add multithreaded bench

bench: 2638316

README.md

@@ -9,6 +9,8 @@ This code is alpha status right now, and is intended for experiment/testing only
 * The search is wider than upstream.
 * Single file build for quicker compiles.
 * Most reductions/extensions are removed.
+* Increased minimum hash size to 32MB.
+* Uses multi threaded bench for PGO builds
 
 ## Future plans
 

generate

@@ -1,5 +1,5 @@
 #!/bin/bash
-#below line commented out because it only needs to run once and has already been run
+#below line technically only needs to run once
 guard2once -r .
 rm -rf build
 cp -r src build

src/Makefile

@@ -49,7 +49,7 @@ PREFIX = /usr/local
 BINDIR = $(PREFIX)/bin
 
 ### Built-in benchmark for pgo-builds
-PGOBENCH = $(WINE_PATH) ./$(EXE) bench
+PGOBENCH = $(WINE_PATH) ./$(EXE) bench 64 3 14
 
 ### Source and object files
 SRCS = main_gen.cpp
@@ -1037,7 +1037,7 @@ gcc-profile-make:
 
 gcc-profile-use:
 	$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
-	EXTRACXXFLAGS='-fprofile-use=profdir -fno-peel-loops -fno-tracer' \
+	EXTRACXXFLAGS='-fprofile-use=profdir -fno-peel-loops -fno-tracer -fprofile-correction' \
 	EXTRACXXFLAGS+=$(EXTRAPROFILEFLAGS) \
 	EXTRALDFLAGS='-lgcov' \
 	all

src/main_gen.cpp

@@ -8725,17 +8725,6 @@ Value Search::Worker::search(
                 ? ss->staticEval > (ss - 2)->staticEval
                 : (ss - 4)->staticEval != VALUE_NONE && ss->staticEval > (ss - 4)->staticEval;
 
-    // Step 7. Razoring (~1 Elo)
-    // If eval is really low check with qsearch if it can exceed alpha, if it can't,
-    // return a fail low.
-    // Adjust razor margin according to cutoffCnt. (~1 Elo)
-    if (eval < alpha - 438 - (332 - 154 * ((ss + 1)->cutoffCnt > 3)) * depth * depth)
-    {
-        value = qsearch<NonPV>(pos, ss, alpha - 1, alpha);
-        if (value < alpha)
-            return value;
-    }
-
     // Step 8. Futility pruning: child node (~40 Elo)
     // The depth condition is important for mate finding.
     if (!ss->ttPv && depth < 11