short to int data re-alignment

Question

0.00/5 (No votes)

See more:

how do a load b1 as an array of 4 rows?

b1 = -114 -113 -113 -112 row 0
-93 -95 -90 -92
-25 -26 -23 -25
36 35 35 38 row 3
in my codes f0 is loaded with 8 elements (column 1 and 2 of b1)
and f1 is loaded with 8 elements as well(columns 3 and 4).what i want is to change the loading to 4 elements according to rows.something like:
f0 =114 -113 -113 -112
f1 =-93 -95 -90 -92
f2 = -25 -26 -23 -25
f3 =36 35 35 38
how do i achieve this given my snippet codes

C++

void tom::add(void* btr)
{

__declspec(align(16))short* b =(short*)btr;

  _declspec(align(16))__m128i* b1=(__m128i*)b;
  __m128i f0,f1,f2,f3,temp;
            temp = (_mm_set_epi32(0,0,0,0));

    f0 = _mm_load_si128(b1);//-114,-93,-25,36,-113,-95,-26,35
    f1 = _mm_load_si128(b1+1);//-113,-90,-23,35,-112,-92,-25,38<</xml>

Posted 16-Aug-10 0:31am

SMART LUBOBYA

Updated 16-Aug-10 0:35am

v2

Add a Solution

1 solution

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Eugen Podsypalnikov · Accepted Answer · 2010-08-16T01:37:00

Solution 1

May be a variant :) :

void tom::add(short* pshData, int iCount)
{
  if (4 * 4 <= iCount) {
    int* piData = new int[iCount];
    while (iCount--) {
      piData[iCount] = pshData[iCount];
    }

    _m128i* p128Data = (_m128i*) piData;
    _m128i f0 = _mm_load_si128(p128Data++);
    _m128i f1 = _mm_load_si128(p128Data++);
    _m128i f2 = _mm_load_si128(p128Data++);
    _m128i f3 = _mm_load_si128(p128Data);

    //..

    delete[] piData;
  }
}

Posted 16-Aug-10 1:37am

Eugen Podsypalnikov

Updated 16-Aug-10 1:56am

v2

Comments

SMART LUBOBYA 16-Aug-10 9:07am

i did this but i get error C2661, see codes below
void tom::add (short* ptr,int icount)
{
if (4*4<=icount){
int*block2 = new int[icount];
for(int i = 0; i < icount; i++)
{
block2[i]=ptr[i];
}
__m128i* block3=(__m128i*)block2;
__m128i f0 = *block3++;
__m128i f1 = *block3++;
__m128i f2 = *block3++;
__m128i f3 = *block3;
__m128i s0 = _mm_add_epi16(f0,f3);
__m128i s3 = _mm_sub_epi16(f0,f3);

delete[]block2;
}
}
void tom::add (void* pIn, void* pCoeff)
{

memcpy(pCoeff, pIn, sizeof(short) * 16);
add(pCoeff); //tom.cpp(280) : error C2661: ' tom::add' : no overloaded function takes 1 arguments
}

Eugen Podsypalnikov 16-Aug-10 9:11am

Try to extend your call: add(pCoeff, 16); :)

SMART LUBOBYA 16-Aug-10 11:35am

is there a way of achieving the same result changing the declaration of add from:
void tom::add(void* btr) to void tom::add (short* ptr,int icount)
{
i have several files connected to this .cpp file. kindly help.

SMART LUBOBYA 16-Aug-10 11:36am

is there a way of achieving the same result without changing the declaration of add from:
void tom::add(void* btr) to void tom::add (short* ptr,int icount)
{
i have several files connected to this .cpp file. kindly help

Eugen Podsypalnikov 16-Aug-10 14:13pm

Yes, there is...
...but the function will not be "safe"
(it will access up to short[15] without any knowledge, how many shorts were passed :)) :
void tom::add(void* pData)
{
short* pshData = (short*) pData;
int* piData = new int[16];
for (int i = 0; i < 16; i++) {
piData[i] = pshData[i]; // possible access outside of the bounderies (!)
}

//...

delete[] piData;
}

SMART LUBOBYA 17-Aug-10 4:17am

Reason for my vote of 5
now i know how to convert short arrays(8 elements) to int arrays(4 elements)

SMART LUBOBYA 17-Aug-10 4:48am

thanks for your advice. when manupulating fo-f3, the data goes back to short of 8 element; see s0 and s3. how can i main data to 4 elements?

void tom::add(void* btr)
{

short* b =(short*)btr;

int *B1 = new int[16];
for( j = 0; j < 16;j++)
{
B1[j]=b[j];
}
__m128i* B2=(__m128i*)B1;
__m128i f0 = _mm_load_si128(B2++);// -114,-93,-25,36
__m128i f1 = _mm_load_si128(B2++);//-113,-95,-26,35
__m128i f2 = _mm_load_si128(B2++);//-113,-90,-23,35
__m128i f3 = _mm_load_si128(B2);// -112,-92,-25,38

__m128i s0 = _mm_add_epi16(f0,f3);//-226,-2,-185,-2,-50,-2,74
__m128i s3 = _mm_sub_epi16(f0,f3);//-2,0,-1,0,0,0,-2,0

SMART LUBOBYA 17-Aug-10 4:50am

how can maintain data ie s0, s3 to 4 elements?

Eugen Podsypalnikov 17-Aug-10 4:57am

I can see the suffix "16" in your manipulations... But it would be only a half of an int (16 = short length). You should find something with the suffix "32" (int length, 128 = 32 * 4) :)

SMART LUBOBYA 17-Aug-10 7:45am

its done, thank very much again

SMART LUBOBYA 3-Sep-10 13:23pm

how do i align B1 or btr to 16 byte ?

Eugen Podsypalnikov 3-Sep-10 16:47pm

Try it please :) :
void tom::add(void* pData /*assumed there are 16 shorts inside*/)
{
short* pShortData = (short*) pData;

BYTE pbyDump = new BYTE[16 * 16]; // 16 shorts, each on 16 bytes
memset(pbyDump, 0, 16 * 16); // init the dump by nulls
for (int i = 0; i < 16; i++) { // for each short
memcpy(pbyDump + (i * 16), // offset for the short destination
&pShortData[i], // source short element
sizeof(short)); // length to copy, other bytes stay by nulls
}

// do something with the dump
//..

delete[] pbyDump;
}