diff --git a/include/mtl/utility.hpp b/include/mtl/utility.hpp
index e6aa5cd..6e5ac9c 100644
--- a/include/mtl/utility.hpp
+++ b/include/mtl/utility.hpp
@@ -17,10 +17,11 @@ void mtl_hybridmove(void* dst, const void* src, size_t num8);
 
 uint32_t mtl_udiv10(uint32_t x);
 uint32_t mtl_umod10(uint32_t x);
-uint32_t mtl_utostr(uint32_t x, char* buf, size_t buf_size, char** buf_out);
-uint32_t mtl_utostrx(uint32_t x, char* buf, size_t buf_size, char** buf_out);
-uint32_t mtl_itostr(uint32_t x, char* buf, size_t buf_size, char** buf_out);
-uint32_t mtl_itostrx(uint32_t x, char* buf, size_t buf_size, char** buf_out);
+
+uint32_t mtl_utostr(uint32_t x, char* buf);
+uint32_t mtl_utostrx(uint32_t x, char* buf);
+uint32_t mtl_itostr(int32_t x, char* buf);
+uint32_t mtl_itostrx(int32_t x, char* buf);
 #endif
 }
 
diff --git a/src/armv4t/to.s b/src/armv4t/to.s
index 87f5c1a..1546a23 100644
--- a/src/armv4t/to.s
+++ b/src/armv4t/to.s
@@ -2,244 +2,350 @@
 
 .include "mtl/armv4t/asm/math.s"
 
-.section .rodata
-
-.align 1
-hex_digits_lo:
-.string	"0123456789abcdef"
-hex_digits_up:
-.string "0123456789ABCDEF"
-
 .section .iwram, "ax", %progbits
 .arm
 .align 2
 
-/*
- * Writes the ASCII representation of the given unsigned value X to the buffer,
- * writes the beginning address of the string to an output ptr, and returns
- * the number of characters written. The value is written in hexidecimal
- * format.
- *
- * r0 = x
- * r1 = buffer start
- * r2 = buffer size
- * r3 = string start out ptr
- * ret= num of characters written
- */
-.global mtl_utostrx
-.type mtl_utostrx STT_FUNC
-mtl_utostrx:
-	// Store variables
-	push	{r4-r7}
-	// Move buffer to end
-	add	r1, r2
-	sub	r1, $1
-	// Add null terminator and decrement
-	mov	r6, $0
-	str	r6, [r1], $-1
-	// Load digit address
-	ldr	r4, =hex_digits_lo
-	// Zero num of characters
-	mov	r7, $0
-	.Ldecode_digitx:
-		// Increment number of characters
-		add	r7, $1
-		// Get least significant octet
-		and	r6, r0, 0xF
-		// Write the ascii digit to the string
-		ldrb	r5, [r4, r6]
-		strb	r5, [r1], $-1
-		// Shift to remove the least significant octet
-		lsrs	r0, $4
-		// Repeat if there are more digits left
-		bne	.Ldecode_digitx
-	// Move to the start of the string and account for "0x",
-	// currently one character before
-	sub	r1, $1
-	//add	r1, $1
-	// Write "0x" to the beginning of the string
-	mov	r6, '0'
-	strb	r6, [r1]
-	mov	r6, 'x'
-	strb	r6, [r1, $1]
-	// Write the start of the string to the pointer
-	str	r1, [r3]
-	// Return the number of characters written, account for "0x"
-	add	r0, r7, $2
-	// Restore variables
-	pop	{r4-r7}
-bx	lr
-
-/*
- * Writes the ASCII representation of the given unsigned value X to the buffer,
- * writes the beginning address of the string to an output ptr, and returns
- * the number of characters written. The value is written in decimal format.
- *
- * r0 = x
- * r1 = buffer start
- * r2 = buffer size
- * r3 = string start out ptr
- * ret= num of characters written
- */
-.global mtl_utostr
-.type mtl_utostr STT_FUNC
-mtl_utostr:
-	// Store variables
-	push	{r4-r7}
-	// Move buffer to end
-	add	r1, r2
-	sub	r1, $1
-	// Add null terminator and decrement
-	mov	r6, $0
-	str	r6, [r1], $-1
-	// Zero num of characters
-	mov	r7, $0
-	.Ldecode_digit:
-		// Increment number of characters
-		add	r7, $1
-		// Use % 10 to get the least significant digit
-		umod10	r6, r0, r12
-		// Write the ascii digit to the string
-		add	r5, r6, '0'
-		strb	r5, [r1], $-1
-		// Divide by 10 to remove the least significant digit
-		udiv10	r6, r0, r12
-		mov	r0, r6
-		// Repeat if there are more digits left
-		bne	.Ldecode_digit
-	// Move to the start of the string, currently one char before
-	add	r1, $1
-	// Write the start of the string to the pointer
-	str	r1, [r3]
-	// Return the number of characters written
-	mov	r0, r7
-	// Restore variables
-	pop	{r4-r7}
-bx	lr
-
 /*
  * Writes the ASCII representation of the given signed value X to the buffer,
- * writes the beginning address of the string to an output ptr, and returns
- * the number of characters written. The value is written in hexidecimal
- * format.
