diff --git a/Algo256/bmw.cu b/Algo256/bmw.cu
index 3b5ed21..fd12f02 100644
--- a/Algo256/bmw.cu
+++ b/Algo256/bmw.cu
@@ -33,10 +33,6 @@ extern "C" void bmw_hash(void *state, const void *input)
 
 static bool init[MAX_GPUS] = { 0 };
 
-static __inline uint32_t swab32_if(uint32_t val, bool iftrue) {
-	return iftrue ? swab32(val) : val;
-}
-
 extern "C" int scanhash_bmw(int thr_id, struct work* work, uint32_t max_nonce, unsigned long *hashes_done)
 {
 	uint32_t _ALIGN(64) endiandata[20];
diff --git a/README.txt b/README.txt
index b7bcc04..5d45c46 100644
--- a/README.txt
+++ b/README.txt
@@ -269,6 +269,13 @@ features.
 
 >>> RELEASE HISTORY <<<
 
+  Jan. 2017 (WIP)
+                  Handle cryptonight, wildkeccak and cryptonight-lite
+                  Rework the common skein512 (used in most algos except skein ;)
+                  Store the share diff of second nonce(s) in most algos
+                  Hardware monitoring thread to get more accurate power readings
+                  Small changes for the quiet mode
+
   Dec. 21th 2016  v1.8.4
                   Improve streebog based algos, veltor and sib (from alexis work)
                   Blake2s greetly improved (3x), thanks to alexis too...
diff --git a/crypto/wildkeccak.cu b/crypto/wildkeccak.cu
index 3598d5f..a237715 100644
--- a/crypto/wildkeccak.cu
+++ b/crypto/wildkeccak.cu
@@ -275,7 +275,7 @@ extern "C" int scanhash_wildkeccak(int thr_id, struct work* work, uint32_t max_n
 		init[thr_id] = true;
 	}
 
-	throughput = cuda_default_throughput(thr_id, WK_CUDABlocks*WK_CUDAThreads);
+	throughput = WK_CUDABlocks * WK_CUDAThreads;
 
 	cudaMemcpy(d_input[thr_id], pdata, 88, cudaMemcpyHostToDevice);
 //	cudaMemset(d_retnonce[thr_id], 0xFF, 2*sizeof(uint32_t));
@@ -291,7 +291,7 @@ extern "C" int scanhash_wildkeccak(int thr_id, struct work* work, uint32_t max_n
 		uint32_t h_retnonce[2] = { UINT32_MAX, UINT32_MAX };
 		uint2 target = make_uint2(ptarget[6], ptarget[7]);
 
-		wildkeccak_kernel(thr_id, throughput, nonce, target, h_retnonce);
+		wildkeccak_kernel(thr_id, throughput, (uint32_t) nonce, target, h_retnonce);
 		/*
 		wk <<<block, thread, 0, kernel_stream[thr_id]>>> (d_retnonce[thr_id], d_input[thr_id], d_scratchpad[thr_id],
 			(uint32_t)(scratchpad_size >> 2), nonce, ptarget[7]);