder(t_rr_graph_storage* node_storage);
+    RRGraphBuilder();
 
     /* Disable copy constructors and copy assignment operator
      * This is to avoid accidental copy because it could be an expensive operation considering that the 
@@ -35,7 +35,7 @@ class RRGraphBuilder {
     /* -- Mutators -- */
   public:
     /** @brief Return a writable object for rr_nodes */
-    t_rr_graph_storage& node_storage();
+    t_rr_graph_storage& rr_nodes();
     /** @brief Return a writable object for update the fast look-up of rr_node */
     RRSpatialLookup& node_lookup();
     /** .. warning:: The Metadata should stay as an independent data structure than rest of the internal data,
@@ -212,6 +212,11 @@ class RRGraphBuilder {
     inline void emplace_back_edge(RRNodeId src, RRNodeId dest, short edge_switch) {
         node_storage_.emplace_back_edge(src, dest, edge_switch);
     }
+    /** @brief Append 1 more RR node to the RR graph. */
+    inline void emplace_back() {
+        // No edges can be assigned if mutating the rr node array.
+        node_storage_.emplace_back();
+    }
 
     /** @brief alloc_and_load_edges; It adds a batch of edges.  */
     inline void alloc_and_load_edges(const t_rr_edge_info_set* rr_edges_to_create) {
@@ -316,7 +321,7 @@ class RRGraphBuilder {
      * That explains why the reference is used here temporarily
      */
     /* node-level storage including edge storages */
-    t_rr_graph_storage& node_storage_;
+    t_rr_graph_storage node_storage_;
     /* Fast look-up for rr nodes */
     RRSpatialLookup node_lookup_;
 

vpr/src/device/rr_graph_view.h

@@ -69,13 +69,23 @@ class RRGraphView {
      *      } 
      */
     inline vtr::StrongIdRange<RRNodeId> nodes() const {
-        return vtr::StrongIdRange<RRNodeId>(RRNodeId(0), RRNodeId(size()));
+        return vtr::StrongIdRange<RRNodeId>(RRNodeId(0), RRNodeId(num_nodes()));
     }
 
     /** @brief Return number of nodes. This function is inlined for runtime optimization. */
-    inline size_t size() const {
+    inline size_t num_nodes() const {
         return node_storage_.size();
     }
+    /** @brief Is the RR graph currently empty? */
+    inline bool empty() const {
+        return node_storage_.empty();
+    }
+
+    /** @brief Returns a range of RREdgeId's belonging to RRNodeId id.
+     * If this range is empty, then RRNodeId id has no edges.*/
+    inline vtr::StrongIdRange<RREdgeId> edge_range(RRNodeId id) const {
+        return vtr::StrongIdRange<RREdgeId>(node_storage_.first_edge(id), node_storage_.last_edge(id));
+    }
 
     /** @brief Get the type of a routing resource node. This function is inlined for runtime optimization. */
     inline t_rr_type node_type(RRNodeId node) const {
@@ -421,6 +431,11 @@ class RRGraphView {
     const RRSpatialLookup& node_lookup() const {
         return node_lookup_;
     }
+    /** @brief Return the node-level storage structure for queries from client functions */
+    inline const t_rr_graph_storage& rr_nodes() const {
+        return node_storage_;
+    }
+
     /** .. warning:: The Metadata should stay as an independent data structure than rest of the internal data,
      *  e.g., node_lookup! */
     MetadataStorage<int> rr_node_metadata_data() const {
@@ -464,6 +479,7 @@ class RRGraphView {
     const MetadataStorage<std::tuple<int, int, short>>& rr_edge_metadata_;
     /* rr_indexed_data_ and rr_segments_ are needed to lookup the segment information in  node_coordinate_to_string() */
     const vtr::vector<RRIndexedDataId, t_rr_indexed_data>& rr_indexed_data_;
+
     /* Segment info for rr nodes */
     const vtr::vector<RRSegmentId, t_segment_inf>& rr_segments_;
     /* switch info for rr nodes */

vpr/src/draw/draw.cpp

@@ -929,7 +929,7 @@ void alloc_draw_structs(const t_arch* arch) {
     /* Space is allocated for draw_rr_node but not initialized because we do *
      * not yet know information about the routing resources.				  */
     draw_state->draw_rr_node = (t_draw_rr_node*)vtr::malloc(
-        device_ctx.rr_nodes.size() * sizeof(t_draw_rr_node));
+        device_ctx.rr_graph.num_nodes() * sizeof(t_draw_rr_node));
 
     draw_state->arch_info = arch;
 
