Text only | Text with Images

The MASM Forum

General => The Laboratory => Topic started by: frktons on January 21, 2013, 04:32:29 AM

Title: Fastest way to move 3 bytes into a dword
Post by: frktons on January 21, 2013, 04:32:29 AM
Next challenge on the way to enlightenment  is:

What is the fastest way to move a 3 bytes variable
into a dword variable/register?

And the reverse, of course.

Let's see what shows up this time  :P

I bet Dave will enjoy this one.  :lol:
Title: Re: Fastest way to move 3 bytes into a dword
Post by: dedndave on January 21, 2013, 04:40:08 AM
Code Select
mov eax,dword ptr var3bytes
and eax,0FFFFFFh
:P
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 21, 2013, 05:28:14 AM
Quote from: dedndave on January 21, 2013, 04:40:08 AM
Code Select
mov eax,dword ptr var3bytes
and eax,0FFFFFFh
:P

Good idea Dave, I like it.  :t

The reverse operation is still missing. Dave, give us the light.

I'm very curious  to see if more solutions come up & what kind
of bit-imagination assembly programmers have developed  :lol:
Title: Re: Fastest way to move 3 bytes into a dword
Post by: Siekmanski on January 21, 2013, 07:43:45 AM
Quote from: Siekmanski on January 21, 2013, 07:30:41 AM

24bit to 32 bit,


* code modified,

Code Select

.data
align 16
Buffer32               db 32*1024 dup (0)
ByteAndMask32          db -1,-1,-1,0,-1,-1,-1,0,-1,-1,-1,0,-1,-1,-1,0
ByteMask24BitSSE3      db 0,1,2,0,3,4,5,0,6,7,8,0,9,10,11,0

Bytes24bit             db 1,2,3,4,5,6,7,8,9,10,11,12 ; etc.

.code

lea eax,ByteMask24BitSSE3
movdqa xmm1,[eax]
lea eax,ByteAndMask32
movdqa xmm2,[eax]

lea esi,Bytes24bit
lea edi,Buffer32

; mov ecx,1024
align 16
conversion_loop:
movdqu xmm0,[esi]
pshufb xmm0,xmm1
pand xmm0,xmm2
movdqa [edi],xmm0
; movdqu xmm0,[esi+12]
; pshufb xmm0,xmm1
; pand xmm0,xmm2
; movdqa [edi+16],xmm0
; add esi,24
; add edi,32
; dec ecx
; jnz conversion_loop



1,2,3,4,5,6,7,8,9,10,11,12
results in:
00030201h 00060504h 00090807h 000c0b0ah
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 21, 2013, 08:47:07 AM
Quote from: Siekmanski on January 21, 2013, 07:43:45 AM

Code Select

.data
align 16
Buffer32               db 32*1024 dup (0)
ByteAndMask32          db -1,-1,-1,0,-1,-1,-1,0,-1,-1,-1,0,-1,-1,-1,0
ByteMask24BitSSE3      db 0,1,2,0,3,4,5,0,6,7,8,0,9,10,11,0

Bytes24bit             db 1,2,3,4,5,6,7,8,9,10,11,12 ; etc.

.code

lea eax,ByteMask24BitSSE3
movdqa xmm1,[eax]
lea eax,ByteAndMask32
movdqa xmm2,[eax]

lea esi,Bytes24bit
lea edi,Buffer32

; mov ecx,1024
align 16
conversion_loop:
movdqu xmm0,[esi]
pshufb xmm0,xmm1
pand xmm0,xmm2
movdqa [edi],xmm0
; movdqu xmm0,[esi+12]
; pshufb xmm0,xmm1
; pand xmm0,xmm2
; movdqa [edi+16],xmm0
; add esi,24
; add edi,32
; dec ecx,1
; jnz conversion_loop



1,2,3,4,5,6,7,8,9,10,11,12
results in:
00030201h 00060504h 00090807h 000c0b0ah

Very interesting Siekmanski, I want to test it to see the performance
of SSE code against traditional 32 bit code.   :t
Title: Re: Fastest way to move 3 bytes into a dword
Post by: Siekmanski on January 21, 2013, 09:05:40 AM
32bit to 24 bit,

:biggrin:

Code Select


.data
align 16
Bytes32bit             dd 00030201h,00060504h,00090807h,000c0b0ah ; etc.
ByteMask24BitSSE3_2    db 0,1,2,4,5,6,8,9,10,12,13,14,3,3,3,3
Buffer24        db 24*1024+4 dup (0)


