diff --git a/CREDITS.md b/CREDITS.md
index 72143f5f..a1ebe654 100644
--- a/CREDITS.md
+++ b/CREDITS.md
@@ -6,6 +6,7 @@
 * TWW/CTR - for his keyboard handling code
 * Graham - for his PAL/NTSC detection routine
 * enthusi/Onslaught - for his example Turbo Tape 64 compatible loader
+* djmips - for fast 8x8 multiplication routine
 
 ## for documentation
 
diff --git a/c64/basic/math/mul_FAC2_FAC1.s b/c64/basic/math/mul_FAC2_FAC1.s
index 35b71b39..44c21c17 100644
--- a/c64/basic/math/mul_FAC2_FAC1.s
+++ b/c64/basic/math/mul_FAC2_FAC1.s
@@ -88,3 +88,63 @@ mul_FAC2_FAC1_sub_exp_bias:
 mul_FAC2_FAC1_done:
 
 	rts
+
+
+mul_8x8: // XXX we should probably move this subroutine to separate file
+
+	// Routine based on code by djmips, taken from:
+	// - https://codebase64.org/doku.php?id=base:8bit_multiplication_16bit_product_fast_no_tables
+	//
+	// input: mul1, mul2
+	// optput: .A (high byte), mul1 (low byte)
+
+.label mul1 = INDEX+2
+.label mul2 = INDEX+3
+
+	lda #$00
+
+	dec mul2	// decrement because we will be adding with carry set for speed (an extra one)
+	ror mul1
+	bcc !+
+	adc mul2
+!:
+	ror
+	ror mul1
+	bcc !+
+	adc mul2
+!:
+	ror
+	ror mul1
+	bcc !+
+	adc mul2
+!:
+	ror
+	ror mul1
+	bcc !+
+	adc mul2
+!:
+	ror
+	ror mul1
+	bcc !+
+	adc mul2
+!:
+	ror
+	ror mul1
+	bcc !+
+	adc mul2
+!:
+	ror
+	ror mul1
+	bcc !+
+	adc mul2
+!:
+	ror
+	ror mul1
+	bcc !+
+	adc mul2
+!:
+	ror
+	ror mul1
+	inc mul2
+
+	rts