@@ -980,10 +980,10 @@ void init_draw_coords(float width_val) {
         return; //do not initialize only if --disp off and --save_graphics off
     /* Each time routing is on screen, need to reallocate the color of each *
      * rr_node, as the number of rr_nodes may change.						*/
-    if (device_ctx.rr_nodes.size() != 0) {
+    if (rr_graph.num_nodes() != 0) {
         draw_state->draw_rr_node = (t_draw_rr_node*)vtr::realloc(
             draw_state->draw_rr_node,
-            (device_ctx.rr_nodes.size()) * sizeof(t_draw_rr_node));
+            (rr_graph.num_nodes()) * sizeof(t_draw_rr_node));
         /*FIXME: the type cast should be eliminated by making draw_rr_node adapt RRNodeId */
         for (const RRNodeId& rr_id : rr_graph.nodes()) {
             draw_state->draw_rr_node[(size_t)rr_id].color = DEFAULT_RR_NODE_COLOR;
@@ -1310,7 +1310,7 @@ static void draw_routing_costs(ezgl::renderer* g) {
 
     float min_cost = std::numeric_limits<float>::infinity();
     float max_cost = -min_cost;
-    std::vector<float> rr_node_costs(device_ctx.rr_nodes.size(), 0.);
+    std::vector<float> rr_node_costs(device_ctx.rr_graph.num_nodes(), 0.);
 
     for (const RRNodeId& rr_id : device_ctx.rr_graph.nodes()) {
         float cost = 0.;
@@ -1694,7 +1694,7 @@ static void draw_rr_edges(int inode, ezgl::renderer* g) {
         to_node = size_t(rr_graph.edge_sink_node(rr_node, iedge));
         to_type = rr_graph.node_type(RRNodeId(to_node));
         to_ptc_num = rr_graph.node_ptc_num(RRNodeId(to_node));
-        bool edge_configurable = device_ctx.rr_nodes[inode].edge_is_configurable(iedge);
+        bool edge_configurable = rr_graph.rr_nodes()[inode].edge_is_configurable(iedge);
 
         switch (from_type) {
             case OPIN:
@@ -2288,7 +2288,7 @@ static void draw_rr_pin(int inode, const ezgl::color& color, ezgl::renderer* g)
  * the physical pin is on.                                                  */
 void draw_get_rr_pin_coords(int inode, float* xcen, float* ycen, const e_side& pin_side) {
     auto& device_ctx = g_vpr_ctx.device();
-    draw_get_rr_pin_coords(device_ctx.rr_nodes[inode], xcen, ycen, pin_side);
+    draw_get_rr_pin_coords(device_ctx.rr_graph.rr_nodes()[inode], xcen, ycen, pin_side);
 }
 
 void draw_get_rr_pin_coords(const t_rr_node& node, float* xcen, float* ycen, const e_side& pin_side) {
@@ -2357,7 +2357,7 @@ static void draw_rr_src_sink(int inode, ezgl::color color, ezgl::renderer* g) {
     const auto& rr_graph = device_ctx.rr_graph;
 
     float xcen, ycen;
-    draw_get_rr_src_sink_coords(device_ctx.rr_nodes[inode], &xcen, &ycen);
+    draw_get_rr_src_sink_coords(rr_graph.rr_nodes()[inode], &xcen, &ycen);
 
     g->set_color(color);
 
@@ -2786,7 +2786,7 @@ static int draw_check_rr_node_hit(float click_x, float click_y) {
             case SOURCE:
             case SINK: {
                 float xcen, ycen;
-                draw_get_rr_src_sink_coords(device_ctx.rr_nodes[inode], &xcen, &ycen);
+                draw_get_rr_src_sink_coords(rr_graph.rr_nodes()[inode], &xcen, &ycen);
 
                 // Now check if we clicked on this pin
                 if (click_x >= xcen - draw_coords->pin_size && click_x <= xcen + draw_coords->pin_size && click_y >= ycen - draw_coords->pin_size && click_y <= ycen + draw_coords->pin_size) {
@@ -2829,7 +2829,7 @@ void draw_expand_non_configurable_rr_nodes_recurr(int from_node,
 
     for (t_edge_size iedge = 0;
          iedge < rr_graph.num_edges(RRNodeId(from_node)); ++iedge) {
-        bool edge_configurable = device_ctx.rr_nodes[from_node].edge_is_configurable(iedge);
+        bool edge_configurable = rr_graph.rr_nodes()[from_node].edge_is_configurable(iedge);
         int to_node = size_t(rr_graph.edge_sink_node(RRNodeId(from_node), iedge));
 