+ * returns the number of characters written. The value is written in decimal
+ * format and is null terminated. The string will always begin at the start of
+ * the buffer.
  *
  * r0 = x
- * r1 = buffer start
- * r2 = buffer size
- * r3 = string start out ptr
- * ret= num of characters written
- */
-.global mtl_itostrx
-.type mtl_itostrx STT_FUNC
-mtl_itostrx:
-	// Store variables
-	push	{r4-r8}
-	// r8 = 1 if negative, 0 if positive. Negate if negative
-	lsrs	r8, r0, $31
-	mvnne	r0, r0
-	addne	r0, $1
-	// Move buffer to end
-	add	r1, r2
-	sub	r1, $1
-	// Add null terminator and decrement
-	mov	r6, $0
-	str	r6, [r1], $-1
-	// Load digit address
-	ldr	r4, =hex_digits_lo
-	// Zero num of characters
-	mov	r7, $0
-	.Lsdecode_digitx:
-		// Increment number of characters
-		add	r7, $1
-		// Get least significant octet
-		and	r6, r0, 0xF
-		// Write the ascii digit to the string
-		ldrb	r5, [r4, r6]
-		strb	r5, [r1], $-1
-		// Shift to remove the least significant octet
-		lsrs	r0, $4
-		// Repeat if there are more digits left
-		bne	.Lsdecode_digitx
-	// Move to the start of the string and account for "0x",
-	// currently one character before
-	sub	r1, $1
-	// Write "0x" to the beginning of the string
-	mov	r6, '0'
-	strb	r6, [r1]
-	mov	r6, 'x'
-	strb	r6, [r1, $1]
-	// Check if number is negative
-	cmp	r8, $0
-	// Add the negative sign to the string and increase number of
-	// characters written if needed
-	subne	r1, $1
-	movne	r6, '-'
-	strbne	r6, [r1]
-	addne	r7, $1
-	// Write the start of the string to the pointer
-	str	r1, [r3]
-	// Return the number of characters written, account for "0x"
-	add	r0, r7, $2
-	// Restore variables
-	pop	{r4-r8}
-bx	lr
-/*
- * Writes the ASCII representation of the given signed value X to the buffer,
- * writes the beginning address of the string to an output ptr, and returns
- * the number of characters written. The value is written in decimal format.
- *
- * r0 = x
- * r1 = buffer start
- * r2 = buffer size
- * r3 = string start out ptr
+ * r1 = buffer
  * ret= num of characters written
  */
 .global mtl_itostr
 .type mtl_itostr STT_FUNC
 mtl_itostr:
-	// Store variables
-	push	{r4-r8}
-	// r8 = 1 if negative, 0 if positive. Negate if negative
-	lsrs	r8, r0, $31
-	mvnne	r0, r0
-	addne	r0, $1
-	// Move buffer to end
-	add	r1, r2
-	sub	r1, $1
-	// Add null terminator and decrement
-	mov	r6, $0
-	str	r6, [r1], $-1
-	// Zero num of characters
-	mov	r7, $0
-	.Lsdecode_digit:
-		// Increment number of characters
-		add	r7, $1
-		// Use % 10 to get the least significant digit
-		umod10	r6, r0, r12
-		// Write the ascii digit to the string
-		add	r5, r6, '0'
-		strb	r5, [r1], $-1
-		// Divide by 10 to remove the least significant digit
-		udiv10	r6, r0, r12
-		mov	r0, r6
-		// Repeat if there are more digits left
-		bne	.Lsdecode_digit
-	// Check if number is negative
-	cmp	r8, $0
-	// Add the negative sign to the string and increase number of
-	// characters written if needed
-	movne	r6, '-'
-	strbne	r6, [r1]
-	addne	r7, $1
-	// If the number was positive, currently positioned one char before str
-	// If the number was negative, currently at the correct position
-	// Move one character forward if needed
-	addeq	r1, $1
-	// Write the start of the string to the pointer
-	str	r1, [r3]
-	// Return the number of characters written
-	mov	r0, r7
-	// Restore variables
-	pop	{r4-r8}
+	// Branch immediately if the number is positive
+	lsrs	r12, r0, $31
+	beq	mtl_utostr
+	// Write the negative sign to the string
+	mvn	r0, r0
+	add	r0, $1
+	// Get the absolute value of the number
+	mov	r12, '-'
+	strb	r12, [r1], $1
+	// Write abs(x) to the string
+	push	{fp, lr}
+	bl	mtl_utostr
+	pop	{fp, lr}
+	// One more character was written to the string ('-')
+	add	r0, $1
+bx	lr
+
+/*
+ * Writes the ASCII representation of the given signed value X to the buffer,
+ * returns the number of characters written. The value is written in hexadecimal
+ * format and is null terminated. The string will always begin at the start of
+ * the buffer.