.code

lea eax,ByteMask24BitSSE3_2
movdqa xmm1,[eax]

lea esi,Bytes32bit
lea edi,Buffer24

; mov ecx,1024
align 16
conversion_loop2:
movdqa xmm0,[esi]
pshufb xmm0,xmm1
movdqu [edi],xmm0
; movdqa xmm0,[esi+16]
; pshufb xmm0,xmm1
; movdqu [edi+12],xmm0
; add esi,32
; add edi,24
; dec ecx
; jnz conversion_loop2



00030201h 00060504h 00090807h 000c0b0ah
results in:
1,2,3,4,5,6,7,8,9,10,11,12
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 21, 2013, 09:19:17 AM
Quote from: Siekmanski on January 21, 2013, 09:05:40 AM
32bit to 24 bit,

:biggrin:

Code Select


.data
align 16
Bytes32bit             dd 00030201h,00060504h,00090807h,000c0b0ah ; etc.
ByteMask24BitSSE3_2    db 0,1,2,4,5,6,8,9,10,12,13,14,3,3,3,3
Buffer24        db 24*1024+4 dup (0)


.code

lea eax,ByteMask24BitSSE3_2
movdqa xmm1,[eax]

lea esi,Bytes32bit
lea edi,Buffer24

; mov ecx,1024
align 16
conversion_loop2:
movdqa xmm0,[esi]
pshufb xmm0,xmm1
movdqu [edi],xmm0
; movdqa xmm0,[esi+16]
; pshufb xmm0,xmm1
; movdqu [edi+12],xmm0
; add esi,32
; add edi,24
; dec ecx,1
; jnz conversion_loop2



00030201h 00060504h 00090807h 000c0b0ah
results in:
1,2,3,4,5,6,7,8,9,10,11,12


Yes, this was the missing part. Very nice. I'll test them ASAP.   :t
Title: Re: Fastest way to move 3 bytes into a dword
Post by: Siekmanski on January 21, 2013, 09:31:47 AM
 :icon_redface: Made a stupid mistake,

dec ecx,1

must be: dec ecx

sources modified....   :biggrin:
Title: Re: Fastest way to move 3 bytes into a dword
Post by: Siekmanski on January 21, 2013, 10:02:06 AM
To get faster results,unrole the conversionloop 3 times to run in L1 cache (64 byte)
This 24 bit to 32 bit routine is now 60 bytes long and fits in the L1 cache.
( The 32 bit to 24 bit routine unroled 3 times is 65 byte so 1 byte to big to fit the L1 cache.)

Code Select

mov ecx,128
align 16
conversion_loop2:
movdqa xmm0,[esi]
pshufb xmm0,xmm1
movdqu [edi],xmm0
movdqa xmm0,[esi+16]
pshufb xmm0,xmm1
movdqu [edi+12],xmm0
movdqa xmm0,[esi+32]
pshufb xmm0,xmm1
movdqu [edi+24],xmm0
add esi,48
add edi,36
dec ecx
jnz conversion_loop2


Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 21, 2013, 10:15:27 AM
Quote from: Siekmanski on January 21, 2013, 10:02:06 AM
To get faster results,unrole the conversionloop 3 times to run in L1 cache (64 byte)
This 24 bit to 32 bit routine is now 60 bytes long and fits in the L1 cache.
( The 32 bit to 24 bit routine unroled 3 times is 65 byte so 1 byte to big to fit the L1 cache.)

Code Select

mov ecx,128
align 16
conversion_loop2:
movdqa xmm0,[esi]
pshufb xmm0,xmm1
movdqu [edi],xmm0
movdqa xmm0,[esi+16]
pshufb xmm0,xmm1
movdqu [edi+12],xmm0
movdqa xmm0,[esi+32]
pshufb xmm0,xmm1
movdqu [edi+24],xmm0
add esi,48
add edi,36
dec ecx
jnz conversion_loop2




I'm preparing the test program, but I'm not sure it'll be ready soon.
It is night and I'm almost sleeping.   :dazzled:
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 21, 2013, 10:32:25 AM
This is the structure of the test program:
Code Select

