Limit the number of thinlto jobs to 8; also support limits in lld.

build/config/compiler/BUILD.gn

 # Copyright (c) 2013 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 import("//build/config/android/config.gni")
 import("//build/config/chrome_build.gni")
 import("//build/config/compiler/compiler.gni")
 import("//build/config/nacl/config.gni")
 import("//build/toolchain/cc_wrapper.gni")
 import("//build/toolchain/toolchain.gni")
@@ skipping 423 matching lines @@
   # optimizations/transformations that require whole-program visibility at link
   # time, so they need to be applied to all translation units, and we may end up
   # with miscompiles if only part of the program is compiled with LTO flags. For
   # that reason, we cannot allow targets to enable or disable these flags, for
   # example by disabling the optimize configuration.
   # TODO(pcc): Make this conditional on is_official_build rather than on gn
   # flags for specific features.
   if (!is_debug && (allow_posix_link_time_opt || is_cfi) && !is_nacl) {
     if (use_thin_lto) {
       cflags += [ "-flto=thin" ]
-      ldflags += [
-        "-flto=thin",
+      ldflags += [ "-flto=thin" ]
 
-        # Limit the parallelism to the sweet spot on most of the machines.
-        # As of now, ThinLTO does not scale beyond 16 cores anyway.
-        "-Wl,-plugin-opt,jobs=16",
-      ]
+      # Limit the parallelism to avoid too agressive competition between
+      # linker jobs. This is still suboptimal to a potential dynamic
+      # resource allocation scheme, but should be good enough.

[Nico, 2016/12/06 21:08:48]

Shouldn't the OS scheduler be ok with linker threads being oversubscribed, as
long as there's not a lot of memory use per thread? (Is there?)

+      if (use_lld) {
+        ldflags += [ "-Wl,--thinlto-jobs=8" ]
+      } else {
+        ldflags += [ "-Wl,-plugin-opt,jobs=8" ]
+      }
     } else {
       # Note: ThinLTO does not currently have this feature implemented
       # For Full LTO, it provides a measurable runtime speedup of Chrome.
       cflags += [
         "-flto",
         "-fwhole-program-vtables",
       ]
       ldflags += [
         "-flto",
         "-fwhole-program-vtables",
@@ skipping 1174 matching lines @@
 
 if (is_ios || is_mac) {
   # On Mac and iOS, this enables support for ARC (automatic ref-counting).
   # See http://clang.llvm.org/docs/AutomaticReferenceCounting.html.
   config("enable_arc") {
     common_flags = [ "-fobjc-arc" ]
     cflags_objc = common_flags
     cflags_objcc = common_flags
   }
 }