WIP ARM Neon/AVX2 SIMD implementation.

.gitignore

@@ -16,6 +16,7 @@ Gemfile.lock
 .DS_Store
 */**/Makefile
 */**/*.o
+*/**/extconf.h
 */**/*.class
 */**/*.jar
 .byebug_history

Rakefile

@@ -86,7 +86,7 @@ end
 
 file EXT_GENERATOR_DL => EXT_GENERATOR_SRC do
   cd EXT_GENERATOR_DIR do
-    ruby 'extconf.rb'
+    ruby "extconf.rb #{ENV['JSON_GENERATOR_CONFIGURE_OPTS']}"
     sh MAKE
   end
   cp "#{EXT_GENERATOR_DIR}/generator.#{CONFIG['DLEXT']}", EXT_ROOT_DIR

benchmark/encoder-simple.rb

@@ -0,0 +1,58 @@
+require "benchmark/ips"
+require "json"
+require "date"
+require "oj"
+
+Oj.default_options = Oj.default_options.merge(mode: :compat)
+
+if ENV["ONLY"]
+  RUN = ENV["ONLY"].split(/[,: ]/).map{|x| [x.to_sym, true] }.to_h
+  RUN.default = false
+elsif ENV["EXCEPT"]
+  RUN = ENV["EXCEPT"].split(/[,: ]/).map{|x| [x.to_sym, false] }.to_h
+  RUN.default = true
+else
+  RUN = Hash.new(true)
+end
+
+def implementations(ruby_obj)
+  state = JSON::State.new(JSON.dump_default_options)
+  {
+    json: ["json", proc { JSON.generate(ruby_obj) }],
+    oj: ["oj", proc { Oj.dump(ruby_obj) }],
+  }
+end
+
+def benchmark_encoding(benchmark_name, ruby_obj, check_expected: true, except: [])
+  json_output = JSON.dump(ruby_obj)
+  puts "== Encoding #{benchmark_name} (#{json_output.bytesize} bytes)"
+
+  impls = implementations(ruby_obj).select { |name| RUN[name] }
+  except.each { |i| impls.delete(i) }
+
+  Benchmark.ips do |x|
+    expected = ::JSON.dump(ruby_obj) if check_expected
+    impls.values.each do |name, block|
+      begin
+        result = block.call
+        if check_expected && expected != result
+          puts "#{name} does not match expected output. Skipping"
+          puts "Expected:" + '-' * 40
+          puts expected
+          puts "Actual:" + '-' * 40
+          puts result
+          puts '-' * 40
+          next
+        end
+      rescue => error
+        puts "#{name} unsupported (#{error})"
+        next
+      end
+      x.report(name, &block)
+    end
+    x.compare!(order: :baseline)
+  end
+  puts
+end
+
+benchmark_encoding "long string", (["this is a test of the emergency broadcast system."*5]*500)

ext/json/ext/generator/extconf.rb

@@ -6,5 +6,55 @@
 else
   append_cflags("-std=c99")
   $defs << "-DJSON_GENERATOR"
+
+  if enable_config('use-simd', default=true)
+    if RbConfig::CONFIG['host_cpu'] =~ /^(arm.*|aarch64.*)/
+      $defs.push("-DENABLE_SIMD")
+
+      # Try to compile a small program using NEON instructions
+      if have_header('arm_neon.h')
+        have_type('uint8x16_t', headers=['arm_neon.h']) && try_compile(<<~'SRC')
+          #include <arm_neon.h>
+          int main() {
+              uint8x16_t test = vdupq_n_u8(32);
+              return 0;
+          }
+        SRC
+
+        have_type('uint8x8_t', headers=['arm_neon.h']) && try_compile(<<~'SRC')
+            #include <arm_neon.h>
+            int main() {
+                uint8x8_t test = vdup_n_u8(32);
+                return 0;
+            }
+        SRC
+        end
+      elsif have_header('x86intrin.h')
+
+        if have_type('__m256i', headers=['x86intrin.h']) && try_compile(<<~'SRC', opt='-mavx2')
+          #include <x86intrin.h>
+          int main() {
+              __m256i test = _mm256_set1_epi8(32);
+              return 0;
+          }
+          SRC
+          $defs.push("-DENABLE_SIMD")
+        end
+
+        if have_type('__m128i', headers=['x86intrin.h']) && try_compile(<<~'SRC', opt='-mavx2')
+          #include <x86intrin.h>
+          int main() {
+              __m128i test = _mm_set1_epi8(32);
+              return 0;
+          }
+          SRC
+            $defs.push("-DENABLE_SIMD") unless $defs.include?('-DENABLE_SIMD')
+        end
+      end
+
+      have_header('cpuid.h')
+  end
+
+  create_header
   create_makefile 'json/ext/generator'
 end