format ELF64 executable 3 entry start segment readable executable ;============================================================================== macro IACA_START { ; mov ebx,111 ; db 0x64,0x67,0x90 } macro IACA_END { ; mov ebx,222 ; db 0x64,0x67,0x90 } ;============================================================================== align 16 thread: ;------------------------------------------------------------------------------ virtual at rsp .startx dd ? .starty dd ? end virtual push rbp rbx mov rbp,rsp sub rsp,32 and rsp,-32 .loop_tile: mov r15d,1 lock xadd [image_tile],r15d cmp r15d,TILE_COUNT jge .finish xor edx,edx mov eax,r15d mov edi,TILE_X_COUNT div edi imul eax,TILE_SIZE imul edx,TILE_SIZE mov [.startx],edx mov [.starty],eax imul eax,IMAGE_SIZE add eax,edx shl eax,2 mov rbx,[image_ptr] add rbx,rax xor r14d,r14d align 32 .loop_row: xor r13d,r13d align 32 .loop_pixel: mov edx,[.startx] mov eax,[.starty] add edx,r13d add eax,r14d ;------------------------------------ compute color vmovaps ymm5,[image_size_rcp] vmovaps ymm6,[.c_m0_5] vxorps ymm0,ymm0,ymm0 vxorps ymm2,ymm2,ymm2 vcvtsi2ss xmm0,xmm0,edx vcvtsi2ss xmm2,xmm2,eax vbroadcastss ymm0,xmm0 vbroadcastss ymm2,xmm2 vmovaps ymm1,ymm0 vaddps ymm0,ymm0,[.c_0_to_7] vaddps ymm1,ymm1,[.c_8_to_15] vfmadd213ps ymm0,ymm5,ymm6 vfmadd213ps ymm1,ymm5,ymm6 vfmadd213ps ymm2,ymm5,ymm6 vaddps ymm13,ymm0,ymm0 ; ymm13 = a0 vaddps ymm14,ymm1,ymm1 ; ymm14 = a1 vaddps ymm15,ymm2,ymm2 ; ymm15 = b vaddps ymm13,ymm13,[.c_m0_5] vaddps ymm14,ymm14,[.c_m0_5] vxorps ymm0,ymm0,ymm0 ; ymm0 = x0 vxorps ymm1,ymm1,ymm1 ; ymm1 = x1 vxorps ymm2,ymm2,ymm2 ; ymm2 = y0 vxorps ymm3,ymm3,ymm3 ; ymm3 = y1 vxorps ymm4,ymm4,ymm4 ; ymm4 = x0^2 vxorps ymm5,ymm5,ymm5 ; ymm5 = x1^2 vxorps ymm6,ymm6,ymm6 ; ymm6 = y0^2 vxorps ymm7,ymm7,ymm7 ; ymm7 = y1^2 vmovaps ymm12,[.c_4_0] ; ymm12 = 4.0 vpcmpeqd ymm10,ymm10,ymm10 ; ymm10 = (a0,b) iterations vpcmpeqd ymm11,ymm11,ymm11 ; ymm11 = (a1,b) iterations mov eax,256 align 32 .loop: vaddps ymm0,ymm0,ymm0 ; ymm0 = 2*x0 vaddps ymm1,ymm1,ymm1 ; ymm1 = 2*x1 vfmadd213ps ymm2,ymm0,ymm15 ; ymm2 = y0n = 2*x0*y0 + b vfmadd213ps ymm3,ymm1,ymm15 ; ymm3 = y1n = 2*x1*y1 + b vsubps ymm0,ymm4,ymm6 ; ymm0 = x0^2 - y0^2 vsubps ymm1,ymm5,ymm7 ; ymm1 = x1^2 - y1^2 vaddps ymm0,ymm0,ymm13 ; ymm0 = x0n = x0^2 - y0^2 + a vaddps ymm1,ymm1,ymm14 ; ymm1 = x1n = x1^2 - y1^2 + a vmulps ymm4,ymm0,ymm0 ; ymm4 = x0n^2 vmulps ymm5,ymm1,ymm1 ; ymm5 = x1n^2 vmulps ymm6,ymm2,ymm2 ; ymm6 = y0n^2 vmulps ymm7,ymm3,ymm3 ; ymm7 = y1n^2 vaddps ymm8,ymm4,ymm6 ; ymm8 = m0 = x0n^2 + y0n^2 vaddps ymm9,ymm5,ymm7 ; ymm9 = m1 = x1n^2 + y1n^2 vcmpltps ymm8,ymm8,ymm12 ; ymm8 = ymm8 < 4.0 vcmpltps ymm9,ymm9,ymm12 ; ymm9 = ymm9 < 4.0 vpsubd ymm10,ymm10,ymm8 ; increment ymm10 when m0 < 4.0 vpsubd ymm11,ymm11,ymm9 ; increment ymm11 when m1 < 4.0 sub eax,1 jnz .loop vmovdqa ymm0,[.c_255] vmovaps ymm2,[.c_1div255] vpsubd ymm10,ymm0,ymm10 vpsubd ymm11,ymm0,ymm11 vcvtdq2ps ymm0,ymm10 vcvtdq2ps ymm1,ymm11 vmulps ymm0,ymm0,ymm2 vmulps ymm1,ymm1,ymm2 vmulps ymm0,ymm0,ymm0 vmulps ymm1,ymm1,ymm1 vmulps ymm0,ymm0,ymm0 vmulps ymm1,ymm1,ymm1 vmulps ymm0,ymm0,ymm0 vmulps ymm1,ymm1,ymm1 vmovaps ymm2,[.c_255_0] vmulps ymm0,ymm0,ymm2 vmulps ymm1,ymm1,ymm2 vcvttps2dq ymm0,ymm0 vcvttps2dq ymm1,ymm1 vmovdqa ymm6,[.c_ff000000] vpslld ymm2,ymm0,8 vpslld ymm4,ymm1,8 vpslld ymm3,ymm0,16 vpslld ymm5,ymm1,16 vpor ymm0,ymm0,ymm2 vpor ymm1,ymm1,ymm4 vpor ymm3,ymm3,ymm6 vpor ymm5,ymm5,ymm6 vpor ymm0,ymm0,ymm3 vpor ymm1,ymm1,ymm5 vmovdqa [rbx+r13*4],ymm0 vmovdqa [rbx+r13*4+32],ymm1 ;------------------------------------ add r13d,16 cmp r13d,TILE_SIZE jne .loop_pixel add rbx,IMAGE_SIZE*4 add r14d,1 cmp