         if (!edge_configurable && !expanded_nodes.count(to_node)) {
@@ -3355,7 +3355,7 @@ static void draw_pin_to_sink(int ipin_node, int sink_node, ezgl::renderer* g) {
         draw_get_rr_pin_coords(ipin_node, &x1, &y1, pin_side);
 
         float x2 = 0, y2 = 0;
-        draw_get_rr_src_sink_coords(device_ctx.rr_nodes[sink_node], &x2, &y2);
+        draw_get_rr_src_sink_coords(rr_graph.rr_nodes()[sink_node], &x2, &y2);
 
         g->draw_line({x1, y1}, {x2, y2});
 
@@ -3370,7 +3370,7 @@ static void draw_source_to_pin(int source_node, int opin_node, ezgl::renderer* g
     const auto& rr_graph = device_ctx.rr_graph;
 
     float x1 = 0, y1 = 0;
-    draw_get_rr_src_sink_coords(device_ctx.rr_nodes[source_node], &x1, &y1);
+    draw_get_rr_src_sink_coords(rr_graph.rr_nodes()[source_node], &x1, &y1);
 
     /* Draw the line for each ipin on different sides */
     for (const e_side& pin_side : SIDES) {
@@ -4141,7 +4141,7 @@ static void draw_router_expansion_costs(ezgl::renderer* g) {
     auto& device_ctx = g_vpr_ctx.device();
     auto& routing_ctx = g_vpr_ctx.routing();
 
-    std::vector<float> rr_costs(device_ctx.rr_nodes.size());
+    std::vector<float> rr_costs(device_ctx.rr_graph.num_nodes());
 
     for (const RRNodeId& rr_id : device_ctx.rr_graph.nodes()) {
         float cost = get_router_expansion_cost(
@@ -4198,11 +4198,11 @@ static void draw_rr_costs(ezgl::renderer* g, const std::vector<float>& rr_costs,
                        || draw_state->show_router_expansion_cost == DRAW_ROUTER_EXPANSION_COST_KNOWN_WITH_EDGES
                        || draw_state->show_router_expansion_cost == DRAW_ROUTER_EXPANSION_COST_EXPECTED_WITH_EDGES);
 
-    VTR_ASSERT(rr_costs.size() == device_ctx.rr_nodes.size());
+    VTR_ASSERT(rr_costs.size() == rr_graph.num_nodes());
 
     float min_cost = std::numeric_limits<float>::infinity();
     float max_cost = -min_cost;
-    for (const RRNodeId& rr_id : device_ctx.rr_graph.nodes()) {
+    for (const RRNodeId& rr_id : rr_graph.nodes()) {
         if (std::isnan(rr_costs[(size_t)rr_id])) continue;
 
         min_cost = std::min(min_cost, rr_costs[(size_t)rr_id]);
@@ -4214,7 +4214,7 @@ static void draw_rr_costs(ezgl::renderer* g, const std::vector<float>& rr_costs,
 
     //Draw the nodes in ascending order of value, this ensures high valued nodes
     //are not overdrawn by lower value ones (e.g-> when zoomed-out far)
-    std::vector<int> nodes(device_ctx.rr_nodes.size());
+    std::vector<int> nodes(rr_graph.num_nodes());
     std::iota(nodes.begin(), nodes.end(), 0);
     auto cmp_ascending_cost = [&](int lhs_node, int rhs_node) {
         if (lowest_cost_first) {

vpr/src/draw/search_bar.cpp

@@ -68,7 +68,7 @@ void search_and_highlight(GtkWidget* /*widget*/, ezgl::application* app) {
         ss >> rr_node_id;
 
         // valid rr node id check
-        if (rr_node_id < 0 || rr_node_id >= int(device_ctx.rr_nodes.size())) {
+        if (rr_node_id < 0 || rr_node_id >= int(device_ctx.rr_graph.num_nodes())) {
             warning_dialog_box("Invalid RR Node ID");
             app->refresh_drawing();
             return;

vpr/src/pack/post_routing_pb_pin_fixup.cpp

@@ -159,7 +159,7 @@ static void update_cluster_pin_with_post_routing_results(const DeviceContext& de
             if (!rr_node) {
                 continue;
             }
-            VTR_ASSERT((size_t)rr_node < device_ctx.rr_nodes.size());
+            VTR_ASSERT((size_t)rr_node < device_ctx.rr_graph.num_nodes());
 
