[PATCH] D60792: Time profiler: optimize json output time

[Anton Afanasyev via Phabricator via llvm-commits]

anton-afanasyev created this revision.
anton-afanasyev added a reviewer: lebedev.ri.
Herald added subscribers: llvm-commits, hiraditya.
Herald added a project: LLVM.

Use llvm::json::Array.reserve() to optimize json output time. Here is motivation:
https://reviews.llvm.org/D60609#1468941. In short: for the json array
with ~32K entries, pushing back each entry takes ~4% of whole time compared
to the method of preliminary memory reservation: (3995-3845)/3995 = 3.75%.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D60792

Files:
  llvm/lib/Support/TimeProfiler.cpp


Index: llvm/lib/Support/TimeProfiler.cpp
===================================================================
--- llvm/lib/Support/TimeProfiler.cpp
+++ llvm/lib/Support/TimeProfiler.cpp
@@ -89,6 +89,7 @@
            "All profiler sections should be ended when calling Write");
 
     json::Array Events;
+    Events.reserve(Entries.size() + CountAndTotalPerName.size() + 1);
 
     // Emit all events for the main flame graph.
     for (const auto &E : Entries) {
@@ -149,6 +150,7 @@
         {"args", json::Object{{"name", "clang"}}},
     });
 
+    assert((Entries.size() + CountAndTotalPerName.size() + 1 == Events.size()) && "Events reserved size should be enough!");
     OS << formatv("{0:2}", json::Value(json::Object(
                                {{"traceEvents", std::move(Events)}})));
   }


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D60792.195439.patch
Type: text/x-patch
Size: 804 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20190416/94696107/attachment.bin>