+ *
+ * r0 = x
+ * r1 = buffer
+ * ret= num of characters written
+ */
+.global mtl_itostrx
+.type mtl_itostrx STT_FUNC
+mtl_itostrx:
+	// Branch immediately if the number is positive
+	lsrs	r12, r0, $31
+	beq	mtl_utostrx
+	// Write the negative sign to the string
+	mov	r12, '-'
+	strb	r12, [r1], $1
+	// Get the absolute value of the number
+	mvn	r0, r0
+	add	r0, $1
+	// Write abs(x) to the string
+	push	{fp, lr}
+	bl	mtl_utostrx
+	pop	{fp, lr}
+	// One more character was written to the string ('-')
+	add	r0, $1
+bx	lr
+
+/*
+ * Writes the ASCII representation of the given unsigned value X to the buffer,
+ * returns the number of characters written. The value is written in decimal
+ * format and is null terminated. The string will always begin at the start of
+ * the buffer.
+ *
+ * Instead of using a loop like the previous implementation, this implementation
+ * uses a manually unrolled loop. This implementation also writes the digits
+ * from left to right instead of right to left. Using this method we can write
+ * the string to the beginning of the buffer and still avoid reversing the string.
+ *
+ * r0 = x
+ * r1 = buffer
+ * ret= num of characters written
+ */
+.global mtl_utostr
+.type mtl_utostr STT_FUNC
+mtl_utostr:
+// Store variables
+push	{r4-r6}
+	// Temporary used to calculate 10^P
+	mov	r5, $10
+	// Minimum of 1 digit
+	mov	r3, $1
+	// If less than 10, one digit, otherwise more
+	mov	r4, $10
+	cmp	r0, r4
+	blo	.Ldigits_1
+	// Now 2 digits
+	add	r3, $1
+	// If less than 100, two digits, otherwise more
+	mul	r4, r5
+	cmp	r0, r4
+	blo	.Ldigits_2
+	// Repeat...
+	add	r3, $1
+	mul	r4, r5
+	cmp	r0, r4
+	blo	.Ldigits_3
+	add	r3, $1
+	mul	r4, r5
+	cmp	r0, r4
+	blo	.Ldigits_4
+	add	r3, $1
+	mul	r4, r5
+	cmp	r0, r4
+	blo	.Ldigits_5
+	add	r3, $1
+	mul	r4, r5
+	cmp	r0, r4
+	blo	.Ldigits_6
+	add	r3, $1
+	mul	r4, r5
+	cmp	r0, r4
+	blo	.Ldigits_7
+	add	r3, $1
+	mul	r4, r5
+	cmp	r0, r4
+	blo	.Ldigits_8
+	add	r3, $1
+	mul	r4, r5
+	cmp	r0, r4
+	blo	.Ldigits_9
+	// 10 digits
+	add	r3, $1
+.Ldigits_10:
+	// Divide by 10^9 to get the most significant digit
+	udiv1000000000	r5, r0, r12
+	// Write the digit to the string
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	// Subtract the most significant digit
+	ldr	r6, =1000000000
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_9:
+	// Repeat...
+	udiv100000000	r5, r0, r12
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	ldr	r6, =100000000
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_8:
+	udiv10000000	r5, r0, r12
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	ldr	r6, =10000000
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_7:
+	udiv1000000	r5, r0, r12
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	ldr	r6, =1000000
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_6:
+	udiv100000	r5, r0, r12
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	ldr	r6, =100000
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_5:
+	udiv10000	r5, r0, r12
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	ldr	r6, =10000
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_4:
+	udiv1000	r5, r0, r12
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	mov	r6, $1000
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_3:
+	udiv100	r5, r0, r12
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	mov	r6, $100
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_2:
+	udiv10	r5, r0, r12
+	add	r6, r5, '0'
+	strb	r6, [r1], $1
+	mov	r6, $10
+	mul	r6, r5
+	sub	r0, r6
+.Ldigits_1:
+	add	r6, r0, '0'
+	strb	r6, [r1], $1
+
+	// Write the null terminator
+	mov	r6, $0
+	strb	r6, [r1]
+	// Return number of digits written
+	mov	r0, r3
+// Restore variables
+pop	{r4-r6}
+bx	lr
+
+/*
+ * Writes the ASCII representation of the given unsigned value X to the buffer,
+ * returns the number of characters written. The value is written in hexadecimal
+ * format and is null terminated. The string will always begin at the start of
+ * the buffer.
