diff --git a/bolt/include/bolt/Profile/DataAggregator.h b/bolt/include/bolt/Profile/DataAggregator.h
--- a/bolt/include/bolt/Profile/DataAggregator.h
+++ b/bolt/include/bolt/Profile/DataAggregator.h
@@ -27,6 +27,9 @@
 class BinaryContext;
 class BoltAddressTranslation;
 
+// The format to use with -o in aggregation mode (perf2bolt)
+enum ProfileOutputKind { PO_Fdata, PO_YAML };
+
 /// DataAggregator inherits all parsing logic from DataReader as well as
 /// its data structures used to represent aggregated profile data in memory.
 ///
diff --git a/bolt/lib/Profile/DataAggregator.cpp b/bolt/lib/Profile/DataAggregator.cpp
--- a/bolt/lib/Profile/DataAggregator.cpp
+++ b/bolt/lib/Profile/DataAggregator.cpp
@@ -78,6 +78,8 @@
   cl::Hidden,
   cl::cat(AggregatorCategory));
 
+extern cl::opt<ProfileOutputKind> ProfileOutput;
+
 cl::opt<bool> ReadPreAggregated(
     "pa", cl::desc("skip perf and read data from a pre-aggregated file format"),
     cl::cat(AggregatorCategory));
@@ -610,7 +612,8 @@
     convertBranchData(Function);
   }
 
-  if (opts::AggregateOnly) {
+  if (opts::AggregateOnly &&
+      opts::ProfileOutput == ProfileOutputKind::PO_Fdata) {
     if (std::error_code EC = writeAggregatedFile(opts::OutputFilename))
       report_error("cannot create output data file", EC);
   }
diff --git a/bolt/lib/Rewrite/RewriteInstance.cpp b/bolt/lib/Rewrite/RewriteInstance.cpp
--- a/bolt/lib/Rewrite/RewriteInstance.cpp
+++ b/bolt/lib/Rewrite/RewriteInstance.cpp
@@ -208,6 +208,15 @@
                clEnumValN(PPP_All, "all", "enable all debugging printout")),
     cl::ZeroOrMore, cl::Hidden, cl::cat(BoltCategory));
 
+cl::opt<ProfileOutputKind> ProfileOutput(
+    "profile-output",
+    cl::desc(
+        "format to dump profile output in aggregation mode, default is fdata"),
+    cl::init(PO_Fdata),
+    cl::values(clEnumValN(PO_Fdata, "fdata", "offset-based plaintext format"),
+               clEnumValN(PO_YAML, "yaml", "dense YAML reprensentation")),
+    cl::ZeroOrMore, cl::Hidden, cl::cat(BoltCategory));
+
 static cl::opt<cl::boolOrDefault> RelocationMode(
     "relocs", cl::desc("use relocations in the binary (default=autodetect)"),
     cl::cat(BoltCategory));
@@ -2796,6 +2805,11 @@
     YAMLProfileWriter PW(opts::SaveProfile);
     PW.writeProfile(*this);
   }
+  if (opts::AggregateOnly &&
+      opts::ProfileOutput == ProfileOutputKind::PO_YAML) {
+    YAMLProfileWriter PW(opts::OutputFilename);
+    PW.writeProfile(*this);
+  }
 
   // Release memory used by profile reader.
   ProfileReader.reset();