r14d,TILE_SIZE jne .loop_row jmp .loop_tile .finish: mov rsp,rbp pop rbx rbp ret align 32 .c_0_to_7: dd 0.0,1.0,2.0,3.0,4.0,5.0,6.0,7.0 .c_8_to_15: dd 8.0,9.0,10.0,11.0,12.0,13.0,14.0,15.0 .c_16_to_23: dd 16.0,17.0,18.0,19.0,20.0,21.0,22.0,23.0 .c_24_to_31: dd 24.0,25.0,26.0,27.0,28.0,29.0,30.0,31.0 .c_m0_5: dd 8 dup -0.5 .c_255: dd 8 dup 255 .c_255_0: dd 8 dup 255.0 .c_4_0: dd 8 dup 4.0 .c_ff000000: dd 8 dup 0xff000000 .c_1div255: dd 8 dup 0.003921569 ;============================================================================== align 16 mem_alloc: ; in: size in bytes ; out: pointer to the allocated memory ;------------------------------------------------------------------------------ mov eax,9 ; sys_mmap mov rsi,rdi ; length xor edi,edi ; addr mov edx,0x1+0x2 ; PROT_READ | PROT_WRITE mov r10d,0x02+0x20 ; MAP_PRIVATE | MAP_ANONYMOUS mov r8,-1 ; fd xor r9d,r9d ; offset syscall ret ;============================================================================== align 16 thread_create: ; in: mutex address ;------------------------------------------------------------------------------ mov edi,4096 call mem_alloc mov rsi,rax add rsi,4096 mov eax,56 ; sys_clone mov edi,0x100+0x200+0x400+0x800+0x10000 xor edx,edx syscall test eax,eax jnz .ret call thread mov dword [rbx],1 mov eax,202 ; sys_futex mov rdi,rbx ; mutex address mov esi,1 ; FUTEX_WAKE mov edx,1 ; wake 1 thread syscall mov eax,60 ; sys_exit xor edi,edi ; exit code syscall .ret: ret ;============================================================================== align 16 thread_wait: ; in: mutex address ;------------------------------------------------------------------------------ mov eax,202 ; sys_futex mov rdi,rbx ; mutex address mov esi,0 ; FUTEX_WAIT mov edx,0 ; mutex 'running' value xor r10d,r10d ; unused but must be zero syscall ret ;============================================================================== align 16 image_save: ;------------------------------------------------------------------------------ push rbx mov eax,85 mov rdi,.tga_name mov esi,110000000b syscall mov rbx,rax mov eax,1 mov rdi,rbx mov rsi,.tga_head mov edx,18 syscall mov eax,1 mov rdi,rbx mov rsi,[image_ptr] mov edx,IMAGE_SIZE*IMAGE_SIZE*4 syscall pop rbx ret .tga_name db 'mandelbrot2.tga',0 .tga_head db 0,0,2,9 dup 0 db (IMAGE_SIZE and 0x00ff),(IMAGE_SIZE and 0xff00) shr 8 db (IMAGE_SIZE and 0x00ff),(IMAGE_SIZE and 0xff00) shr 8,32,0 ;============================================================================== align 16 main: ;------------------------------------------------------------------------------ mov edi,IMAGE_SIZE*IMAGE_SIZE*4 call mem_alloc mov [image_ptr],rax xor r12d,r12d .loop_create: lea rbx,[thread_mutex+r12*4] call thread_create add r12d,1 cmp r12d,THREAD_COUNT jne .loop_create xor r12d,r12d .loop_wait: lea rbx,[thread_mutex+r12*4] call thread_wait add r12d,1 cmp r12d,THREAD_COUNT jne .loop_wait call image_save ret ;============================================================================== align 16 start: ;------------------------------------------------------------------------------ call main mov eax,60 ; sys_exit xor edi,edi ; exit code syscall ;============================================================================== segment readable writeable align 4 THREAD_COUNT = 8 thread_mutex dd 0,0,0,0,0,0,0,0 align 8 image_ptr dq 0 align 4 image_tile dd 0 TILE_SIZE = 64 TILE_X_COUNT = IMAGE_SIZE / TILE_SIZE TILE_Y_COUNT = IMAGE_SIZE / TILE_SIZE TILE_COUNT = TILE_X_COUNT * TILE_Y_COUNT IMAGE_SIZE = 1024 align 32 image_size_rcp: dd 8 dup 0.000976563 ;==============================================================================