Adjust GTP configuration to improve performance

- Set the number of search threads to 16.
- Set the number of max batch size to 8.
- Use two neural network server threads for GPU and Neural Engine.

ios/KataGo iOS/Resources/default_gtp.cfg

@@ -217,7 +217,7 @@ maxTimePondering = 60  # Maximum time to ponder, in seconds. Comment out to make
 lagBuffer = 1.0
 
 # Number of threads to use in search
-numSearchThreads = 2
+numSearchThreads = 16
 
 # Play a little faster if the opponent is passing, for friendliness
 searchFactorAfterOnePass = 0.50
@@ -232,7 +232,7 @@ searchFactorWhenWinningThreshold = 0.95
 # The default value here is roughly equal to numSearchThreads, but you can specify it manually
 # if you are running out of memory, or if you are using multiple GPUs that expect to split
 # up the work.
-# nnMaxBatchSize = <integer>
+nnMaxBatchSize = 8
 
 # Cache up to (2 ** this) many neural net evaluations in case of transpositions in the tree.
 # Uncomment and edit to change if you want to adjust a major component of KataGo's RAM usage.
@@ -251,7 +251,7 @@ searchFactorWhenWinningThreshold = 0.95
 # Metal backend runs the default GPU 0.
 # CoreML backend runs at another two threads.
 # So, if you want to use Metal + CoreML, you should set numNNServerThreadsPerModel to 3.
-numNNServerThreadsPerModel = 1
+numNNServerThreadsPerModel = 2
 
 
 # TENSORRT GPU settings--------------------------------------
@@ -347,8 +347,8 @@ coremlDeviceToUse = 100 # Neural Engine
 
 # IF USING TWO MODEL: Uncomment these two lines
 # (AND also set numNNServerThreadsPerModel = 2 above)
-# coremlDeviceToUseThread0 = 0 # GPU
-# coremlDeviceToUseThread1 = 100 # Neural Engine
+coremlDeviceToUseThread0 = 0 # GPU
+coremlDeviceToUseThread1 = 100 # Neural Engine
 
 # IF USING THREE MODEL: Uncomment these three lines
 # (AND also set numNNServerThreadsPerModel = 3 above)