;==============================================================================
; Test_mov3todw.asm
; ------------------------------------------------------------------------
; Example to test instructions that mov 3 bytes vars into dwords.
; The test uses 48 bytes string to be read 3 bytes each time, into 16 DW.
; ------------------------------------------------------------------------
; Frktons 20-jan-2013 @Masm32 Forum
;==============================================================================
include \masm32\include\masm32rt.inc
;==============================================================================
.nolist
.686
.xmm

include \masm32\macros\timers.asm
; get them from the
;[url=http://www.masm32.com/board/index.php?topic=770.0]Masm32 Laboratory[/url]

AxCPUid_Print PROTO

LOOP_COUNT EQU 1000

include \masm32\include\MyLib.inc


;==============================================================================
.data

    align 16
    Area  DB "Here it is a string with 48 characters inside me",0
    AreaLn = ($ - Area - 1)
    align Four   
    AreaLen    dd 0
    Counter    dd 0
    PtrSource  dd Area
    PtrDest    dd ArrayDW
 

    align Four
    LineSep     db  72 dup("-"),0,0,0,0

    align Four
    PtrLineSep  dd  LineSep
 

.data?

    align 16
    ArrayDW  dd 16 DUP (?)
    align Four
    CPU_Count DD  ?             ; Number of Cycles elapsed   
   


   
.code
;==============================================================================
align Four
MovProc proc

    mov edx, 1000      ; Number of cycles to perform

align Four   
TotCycles:

    mov esi, PtrSource
    mov edi, PtrDest
    mov ecx, 16
align Four   
cycle:
    mov eax,   [esi]
    and eax,   00FFFFFFH
    mov [edi], eax
    add esi,   3
    add edi,   Four

    dec ecx
    jnz cycle

    dec edx
    jnz TotCycles


    ret

MovProc endp

;==============================================================================
align Four
DisplayArrayDW  proc

    mov  ecx, 0
    mov  edx, PtrDest

Display:

    pushad

    print DWORD PTR edx

    popad
   
    add   edx, Four
    inc   ecx
    cmp   ecx, 16
    jnz   Display

    ret

DisplayArrayDW  endp
;==============================================================================
align Four
Main proc



    invoke GetLocaleInfo,LOCALE_USER_DEFAULT,LOCALE_STHOUSAND,offset Tsep,Four
    invoke CharToOem,offset Tsep,offset Tsep

    CALL   FillMyArray

    CALL   FillMyArray0

    INVOKE ConsoleSize, 40, 100

    print PtrLineSep, 13, 10
   
    invoke AxCPUid_Print

    print PtrLineSep, 13, 10

    REPEAT Four

;---------------------------------------------------------------------------------

invoke Sleep, 100
counter_begin LOOP_COUNT, HIGH_PRIORITY_CLASS
       
      CALL MovProc

counter_end

      mov  edi, PtrFmtNum16
      lea  esi, InitString
      movdqa  xmm0, [esi]
      movdqa  [edi], xmm0

      INVOKE FormatNumDW, eax, PtrFmtNum16
     
print PtrFmtNum16, 9,  "cycles for Dave - MOV 4 bytes / AND", 13, 10

;---------------------------------------------------------------------------------

      print PtrLineSep, 13, 10
     
       
    ENDM 
   
;    CALL DisplayArrayDW
 

    ret

Main endp


;-------------------------------------------------------------

    include AxCPUid.inc

;-------------------------------------------------------------

;==============================================================================
start:
;==============================================================================

;==============================================================================
    call Main

    inkey
    exit
;==============================================================================
end start


If you are still awake, try to adapt your code for the task.
Attached the files you need.

Frank
Title: Re: Fastest way to move 3 bytes into a dword
Post by: Siekmanski on January 21, 2013, 12:56:33 PM
Inserted my routines.

Code Select
------------------------------------------------------------------------
Intel(R) Core(TM)2 Quad CPU    Q6600  @ 2.40GHz

