cmd/compile,cmd/preprofile: move logic to shared common package

The processing performed in cmd/preprofile is a simple version of the
same initial processing performed by cmd/compile/internal/pgo. Refactor
this processing into the new IR-independent cmd/internal/pgo package.

Now cmd/preprofile and cmd/compile run the same code for initial
processing of a pprof profile, guaranteeing that they always stay in
sync.

Since it is now trivial, this CL makes one change to the serialization
format: the entries are ordered by weight. This allows us to avoid
sorting ByWeight on deserialization.

Impact on PGO parsing when compiling cmd/compile with PGO:

* Without preprocessing: PGO parsing ~13.7% of CPU time
* With preprocessing (unsorted): ~2.9% of CPU time (sorting ~1.7%)
* With preprocessing (sorted): ~1.3% of CPU time

The remaining 1.3% of CPU time approximately breaks down as:

* ~0.5% parsing the preprocessed profile
* ~0.7% building weighted IR call graph
  * ~0.5% walking function IR to find direct calls
  * ~0.2% performing lookups for indirect calls targets

For #58102.

Change-Id: Iaba425ea30b063ca195fb2f7b29342961c8a64c2
Reviewed-on: https://go-review.googlesource.com/c/go/+/569337
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Pratt <mpratt@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>

1 files changed, 13 insertions, 12 deletions

diff --git a/src/cmd/compile/internal/devirtualize/pgo_test.go b/src/cmd/compile/internal/devirtualize/pgo_test.go
index 84c96df122..6ba8e9f907 100644
--- a/src/cmd/compile/internal/devirtualize/pgo_test.go
+++ b/src/cmd/compile/internal/devirtualize/pgo_test.go
@@ -7,10 +7,11 @@ package devirtualize
 import (
 	"cmd/compile/internal/base"
 	"cmd/compile/internal/ir"
-	"cmd/compile/internal/pgo"
+	pgoir "cmd/compile/internal/pgo"
 	"cmd/compile/internal/typecheck"
 	"cmd/compile/internal/types"
 	"cmd/internal/obj"
+	"cmd/internal/pgo"
 	"cmd/internal/src"
 	"testing"
 )
@@ -32,32 +33,32 @@ func makePos(b *src.PosBase, line, col uint) src.XPos {
 }
 
 type profileBuilder struct {
-	p *pgo.Profile
+	p *pgoir.Profile
 }
 
 func newProfileBuilder() *profileBuilder {
-	// findHotConcreteCallee only uses pgo.Profile.WeightedCG, so we're
+	// findHotConcreteCallee only uses pgoir.Profile.WeightedCG, so we're
 	// going to take a shortcut and only construct that.
 	return &profileBuilder{
-		p: &pgo.Profile{
-			WeightedCG: &pgo.IRGraph{
-				IRNodes: make(map[string]*pgo.IRNode),
+		p: &pgoir.Profile{
+			WeightedCG: &pgoir.IRGraph{
+				IRNodes: make(map[string]*pgoir.IRNode),
 			},
 		},
 	}
 }
 
 // Profile returns the constructed profile.
-func (p *profileBuilder) Profile() *pgo.Profile {
+func (p *profileBuilder) Profile() *pgoir.Profile {
 	return p.p
 }
 
 // NewNode creates a new IRNode and adds it to the profile.
 //
 // fn may be nil, in which case the node will set LinkerSymbolName.
-func (p *profileBuilder) NewNode(name string, fn *ir.Func) *pgo.IRNode {
-	n := &pgo.IRNode{
-		OutEdges: make(map[pgo.NamedCallEdge]*pgo.IREdge),
+func (p *profileBuilder) NewNode(name string, fn *ir.Func) *pgoir.IRNode {
+	n := &pgoir.IRNode{
+		OutEdges: make(map[pgo.NamedCallEdge]*pgoir.IREdge),
 	}
 	if fn != nil {
 		n.AST = fn
@@ -69,13 +70,13 @@ func (p *profileBuilder) NewNode(name string, fn *ir.Func) *pgo.IRNode {
 }
 
 // Add a new call edge from caller to callee.
-func addEdge(caller, callee *pgo.IRNode, offset int, weight int64) {
+func addEdge(caller, callee *pgoir.IRNode, offset int, weight int64) {
 	namedEdge := pgo.NamedCallEdge{
 		CallerName:     caller.Name(),
 		CalleeName:     callee.Name(),
 		CallSiteOffset: offset,
 	}
-	irEdge := &pgo.IREdge{
+	irEdge := &pgoir.IREdge{
 		Src:            caller,
 		Dst:            callee,
 		CallSiteOffset: offset,