             /* If the node has been visited on the other side, we just skip it */
             if (visited_rr_nodes.end() != std::find(visited_rr_nodes.begin(), visited_rr_nodes.end(), RRNodeId(rr_node))) {

vpr/src/power/power.cpp

@@ -1192,7 +1192,7 @@ void power_routing_init(const t_det_routing_arch* routing_arch) {
     }
 
     /* Initialize RR Graph Structures */
-    rr_node_power = (t_rr_node_power*)vtr::calloc(device_ctx.rr_nodes.size(),
+    rr_node_power = (t_rr_node_power*)vtr::calloc(rr_graph.num_nodes(),
                                                   sizeof(t_rr_node_power));
     for (const RRNodeId& rr_id : device_ctx.rr_graph.nodes()) {
         rr_node_power[(size_t)rr_id].driver_switch_type = OPEN;

vpr/src/route/annotate_routing.cpp

@@ -28,7 +28,7 @@ vtr::vector<RRNodeId, ClusterNetId> annotate_rr_node_nets(const DeviceContext& d
     const auto& rr_graph = device_ctx.rr_graph;
 
     vtr::vector<RRNodeId, ClusterNetId> rr_node_nets;
-    rr_node_nets.resize(device_ctx.rr_nodes.size(), ClusterNetId::INVALID());
+    rr_node_nets.resize(rr_graph.num_nodes(), ClusterNetId::INVALID());
 
     for (auto net_id : clustering_ctx.clb_nlist.nets()) {
         if (clustering_ctx.clb_nlist.net_is_ignored(net_id)) {
@@ -52,7 +52,7 @@ vtr::vector<RRNodeId, ClusterNetId> annotate_rr_node_nets(const DeviceContext& d
                  * whose capacity is 1 
                  */
                 if ((rr_node_nets[rr_node])
-                    && (1 == device_ctx.rr_nodes.node_capacity(rr_node))
+                    && (1 == rr_graph.node_capacity(rr_node))
                     && (net_id != rr_node_nets[rr_node])) {
                     VPR_FATAL_ERROR(VPR_ERROR_ANALYSIS,
                                     "Detect two nets '%s' and '%s' that are mapped to the same rr_node '%ld'!\n%s\n",

vpr/src/route/check_route.cpp

@@ -74,8 +74,8 @@ void check_route(enum e_route_type route_type, e_check_route_option check_route_
 
     check_locally_used_clb_opins(route_ctx.clb_opins_used_locally, route_type);
 
-    auto connected_to_route = std::make_unique<bool[]>(device_ctx.rr_nodes.size());
-    std::fill_n(connected_to_route.get(), device_ctx.rr_nodes.size(), false);
+    auto connected_to_route = std::make_unique<bool[]>(rr_graph.num_nodes());
+    std::fill_n(connected_to_route.get(), rr_graph.num_nodes(), false);
 
     max_pins = 0;
     for (auto net_id : cluster_ctx.clb_nlist.nets())
@@ -536,7 +536,7 @@ void recompute_occupancy_from_scratch() {
             /* Will always be 0 for pads or SINK classes. */
             for (ipin = 0; ipin < num_local_opins; ipin++) {
                 inode = route_ctx.clb_opins_used_locally[blk_id][iclass][ipin];
-                VTR_ASSERT(inode >= 0 && inode < (ssize_t)device_ctx.rr_nodes.size());
+                VTR_ASSERT(inode >= 0 && inode < (ssize_t)device_ctx.rr_graph.num_nodes());
                 route_ctx.rr_node_route_inf[inode].set_occ(route_ctx.rr_node_route_inf[inode].occ() + 1);
             }
         }
@@ -592,9 +592,9 @@ static void check_node_and_range(int inode, enum e_route_type route_type) {
 
     auto& device_ctx = g_vpr_ctx.device();
 
-    if (inode < 0 || inode >= (int)device_ctx.rr_nodes.size()) {
+    if (inode < 0 || inode >= (int)device_ctx.rr_graph.num_nodes()) {
         VPR_FATAL_ERROR(VPR_ERROR_ROUTE,
-                        "in check_node_and_range: rr_node #%d is out of legal, range (0 to %d).\n", inode, device_ctx.rr_nodes.size() - 1);
+                        "in check_node_and_range: rr_node #%d is out of legal, range (0 to %d).\n", inode, device_ctx.rr_graph.num_nodes() - 1);
     }
     check_rr_node(inode, route_type, device_ctx);
 }