; Input    
;   eax   dword for print    
;   rdi   buffer    
; Usage    
;   edx    tmp    
PrintToBufferDwordHex:             
repeat 8       
  rol eax,4    
  mov dl,al                            
  and dl,00Fh                          
  cmp dl,00Ah                          
  jb @f                                
  add dl,7                             
@@:                                    
  add dl,'0'                           
  mov [rdi+%-1],dl                     
end repeat                             
                                       
  ret   
    

@@:
processed: @@:
error: symbol already defined.
    

; Input
;   eax   dword for print
;   rdi   buffer
PrintToBufferDwordHex:
  push rdx
  mov edx,eax
rept 8 n:0 {
  rol eax,4
  mov dl,al  
  and dl,00Fh
  cmp dl,00Ah
  jb @f   
  add dl,7
@@:         
  add dl,'0'        
  mov [rdi+n],dl
}
  pop rdx
  ret
    

Found beautiful solution!

Code:

...    

Why not just use plain assembler loops? 8 times more code might in most cases be worse than a small loop.

; Input    
;   eax   dword for print    
;   rdi   buffer    
PrintToBufferHexDWord:    
  push rcx rdx    
  mov ecx,8    
align 16    
.char:    
  shl rdx,8    
  mov dl,al    
  and dl,00Fh    
  cmp dl,00Ah    
  jb @f    
  add dl,'A'-'9'-1                              ; 7    
@@:    
  add dl,'0'    
  ror eax,4    
  loop .char    
  mov [rdi],rdx    
  pop rdx rcx    
  ret