added memory_set'function

Makefile

@@ -7,6 +7,11 @@ LD_FLAGS := --strip-all
 DEBUG_LD_FLAGS := -g
 DEBUG_NASM_FLAGS := -g -F dwarf
 
+# check for avx2 support
+ifeq ($(shell grep -o 'avx2[^ ]*' /proc/cpuinfo | head -n 1),avx2)
+    NASM_FLAGS += -DAVX2
+endif
+
 
 SRC_PATH := src
 OBJ_PATH := build/obj

src/drawing.asm

@@ -6,13 +6,16 @@ section .bss
 	extern term_rows
 	extern term_cols
 
-section .data 
+	simulation_running: RESB 1
+
+section .rodata 
 	clear: 		db ESC_CHAR, "[2J", 0
 	reset:		db ESC_CHAR, "[0m", 0
 
 	statusbar:	db ESC_CHAR, "[100m", "Use arrow keys to move cursor, enter to invert cell, p to       simulation", 0 
 	
-
+	start_str:	db "START", 0
+	stop_str:	db "STOP", 0
 
 section .text 
 extern print_str
@@ -51,4 +54,5 @@ print_game_ui:
 
 
 	
+	
 	ret

src/main.asm

@@ -3,15 +3,17 @@
 section .bss
 	multipurpuse_buf: RESB 8
 	
+	global term_rows
 	term_rows: RESW 1
+	global term_cols
 	term_cols: RESW 1
 	
+	global gameboard_ptr
 	gameboard_ptr: RESQ 1
 
 	extern cursor_rows
 	extern cursor_cols
 
-section .data
 	
 section .text
 extern print_str

src/str.asm

@@ -60,15 +60,83 @@ unsigned_int_to_ascii: ; takes pointer to array in rdi and value stored in rsi D
 
     ret
 
+global string_copy
 string_copy:; takes pointer to destination in rdi and pointer to source in rsi
 	
 	xor rax, rax
 
 	.copy_next_byte:
-	mov byte al, [rdi+rax]
+	mov byte cl, [rsi+rax]
-	mov [rsi+rax], al
+	test cl, cl
+	jz .exit
+	mov [rdi+rax], cl
 	inc rax
-	test rax,rax
+	jmp .copy_next_byte
-	jnz .copy_next_byte
+
+	.exit:
 	ret
 
+global memory_set:
+memory_set:; takes  destination in rdi, byte in sil and lenght in rdx
+	; first check if value is 16 byte alligned
+
+	mov r9, rdi
+
+	mov rax, rdi
+	and rax, 0xF; offset is stored in rax
+	test al, al; check if resault is 0
+	jz .addr_is_16_Byte_alligned
+	mov r8b, 16
+	sub r8b, al; now offset to first higher 16 byte alligned address is stored in r8
+	sub rdx, r8; we will write these bytes now
+	
+	movzx rax, sil
+	imul rax, 0x01010101; to extend across whoule register
+	shl rax, 32; to extend across whoule register
+
+	;add rdi, rdx
+	; we know that rdi has initial address and rdx offset so well fill just add to it
+	mov rcx, 1; we will allwais copy only once
+	
+
+	.check_qword:; check if offset is more than qword
+	cmp r8b, 8
+	jl .check_dword
+	rep stosq
+
+	.check_dword:
+	cmp r8b, 4
+	jl .check_word
+	rep stosd
+
+	.check_word:
+	cmp r8b, 2
+	jl .check_byte
+	rep stosw
+
+	.check_byte:
+	test r8b, r8b; check if offset is 1 or 0
+	jz .addr_is_16_Byte_alligned
+	rep stosb
+	
+	
+	.addr_is_16_Byte_alligned:
+	shr rdx, 4; set it to how many 128bit(16Byte) chunk we need 
+	
+	%ifdef AVX2
+		vpbroadcastq xmm8, rax
+	%else
+		movq xmm8, rax
+		shufpd xmm8, xmm8, 0x00
+	%endif
+
+	.move_16_bytes:
+	movdqa [rdi], xmm8
+	add rdi, 16
+	dec rdx
+
+	test rdx,rdx; test if rdx is zero 
+	jnz .move_16_bytes
+	
+	mov rax, r9; return pointer to memory area same as memset in libc
+	ret