diff --git a/desmume/src/GPU.cpp b/desmume/src/GPU.cpp
index 8f5070a69..fbdc9c885 100644
--- a/desmume/src/GPU.cpp
+++ b/desmume/src/GPU.cpp
@@ -1673,7 +1673,7 @@ FORCEINLINE void GPUEngineBase::_RenderPixel_CheckWindows(const size_t srcX, boo
 #ifdef ENABLE_SSE2
 
 template <GPULayerID LAYERID, bool ISCUSTOMRENDERINGNEEDED>
-void GPUEngineBase::_RenderPixel_CheckWindows16_SSE2(const size_t dstX, __m128i &didPassWindowTest, __m128i &enableColorEffect) const
+FORCEINLINE void GPUEngineBase::_RenderPixel_CheckWindows16_SSE2(const size_t dstX, __m128i &didPassWindowTest, __m128i &enableColorEffect) const
 {
 	// If no windows are enabled, then we don't need to perform any window tests.
 	// In this case, the pixel always passes and the color effect is always processed.
diff --git a/desmume/src/GPU.h b/desmume/src/GPU.h
index cac243b03..b1993ae37 100644
--- a/desmume/src/GPU.h
+++ b/desmume/src/GPU.h
@@ -1359,7 +1359,7 @@ protected:
 	template<NDSColorFormat COLORFORMATA, NDSColorFormat COLORFORMATB> FORCEINLINE __m128i _ColorEffectBlend3D(const __m128i &colA_Lo, const __m128i &colA_Hi, const __m128i &colB);
 	template<NDSColorFormat COLORFORMAT> FORCEINLINE __m128i _ColorEffectIncreaseBrightness(const __m128i &col, const __m128i &blendEVY);
 	template<NDSColorFormat COLORFORMAT> FORCEINLINE __m128i _ColorEffectDecreaseBrightness(const __m128i &col, const __m128i &blendEVY);
-	template<GPULayerID LAYERID, bool ISCUSTOMRENDERINGNEEDED> void _RenderPixel_CheckWindows16_SSE2(const size_t dstX, __m128i &didPassWindowTest, __m128i &enableColorEffect) const;
+	template<GPULayerID LAYERID, bool ISCUSTOMRENDERINGNEEDED> FORCEINLINE void _RenderPixel_CheckWindows16_SSE2(const size_t dstX, __m128i &didPassWindowTest, __m128i &enableColorEffect) const;
 	template<GPULayerID LAYERID, bool ISDEBUGRENDER, bool NOWINDOWSENABLEDHINT, bool COLOREFFECTDISABLEDHINT, bool ISCUSTOMRENDERINGNEEDED> FORCEINLINE void _RenderPixel16_SSE2(const size_t dstX, const __m128i &srcColorHi_vec128, const __m128i &srcColorLo_vec128, const u8 *__restrict srcAlpha, void *__restrict dstColorLine, u8 *__restrict dstLayerIDLine, __m128i &passMask8);
 	template<NDSColorFormat OUTPUTFORMAT> FORCEINLINE void _RenderPixel3D_SSE2(const NDSColorFormat srcFormat, const __m128i &src3, const __m128i &src2, const __m128i &src1, const __m128i &src0, __m128i &dst3, __m128i &dst2, __m128i &dst1, __m128i &dst0, __m128i &dstLayerID, __m128i &passMask8, __m128i &enableColorEffectMask);
 #endif