simplify bfs, other code review comments

This commit is contained in:
Ryan Berger 2023-05-08 15:16:56 -06:00
parent 1c2d14836e
commit 4846c63a97
1 changed files with 50 additions and 56 deletions

View File

@ -4,22 +4,36 @@
package ssa
// reassociate balances trees of commutative computation
// to better group expressions to expose easy optimizations in
// cse, cancelling/counting/factoring expressions, etc.
func reassociate(f *Func) {
visited := f.newSparseSet(f.NumValues())
for _, b := range f.Postorder() {
for i := len(b.Values) - 1; i >= 0; i-- {
val := b.Values[i]
rebalance(val, visited)
}
}
}
// balanceExprTree repurposes all nodes and leaves into a well-balanced expression tree.
// It doesn't truly balance the tree in the sense of a BST, rather it
// It doesn't truly balance the tree in the sense of a BST, rather it
// prioritizes pairing up innermost (rightmost) expressions and their results and only
// pairing results of outermost (leftmost) expressions up with them when no other nice pairing exists
func balanceExprTree(v *Value, visited map[*Value]bool, nodes, leaves []*Value) {
// pairing results of outermost (leftmost) expressions up with them when no other nice pairing exists
func balanceExprTree(v *Value, visited *sparseSet, nodes, leaves []*Value) {
// reset all arguments of nodes to help rebalancing
for i, n := range nodes {
n.reset(n.Op)
// sometimes nodes in the tree are in different blocks
// so pull them in into a common block (v's block)
// to make sure nodes don't end up dominating their leaves
// to make sure nodes don't end up dominating their leaves TODO(ryan-berger), not necessary
if v.Block != n.Block {
copied := n.copyInto(v.Block)
n.Op = OpInvalid
visited[n] = true // "revisit" the copied node
visited.add(copied.ID) // "revisit" the copied node
nodes[i] = copied
}
}
@ -33,24 +47,24 @@ func balanceExprTree(v *Value, visited map[*Value]bool, nodes, leaves []*Value)
// rebuild expression trees from the bottom up, prioritizing
// right grouping.
// if the number of leaves is not even, skip the first leaf
// if the number of leaves is not even, skip the first leaf
// and add it to be paired up later
i := 0
subTrees := leaves
for len(subTrees) != 1 {
nextSubTrees := make([]*Value, 0, (len(subTrees)+1)/2)
start := len(subTrees)%2
start := len(subTrees) % 2
if start != 0 {
nextSubTrees = append(nextSubTrees, subTrees[0])
}
for j := start; j < len(subTrees)-1; j+=2 {
for j := start; j < len(subTrees)-1; j += 2 {
nodes[i].AddArg2(subTrees[j], subTrees[j+1])
nextSubTrees = append(nextSubTrees, nodes[i])
i++
}
subTrees = nextSubTrees
}
}
@ -96,78 +110,58 @@ func probablyMemcombine(op Op, leaves []*Value) bool {
// rebalance balances associative computation to better help CPU instruction pipelining (#49331)
// and groups constants together catch more constant folding opportunities.
//
// a + b + c + d compiles to to (a + (b + (c + d)) which is an unbalanced expression tree
// a + b + c + d compiles to to v1:(a + v2:(b + v3:(c + d)) which is an unbalanced expression tree
// Which is suboptimal since it requires the CPU to compute v3 before fetching it use its result in
// v2, and v2 before its use in v1
//
// This optimization rebalances this expression tree to look like (a + b) + (c + d) ,
// which removes such dependencies and frees up the CPU pipeline.
//
// The above optimization is a good starting point for other sorts of operations such as
// The above optimization is also a good starting point for other sorts of operations such as
// turning a + a + a => 3*a, cancelling pairs a + (-a), collecting up common factors TODO(ryan-berger)
func rebalance(v *Value, visited map[*Value]bool) {
func rebalance(v *Value, visited *sparseSet) {
// We cannot apply this optimization to non-commutative operations,
// values that have more than one use, or non-binary ops (would need more log math).
// Try and save time by not revisiting nodes
if visited[v] || !(v.Uses == 1 && opcodeTable[v.Op].commutative) || len(v.Args) > 2 {
if visited.contains(v.ID) || !opcodeTable[v.Op].commutative {
return
}
// The smallest possible rebalanceable expression has 3 nodes and 4 leaves,
// The smallest possible rebalanceable binary expression has 3 nodes and 4 leaves,
// so preallocate the lists to save time if it is not rebalanceable
leaves := make([]*Value, 0, 4)
nodes := make([]*Value, 0, 3)
// Do a bfs on v to keep a nice reverse topological order
haystack := []*Value{v}
for len(haystack) != 0 {
nextHaystack := make([]*Value, 0, len(v.Args)*len(haystack))
for _, needle := range haystack {
// if we are searching a value, it must be a node so add it to our node list
nodes = append(nodes, needle)
for i := 0; i < len(haystack); i++ {
needle := haystack[i]
// if we are searching a value, it must be a node so add it to our node list
nodes = append(nodes, needle)
// Only visit nodes. Leafs may be rebalancable for a different op type
visited[needle] = true
// Only visit nodes. Leafs may be rebalancable for a different op type
visited.add(v.ID)
for _, a := range needle.Args {
// If the ops aren't the same or have more than one use it must be a leaf.
if a.Op != v.Op || a.Uses != 1 {
leaves = append(leaves, a)
continue
}
// nodes in the tree now hold the invariants that:
// - they are of a common associative operation as the rest of the tree
// - they have only a single use (this invariant could be removed with further analysis TODO(ryan-berger)
nextHaystack = append(nextHaystack, a)
for _, a := range needle.Args {
// If the ops aren't the same or have more than one use it must be a leaf.
if a.Op != v.Op || a.Uses != 1 {
leaves = append(leaves, a)
continue
}
// nodes in the tree now hold the invariants that:
// - they are of a common associative operation as the rest of the tree
// - they have only a single use (this invariant could be removed with further analysis TODO(ryan-berger))
haystack = append(haystack, a)
}
haystack = nextHaystack
}
// we need at least 4 leaves for this expression to be rebalanceable,
minLeaves := len(v.Args) * len(v.Args)
// we need at least args^2 leaves for this expression to be rebalanceable,
// and we can't balance a potential load widening (see memcombine)
if len(leaves) < 4 || probablyMemcombine(v.Op, leaves) {
if len(leaves) < minLeaves || probablyMemcombine(v.Op, leaves) {
return
}
balanceExprTree(v, visited, nodes, leaves)
}
// reassociate balances trees of commutative computation
// to better group expressions to expose easy optimizations in
// cse, cancelling/counting/factoring expressions, etc.
func reassociate(f *Func) {
visited := make(map[*Value]bool)
for _, b := range f.Postorder() {
for i := len(b.Values) - 1; i >= 0; i-- {
val := b.Values[i]
rebalance(val, visited)
}
}
for k := range visited {
delete(visited, k)
}
}