Instructions: MMX, SSE1, SSE2, SSE3, SSSE3
------------------------------------------------------------------------
       49.992   cycles for Dave - 48 bytes MOV 4 bytes / AND
------------------------------------------------------------------------
       48.133   cycles for Dave - 48 bytes MOV 4 bytes / AND
------------------------------------------------------------------------
       48.143   cycles for Dave - 48 bytes MOV 4 bytes / AND
------------------------------------------------------------------------
       48.144   cycles for Dave - 48 bytes MOV 4 bytes / AND
------------------------------------------------------------------------
       23.103   cycles for Siekmanski - 48 bytes SSSE3_24_32
------------------------------------------------------------------------
       23.152   cycles for Siekmanski - 48 bytes SSSE3_24_32
------------------------------------------------------------------------
       23.137   cycles for Siekmanski - 48 bytes SSSE3_24_32
------------------------------------------------------------------------
       23.136   cycles for Siekmanski - 48 bytes SSSE3_24_32
------------------------------------------------------------------------
       20.138   cycles for Siekmanski - 48 bytes SSSE3_24_32 unroled
------------------------------------------------------------------------
       20.124   cycles for Siekmanski - 48 bytes SSSE3_24_32 unroled
------------------------------------------------------------------------
       20.130   cycles for Siekmanski - 48 bytes SSSE3_24_32 unroled
------------------------------------------------------------------------
       20.137   cycles for Siekmanski - 48 bytes SSSE3_24_32 unroled
------------------------------------------------------------------------
       23.137   cycles for Siekmanski - 48 bytes SSSE3_32_24
------------------------------------------------------------------------
       23.126   cycles for Siekmanski - 48 bytes SSSE3_32_24
------------------------------------------------------------------------
       23.137   cycles for Siekmanski - 48 bytes SSSE3_32_24
------------------------------------------------------------------------
       23.137   cycles for Siekmanski - 48 bytes SSSE3_32_24
------------------------------------------------------------------------
       19.139   cycles for Siekmanski - 48 bytes SSSE3_32_24 unroled
------------------------------------------------------------------------
       19.130   cycles for Siekmanski - 48 bytes SSSE3_32_24 unroled
------------------------------------------------------------------------
       19.138   cycles for Siekmanski - 48 bytes SSSE3_32_24 unroled
------------------------------------------------------------------------
       19.138   cycles for Siekmanski - 48 bytes SSSE3_32_24 unroled
------------------------------------------------------------------------
Title: Re: Fastest way to move 3 bytes into a dword
Post by: KeepingRealBusy on January 21, 2013, 02:10:30 PM
I looked at some of this code and noticed some new instructions that I had never seen before, and then it dawned on me, I don't have the manuals for my new quad A8 3520M cpu, I only have the manuals for my old dual core system.

Does anyone have a good link to AMD to get the CORRECT manuals for this CPU?

Dave.
Title: Re: Fastest way to move 3 bytes into a dword
Post by: sinsi on January 21, 2013, 02:44:47 PM
Try here: http://developer.amd.com/resources/documentation-articles/developer-guides-manuals/
Title: Re: Fastest way to move 3 bytes into a dword
Post by: dedndave on January 21, 2013, 02:55:21 PM
i have a prescott that supports SSE3, Marinus
crashes at PSHUFB XMM0,XMM1   :P

87,665 cycles for the first test, though
Title: Re: Fastest way to move 3 bytes into a dword
Post by: KeepingRealBusy on January 21, 2013, 03:09:36 PM
Quote from: sinsi on January 21, 2013, 02:44:47 PM
Try here: http://developer.amd.com/resources/documentation-articles/developer-guides-manuals/

Sinsi,

Thank you, exactly what I was looking for.

Dave.
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 21, 2013, 03:37:13 PM
My results so far:
Quote
------------------------------------------------------------------------
Intel(R) Core(TM)2 CPU  6600  @ 2.40GHz

Instructions: MMX, SSE1, SSE2, SSE3, SSSE3
------------------------------------------------------------------------
       72,205   cycles for Dave - 48 bytes MOV 4 bytes / AND
       72,142   cycles for Dave - 48 bytes MOV 4 bytes / AND
       72,138   cycles for Dave - 48 bytes MOV 4 bytes / AND
------------------------------------------------------------------------
       34,586   cycles for Siekmanski - 48 bytes SSSE3_24_32
       34,571   cycles for Siekmanski - 48 bytes SSSE3_24_32
       34,575   cycles for Siekmanski - 48 bytes SSSE3_24_32
------------------------------------------------------------------------
       30,106   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
       30,067   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
       30,111   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
------------------------------------------------------------------------
       34,651   cycles for Siekmanski - 48 bytes SSSE3_32_24
       34,758   cycles for Siekmanski - 48 bytes SSSE3_32_24
       34,736   cycles for Siekmanski - 48 bytes SSSE3_32_24
------------------------------------------------------------------------
       28,661   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
       28,645   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
       28,637   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
------------------------------------------------------------------------

