diff --git a/src/core/arm/nce/arm_nce.cpp b/src/core/arm/nce/arm_nce.cpp
index 1311e66a9..123b3da7e 100644
--- a/src/core/arm/nce/arm_nce.cpp
+++ b/src/core/arm/nce/arm_nce.cpp
@@ -39,7 +39,7 @@ fpsimd_context* GetFloatingPointState(mcontext_t& host_ctx) {
 }
 
 using namespace Common::Literals;
-constexpr u32 StackSize = 32_KiB;
+constexpr u32 StackSize = 128_KiB;
 
 } // namespace
 
diff --git a/src/core/arm/nce/interpreter_visitor.cpp b/src/core/arm/nce/interpreter_visitor.cpp
index 8e81c66a5..ec373fdf2 100644
--- a/src/core/arm/nce/interpreter_visitor.cpp
+++ b/src/core/arm/nce/interpreter_visitor.cpp
@@ -249,6 +249,7 @@ bool InterpreterVisitor::LDR_lit_fpsimd(Imm<2> opc, Imm<19> imm19, Vec Vt) {
         return false;
     }
 
+    // Size in bytes
     const u64 size = 4 << opc.ZeroExtend();
     const u64 offset = imm19.SignExtend<u64>() << 2;
     const u64 address = this->GetPc() + offset;
@@ -530,7 +531,7 @@ bool InterpreterVisitor::SIMDImmediate(bool wback, bool postindex, size_t scale,
     }
     case MemOp::Load: {
         u128 data{};
-        m_memory.ReadBlock(address, &data, datasize);
+        m_memory.ReadBlock(address, &data, datasize / 8);
         this->SetVec(Vt, data);
         break;
     }