SIMT challenge

[daydreamer]

I think I return to SIMT task I earlier solved with SIMD + LUT, and follow the student rules to solve it with SIMT
First a way to check with api or user enter number of cores on their computer
1: generate x million random numbers
Probably call rand in a loop
2: check if prime number
3: add those prime numbers together

Main start x number of threads,sending 1,2,3 to thread
On 16 cores cpu, 1,2,3 is choice depending on which is most computing expensive and least computing expensive
Each thread starts with Switch/case
Case 1
Invoke random
Case 2
Invoke primes
Case 3
Invoke sum

Option is return when finished task 1 and 2 and thread joins helping other threads that aren't finished yet

Other ideas is
Two threads random numbers are 1 thread starts from beginning of array and second thread starts from end of array, stops when they reach same adress

[NoCforMe]

I don't get it; what is the point of this whole exercise?

[daydreamer]

When even old computers have 4 or more cores, it's easier with several threads coding standard x86 gp reg code than exotic SIMD opcodes and more flexible able to run many different Algos /proc on each core than 128 bit SSE 4 x32 bit of same algo

I have 4 cores on my newest and want try learn use all 4

[Siekmanski]

I don't get it; what is the point of this whole exercise?

A challenge to excel in programming by using your logical and analytical part of your brain. 

[zedd151]

All I see here is a word salad. 

And only a 'concept' of an 'idea'.   

[daydreamer]

I have no scalar code prime testing code yet,sieve and divide, I only have experimental SSE : real4's using many divps using 2.0 to around 300.0 with a bug :
its not capable of detect all those lower 2.0-300.0 primes and limited to REAL4 range of ca 9 million

 

[sinsi]

Some older (but still recent) CPUs would reduce the clock if full AVX was used on multi cores because of the heat.
With Intel CPUs you also need to pick a CPU that's not hyper-threaded.

Multithreading isn't easy, that's why most apps don't use it.
Also, CPU multithreading is a whole lot different to using an OS's thread functions.

[daydreamer]

so far:
scalar SSE compare floats
after that
packed SSE compare floats

Code Select Expand

// primesx.cpp

#include "pch.h"
#include <iostream>
using namespace std;
float zero = 0.0;
int pflag = 0;
alignas(16) float flut[]{ 2.0,3.0,5.0,7.0,11.0,13.0,17.0,19.0,0.0,0.0,0.0 };
alignas(16) float arr[]{ 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 };
char lut[]{ 0, 0, 2, 3, 0, 5, 0, 7, 0, 0,0,11,0,13,0,0,0,0,0,0 };
int main()
{
int i, j=3;
float f = 3.0;
float fresult = 0;
    cout << "Primesx\n";
for (i = 0; i < 14; i++) {
cout << i << " ";
if (lut[i] != 0) cout << "prime " << (int)lut[i] << " ";
cout << f << " ";
//f = f + 1.0;
}
cout << f << " ";
cout << "\n\n\n\n";
for (j = 1; j < 14; j=j+1) {
f = (float)j; //test all floats
_asm {
push ebx

mov ecx, 14
mov ebx, 1
; lea ebx, [ebx * 4]
lea ebx, [flut + ebx * 4]
; lea ebx, flut
L2 :
movss xmm0, f
movss xmm1, [ebx]
; subss xmm0, xmm1
ucomiss xmm0, xmm1
jne L1; found prime
ja L4s;ja jump if above,jb jump if below
mov eax, 1
mov pflag, eax
movss xmm0, [ebx]
movss fresult, xmm0
jmp l3;have found prime, jump out of loop
L4s:
mov eax,0ffffffffh
jmp L3
L1 :
xorps xmm0, xmm0
mov eax,0
mov pflag,eax
; movss fresult, xmm0
add ebx, 4
dec ecx
jne L2
L3:
pop ebx


}
fresult = fresult * pflag;
cout << fresult << " ";
}//j
cout << "\n";
for (j = 2; j < 20; j++) {
f = (float)j;

_asm {
push ebx
lea ebx, flut
movss xmm0, f
shufps xmm0, xmm0, 0
movups xmm1, [ebx]
movaps xmm3, xmm1
CMPEQPS xmm1, xmm0
pand xmm1, xmm3
movaps xmm7, xmm1
add ebx, 16
movups xmm1, [ebx]
movaps xmm3, xmm1
CMPEQPS xmm1, xmm0
pand xmm1, xmm3
por xmm1, xmm7
movups arr, xmm1
haddps xmm1, xmm1
haddps xmm1, xmm1
movss fresult, xmm1


pop ebx
}
cout << "xmm reg float 0,float 1,float 2,float 3 : "<< arr[0] << " " << arr[1] << " " << arr[2] << " " << arr[3] << "\n";
cout << fresult << " zero = non prime\n";

}//second time j


}