OpenCL: internal implementation tests in MQL5 - page 45

[Andrey Dik]

MetaDriver:

.........
"      out[i]=0;         \r\n" // после десятитысячного шага итерации этот оператор начнёт выполняться в каждом цикле.
"   }                    \r\n" // т.е. на выходе из цикла по любому будем иметь out[i] = 0;
......... 

Why?

[Vladimir Gomonov]

joo:
Why?

out[i]+=u;       // если к неотрицательному числу прибавить число u большее 10000
if(out[i]>10000)  // то результат будет больше 10000 (true)
  out[i]=0;      // тогда ... = 0

[Andrey Dik]

И?...

I can't keep up with your thinking, I'm sorry.

[Vladimir Gomonov]

joo:

И?...

И... then...

"      out[i]=0;         \r\n" // после десятитысячного шага итерации этот оператор начнёт выполняться в каждом цикле.
"   }                    \r\n" // т.е. на выходе из цикла по любому будем иметь out[i] = 0;

[Sceptic Philozoff]

Andrey, it's simple: out[ i ] cannot be higher than 10000, and as soon as it is higher, the array value is zeroed - by convention. Starting from u=10001, it will always be zeroed.

From here on, we will start adding even larger values of u to zero. But not too big, no more than 4 billion for sure. Hence, the value of the array is positive and exactly greater than 10000. Therefore, it is zero again. And so on. And then, outside the loop, we add i to zero.

[Andrey Dik]

This is understandable. What's not clear is why my Out[] array will contain the same thing as Vladimir's array:

"__kernel void MFractal(                                    \r\n"
"                       __global int *out                   \r\n"
"                      )                                    \r\n"
"  {                                                        \r\n"
"   out[get_global_id(0)]= get_global_id(0);                \r\n"
"  }                                                        \r\n";

[Sceptic Philozoff]

joo: It's clear. What's not clear is why my array Out[] will contain the same value as Vladimir's array:

Because outside the loop, when the array value is already zero, i = get_global_id(0) is added to it.

Well go ahead and print out all the values.

[Sergey]

I've almost got the sli of the 590 together when's the next test?

[Vladimir Gomonov]

casinonsk:
I've almost got the sli of the 590 together, when's the next test?

Now it's your turn to write something. )

[Sceptic Philozoff]

casinonsk:
I've almost got the sli of the 590 together when is the next test going to take place?

Run the early tests on this sli of yours and post it here. Some fun at least...