diff --git a/SheepShaver/src/kpx_cpu/src/cpu/ppc/ppc-translate.cpp b/SheepShaver/src/kpx_cpu/src/cpu/ppc/ppc-translate.cpp
index fc45d713..3f885720 100644
--- a/SheepShaver/src/kpx_cpu/src/cpu/ppc/ppc-translate.cpp
+++ b/SheepShaver/src/kpx_cpu/src/cpu/ppc/ppc-translate.cpp
@@ -37,6 +37,12 @@
 #include "mon_disass.h"
 #endif
 
+// Disable dynamic translation for floating-point load/store of doubles?
+// XXX: could this due to some rounding while processing these?
+#ifdef _WIN32
+#define DISABLE_FP_DOUBLE_LOAD_STORE 1
+#endif
+
 // Define to enable const branches optimization
 #define FOLLOW_CONST_JUMPS 1
 
@@ -1103,6 +1109,7 @@ powerpc_cpu::compile_block(uint32 entry_point)
 			dg.gen_store_T0_crf(crfD_field::extract(opcode));
 			break;
 		}
+#ifndef DISABLE_FP_DOUBLE_LOAD_STORE
 		case PPC_I(LFD):		// Load Floating-Point Double
 			op.mem.size = 8;
 			op.mem.do_update = 0;
@@ -1123,6 +1130,7 @@ powerpc_cpu::compile_block(uint32 entry_point)
 			op.mem.do_update = 0;
 			op.mem.do_indexed = 1;
 			goto do_fp_load;
+#endif
 		case PPC_I(LFS):		// Load Floating-Point Single
 			op.mem.size = 4;
 			op.mem.do_update = 0;
@@ -1183,6 +1191,7 @@ powerpc_cpu::compile_block(uint32 entry_point)
 			}
 			break;
 		}
+#ifndef DISABLE_FP_DOUBLE_LOAD_STORE
 		case PPC_I(STFD):		// Store Floating-Point Double
 			op.mem.size = 8;
 			op.mem.do_update = 0;
@@ -1203,6 +1212,7 @@ powerpc_cpu::compile_block(uint32 entry_point)
 			op.mem.do_update = 0;
 			op.mem.do_indexed = 1;
 			goto do_fp_store;
+#endif
 		case PPC_I(STFS):		// Store Floating-Point Single
 			op.mem.size = 4;
 			op.mem.do_update = 0;