+ store also ecx and edx after cpuid call with eax=7

  + test VNNI and BIGALG instructions

rtl/i386/cpu.pp

@@ -253,6 +253,8 @@ unit cpu;
                     movl $0,%ecx
                     cpuid
                     movl %ebx,_ebx
+                    movl %ecx,_ecx
+                    movl %edx,_edx
                     popl %ebx
                   end;
                   _AVX2Support:=_AVXSupport and ((_ebx and $20)<>0);

rtl/x86_64/cpu.pp

@@ -228,6 +228,8 @@ unit cpu;
               movl $0,%ecx
               cpuid
               movl %ebx,_ebx
+              movl %ecx,_ecx
+              movl %edx,_edx
             end ['rax','rbx','rcx','rdx'];
             _AVX2Support:=_AVXSupport and ((_ebx and $20)<>0);
             _AVX512FSupport:=(_ebx and $10000)<>0;

tests/test/units/cpu/tcpu1.pp

@@ -40,8 +40,8 @@ begin
     begin
       writeln('yes');
       asm
-       // vpxor %ymm0,%ymm0,%ymm0
-       // vaddpd %zmm0,%zmm0,%zmm0
+       vpxor %ymm0,%ymm0,%ymm0
+       vpdpbusd %zmm0,%zmm0,%zmm0
       end;
     end
   else
@@ -51,8 +51,8 @@ begin
     begin
       writeln('yes');
       asm
-        //vpxor %ymm0,%ymm0,%ymm0
-        //vaddpd %zmm0,%zmm0,%zmm0
+        vpxor %ymm0,%ymm0,%ymm0
+        vpopcntb %zmm0,%zmm0
       end;
     end
   else