+ *
+ * Instead of using a loop like the previous implementation, this implementation
+ * uses a manually unrolled loop. This implementation also writes the digits
+ * from left to right instead of right to left. Using this method we can write
+ * the string to the beginning of the buffer and still avoid reversing the string.
+ *
+ * r0 = x
+ * r1 = buffer
+ * ret= num of characters written
+ */
+.global mtl_utostrx
+.type mtl_utostrx STT_FUNC
+mtl_utostrx:
+// Store variables
+push	{r4-r6}
+	// Write "0x" to the string
+	mov	r6, '0'
+	strb	r6, [r1], $1
+	mov	r6, 'x'
+	strb	r6, [r1], $1
+
+	// Minimum of 3 characters (including "0x")
+	mov	r3, $3
+	// If less than 0x10, one digit, otherwise more
+	mov	r4, $0x10
+	cmp	r0, r4
+	blo	.Lxdigits_1
+	// Now four characters
+	add	r3, $1
+	// If less than 0x100, two digits, otherwise more
+	lsl	r4, $4
+	cmp	r0, r4
+	blo	.Lxdigits_2
+	// Repeat...
+	add	r3, $1
+	lsl	r4, $4
+	cmp	r0, r4
+	blo	.Lxdigits_3
+	add	r3, $1
+	lsl	r4, $4
+	cmp	r0, r4
+	blo	.Lxdigits_4
+	add	r3, $1
+	lsl	r4, $4
+	cmp	r0, r4
+	blo	.Lxdigits_5
+	add	r3, $1
+	lsl	r4, $4
+	cmp	r0, r4
+	blo	.Lxdigits_6
+	add	r3, $1
+	lsl	r4, $4
+	cmp	r0, r4
+	blo	.Lxdigits_7
+	// Eight digits
+	add	r3, $1
+.Lxdigits_8:
+	// Right shift to get the most significant digit
+	lsr	r5, r0, $28
+	// Get digit character
+	add	r6, r5, '0'
+	// If the digit > 9, change to a lowercase letter
+	cmp	r6, '9'
+	addhi	r6, $0x27
+	// Write the digit
+	strb	r6, [r1], $1
+	// Subtract the digit
+	lsl	r5, $28
+	sub	r0, r5
+.Lxdigits_7:
+	// Repeat...
+	lsr	r5, r0, $24
+	add	r6, r5, '0'
+	cmp	r6, '9'
+	addhi	r6, $0x27
+	strb	r6, [r1], $1
+	lsl	r5, $24
+	sub	r0, r5
+.Lxdigits_6:
+	lsr	r5, r0, $20
+	add	r6, r5, '0'
+	cmp	r6, '9'
+	addhi	r6, $0x27
+	strb	r6, [r1], $1
+	lsl	r5, $20
+	sub	r0, r5
+.Lxdigits_5:
+	lsr	r5, r0, $16
+	add	r6, r5, '0'
+	cmp	r6, '9'
+	addhi	r6, $0x27
+	strb	r6, [r1], $1
+	lsl	r5, $16
+	sub	r0, r5
+.Lxdigits_4:
+	lsr	r5, r0, $12
+	add	r6, r5, '0'
+	cmp	r6, '9'
+	addhi	r6, $0x27
+	strb	r6, [r1], $1
+	lsl	r5, $12
+	sub	r0, r5
+.Lxdigits_3:
+	lsr	r5, r0, $8
+	add	r6, r5, '0'
+	cmp	r6, '9'
+	addhi	r6, $0x27
+	strb	r6, [r1], $1
+	lsl	r5, $8
+	sub	r0, r5
+.Lxdigits_2:
+	lsr	r5, r0, $4
+	add	r6, r5, '0'
+	cmp	r6, '9'
+	addhi	r6, $0x27
+	strb	r6, [r1], $1
+	lsl	r5, $4
+	sub	r0, r5
+.Lxdigits_1:
+	add	r6, r0, '0'
+	cmp	r6, '9'
+	addhi	r6, $0x27
+	strb	r6, [r1], $1
+
+	// Write the null terminator
+	mov	r6, $0
+	strb	r6, [r1]
+	// Return the number of digits written
+	mov	r0, r3
+// Restore variables
+pop	{r4-r6}
 bx	lr