I slightly modified the source to avoid the too many separator lines.
Siekmanski is apparently new on the forum and not used to our "traditional"
way of doing tests.  :lol:

Compliments Siekmanski, you made a very good job.  :t

Waiting for our masters version, though.  :P

Attached the modified source.
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 21, 2013, 03:55:49 PM
Quote from: dedndave on January 21, 2013, 02:55:21 PM
i have a prescott that supports SSE3, Marinus
crashes at PSHUFB XMM0,XMM1   :P

87,665 cycles for the first test, though

You need an "S" more on your SSE level, otherwise no PSHUFB
for you.
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 22, 2013, 01:55:23 AM
Somebody messed up my string  :eusa_naughty:
Code Select

----------------------------------------------------------------------------
Intel(R) Core(TM)2 CPU          6600  @ 2.40GHz

Instructions: MMX, SSE1, SSE2, SSE3, SSSE3
----------------------------------------------------------------------------
       48,058   cycles for Dave - 48 bytes MOV 4 bytes / AND
       48,069   cycles for Dave - 48 bytes MOV 4 bytes / AND
       48,036   cycles for Dave - 48 bytes MOV 4 bytes / AND
----------------------------------------------------------------------------
       54,914   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
       58,549   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
       60,832   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
----------------------------------------------------------------------------
       55,680   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
       55,653   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
       55,674   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
----------------------------------------------------------------------------
       34,575   cycles for Siekmanski - 48 bytes SSSE3_24_32
       34,625   cycles for Siekmanski - 48 bytes SSSE3_24_32
       34,540   cycles for Siekmanski - 48 bytes SSSE3_24_32
----------------------------------------------------------------------------
       30,030   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
       30,036   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
       30,037   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
----------------------------------------------------------------------------
       34,534   cycles for Siekmanski - 48 bytes SSSE3_32_24
       34,533   cycles for Siekmanski - 48 bytes SSSE3_32_24
       34,575   cycles for Siekmanski - 48 bytes SSSE3_32_24
----------------------------------------------------------------------------
       28,543   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
       28,533   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
       28,538   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
----------------------------------------------------------------------------
Here it is a string with 48 characters inside me
Her itis  stingwit 48chaactrs nsie meeeearaitis  stingwit 48chaactrs nsie meeeeara  stingwit 48chaac
trs nsie meeeearaingwit 48chaactrs nsie meeeearait 48chaactrs nsie meeeeara8chaactrs nsie meeeearaac
trs nsie meeeearas nsie meeeearaie meeeearaeeeearaaracte


Was it mr. Siekmansky  ::)
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 22, 2013, 02:38:22 AM
Well Siekmansky, this report can help you find the bugs:
Code Select

----------------------------------------------------------------------------
Intel(R) Core(TM)2 CPU          6600  @ 2.40GHz

Instructions: MMX, SSE1, SSE2, SSE3, SSSE3
----------------------------------------------------------------------------
       48,125   cycles for Dave - 48 bytes MOV 4 bytes / AND
       48,093   cycles for Dave - 48 bytes MOV 4 bytes / AND
       48,180   cycles for Dave - 48 bytes MOV 4 bytes / AND
Destination data Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       54,307   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
       57,137   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
       57,148   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
Destination data Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       54,605   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
       44,780   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
       55,807   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
Destination data Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       34,573   cycles for Siekmanski - 48 bytes SSSE3_24_32
       34,576   cycles for Siekmanski - 48 bytes SSSE3_24_32
       34,565   cycles for Siekmanski - 48 bytes SSSE3_24_32
Destination data  strinaracte
----------------------------------------------------------------------------
       30,059   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
       30,071   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
       30,095   cycles for Siekmanski - 48 bytes SSSE3_24_32 unrolled
Destination data aracte
----------------------------------------------------------------------------
       34,573   cycles for Siekmanski - 48 bytes SSSE3_32_24
       34,611   cycles for Siekmanski - 48 bytes SSSE3_32_24
       34,572   cycles for Siekmanski - 48 bytes SSSE3_32_24
Destination data HHHHHHHHHHHHiiiiiiiiiiiiaaaaaaaaaaaaaaaaaraHHHHHHHHiiiiiiiiiiiiaaaaaaaaaaaaaaaaara
HHHHiiiiiiiiiiiiaaaaaaaaaaaaaaaaaraiiiiiiiiiiiiaaaaaaaaaaaaaaaaaraiiiiiiiiaaaaaaaaaaaaaaaaaraiiiiaaa
aaaaaaaaaaaaaaraaaaaaaaaaaaaaaaaaraaaaaaaaaaaaaaraaaaaaaaaaraaaaaaraaracte
----------------------------------------------------------------------------
       28,563   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
       28,565   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
       28,603   cycles for Siekmanski - 48 bytes SSSE3_32_24 unrolled
Destination data Her itis  stingwit 48chaactrs nsie meeeearaitis  stingwit 48chaactrs nsie meeeeara
  stingwit 48chaactrs nsie meeeearaingwit 48chaactrs nsie meeeearait 48chaactrs nsie meeeeara8chaact
rs nsie meeeearaactrs nsie meeeearas nsie meeeearaie meeeearaeeeearaaracte
----------------------------------------------------------------------------


