X compilebench: Don't spend time measuring compile and link separately

We know from past benchmarks that most of the time is in compile.

t/compilebench/doit.sh

@@ -69,15 +69,17 @@ for CC in gcc tcc; do
 	echo "CC=$CC"
 	rm -f pystone_pypyx.{o,so}
 
-	for i in `seq $N`; do
-		export TIMEFORMAT="Benchmark${CC}compile 1 %R s/op"
-		time compile pystone_pypyx
-	done
-
-	for i in `seq $N`; do
-		export TIMEFORMAT="Benchmark${CC}link 1 %R s/op"
-		time link pystone_pypyx
-	done
+	# don't spend time measuring compile and link separatelly: we know from
+	# past benchmarks that most of the time is in compile.
+	#for i in `seq $N`; do
+	#	export TIMEFORMAT="Benchmark${CC}compile 1 %R s/op"
+	#	time compile pystone_pypyx
+	#done
+	#
+	#for i in `seq $N`; do
+	#	export TIMEFORMAT="Benchmark${CC}link 1 %R s/op"
+	#	time link pystone_pypyx
+	#done
 
 	for i in `seq $N`; do
 		export TIMEFORMAT="Benchmark${CC}compile+link 1 %R s/op"