Attached the program to test the correctness of destination data.
Title: Re: Fastest way to move 3 bytes into a dword
Post by: dedndave on January 22, 2013, 02:50:45 AM
Code Select
    mov     eax,[esi]
    and     eax,0FFFFFFh

AND is dependant on MOV being completed
it would help if you could put a "do something else" instruction in there
it would also help if 0FFFFFFh is in a register
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 22, 2013, 03:00:59 AM
Quote from: dedndave on January 22, 2013, 02:50:45 AM
Code Select
    mov     eax,[esi]
    and     eax,0FFFFFFh

AND is dependant on MOV being completed
it would help if you could put a "do something else" instruction in there
it would also help if 0FFFFFFh is in a register

Well, that's right, why don't you use the keyboard to do that?  :P
Title: Re: Fastest way to move 3 bytes into a dword
Post by: Siekmanski on January 22, 2013, 04:49:37 AM
It's not easy to understand someone elses code  :lol:
I made it work OK now and exchanged the in and output.
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 22, 2013, 05:16:53 AM
Quote from: Siekmanski on January 22, 2013, 04:49:37 AM
It's not easy to understand someone elses code  :lol:
I made it work OK now and exchanged the in and output.


Grab the code I posted in previous post and test your code, something is
still buggy, as it appears by the output.
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 22, 2013, 11:21:44 AM
Added my routine with XMM/SSE instructions. It requires a SSSE3
capable machine.
Quote
----------------------------------------------------------------------------
Intel(R) Core(TM)2 CPU          6600  @ 2.40GHz

Instructions: MMX, SSE1, SSE2, SSE3, SSSE3
----------------------------------------------------------------------------
       70,248   cycles for Dave - 48 bytes MOV 4 bytes / AND
       48,078   cycles for Dave - 48 bytes MOV 4 bytes / AND
       48,171   cycles for Dave - 48 bytes MOV 4 bytes / AND
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       50,346   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
       59,081   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
       56,535   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       47,055   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
       57,040   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
       57,420   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       25,662   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled
       38,100   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled
       38,110   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
----------------------------------------------------------------------------

I skipped Siekmansky code for the time being, waiting for his
tests with the attached program.

Frank
Title: Re: Fastest way to move 3 bytes into a dword
Post by: frktons on January 22, 2013, 01:10:50 PM
Code Select
And the last entry:
[quote]
----------------------------------------------------------------------------
Intel(R) Core(TM)2 CPU          6600  @ 2.40GHz

Instructions: MMX, SSE1, SSE2, SSE3, SSSE3
----------------------------------------------------------------------------
       50,769   cycles for Dave - 48 bytes MOV 4 bytes / AND
       51,653   cycles for Dave - 48 bytes MOV 4 bytes / AND
       52,787   cycles for Dave - 48 bytes MOV 4 bytes / AND
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       43,982   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
       43,484   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
       44,191   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       44,852   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
       48,817   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
       52,075   cycles for Dave - 48 bytes MOV 4 bytes / AND - Unrolled 2
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       30,254   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled
       25,419   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled
       29,016   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
       25,407   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled II
       28,221   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled II
       28,249   cycles for Frank - 48 bytes XMM/SSSE3 - Unrolled II
Destination data:Here it is a string with 48 characters inside me
----------------------------------------------------------------------------
----------------------------------------------------------------------------


Text only | Text with Images