BASM for beginners, lesson 5

Welcome to lesson number 5. Today’s topic is loops. We will take a look on how the compiler implements for loops, and which optimizations it applies on them. We will evaluate the efficiency of these optimizations.

 function ForLoop(Start, Stop : Integer) : Integer;
 var
 I : Integer;

 begin
 Result := 0;
 for I := Start to Stop do
 begin
 Result := Result + I;
 end;
 end;

 This example function does nothing useful except giving us an example for loop to examine. Let us see what the compiler translates the function into. In this lesson we try something new and compile with optimizations off.

 function ForLoopNonOpt(Start, Stop : Integer) : Integer;
 var
 I : Integer;

 begin
 {
 push ebp
 mov ebp,esp
 add esp,-$14
 mov [ebp-$08],edx
 mov [ebp-$04],eax
 }
 Result := 0;
 {
 xor eax,eax
 mov [ebp-$0c],eax
 }
 for I := Start to Stop do
 {
 mov eax,[ebp-$04]
 mov edx,[ebp-$08]
 sub edx,eax
 jl +$15
 inc edx
 mov [ebp-$14],edx
 mov [ebp-$10],eax
 }
 begin
 Result := Result + I;
 {
 mov eax,[ebp-$10]
 add [ebp-$0c],eax
 inc dword ptr [ebp-$10]
 }
 end;
 {
 dec dword ptr [ebp-$14]
 jnz -$0e
 mov eax,[ebp-$0c]
 }
 {
 mov esp,ebp
 pop ebp
 ret
 }
 end;

 It is seen that the compiler generates a lot of code when no or few optimizations are applied. As usual the first 3 lines set up a stackframe. This time it is 20 byte big (16 hex). The two next lines copy the Start and Stop variables unto the stack. Start is transferred in eax and Stop is transferred in edx to the function due to the register calling convention. The next two lines creates a zero in eax and copy it to the stackframe at [ebp-$0c], which is the place of the result variable. Then we are ready to enter the for-loop. Before entering a for-loop it is necessary to test whether the loop will iterate at all. If Stop is bigger than Start this is the case. Start and Stop is fetched from their stack locations into eax & edx. We compute Stop-Start and if this is negative the loop should not be entered and execution is transferred past the end of the loop by the jl (jump low) instruction. The next line increments Stop and then it is copied to location [ebp-$14]. We have no name for the local variable at this location. The purpose of it also needs some further explanations. It is a variable introduced by the compiler as an optimization and this is weird because we compiled with optimizations off. There is only one more line accessing the NoName variable and this is the dec dword ptr [ebp-$14] line. This line decrement NoName by one at the end of each loop-iteration and the loop closing test is a test that it has not reached zero. The dec instruction sets the flags and the jnz jumps back to the top of the loop if NoName <> 0. We would expect that I was used as the loop counter and that it was running from Start to Stop. It is in fact doing so, but it is not used for control of the loop. This is the sole purpose of NoName. The advantage of this is that it saves an instruction for comparing I to Stop. There is also a cost and this is the dec NoName instruction. On P4 the latency/throughput of cmp is 0.5/0.5 clock cycles and for dec they are 1/0.5. Therefore we would expect this optimization to be a "deoptimization". The latency and throughput numbers for P4 is found in the "Intel Pentium 4 and Xeon Processor Optimization" manual from Intel.
 Back to the mov [ebp-$10],eax line. It copies I to the stack. The loop body consist of only one line of Pascal Result := Result + I;. This is translated into 3 lines of ASM. The first two lines are loading I into eax and then adding it to Result at the stack at [ebp-$0c]. The third line increment I. This ended the explanation of the loop code and only two things are left. The Result must be copied to eax which is the register to use for returning the result from a register calling convention function. The last three lines remove the stackframe and return execution to the line after the line that called the function.
 As an exercise let us change the ASM code such that it follows the Pascal code and our understanding of a for-loop. We start by turning the function into a pure BASM one. This is done by “outcommenting” the Pascal code and "incommenting" the ASM code. Defining the two labels LoopEnd and LoopStart is also necessary. The two jumps are edited such that they jump to the labels.

 function ForLoopBASM1(Start, Stop : Integer) : Integer;
 asm
 push ebp
 mov ebp,esp
 add esp,-$14
 mov [ebp-$08],edx
 mov [ebp-$04],eax
 //Result := 0;
 xor eax,eax
 mov [ebp-$0c],eax
 //for I := Start to Stop do
 mov eax,[ebp-$04]
 mov edx,[ebp-$08]
 sub edx,eax
 jl @LoopEnd
 inc edx
 mov [ebp-$14],edx
 mov [ebp-$10],eax
 //begin
 @LoopStart :
 //Result := Result + I;
 mov eax,[ebp-$10]
 add [ebp-$0c],eax
 inc dword ptr [ebp-$10]
 //end;
 dec dword ptr [ebp-$14]
 jnz @LoopStart
 @LoopEnd :
 mov eax,[ebp-$0c]
 mov esp,ebp
 pop ebp
 //ret
 end;

 The first thing we do is removing the NoName variable.

 function ForLoopBASM2(Start, Stop : Integer) : Integer;
 asm
 push ebp
 push ebx //New
 mov ebp,esp
 add esp,-$14
 mov [ebp-$08],edx
 mov [ebp-$04],eax
 //Result := 0;
 xor eax,eax
 mov [ebp-$0c],eax
 //for I := Start to Stop do
 mov eax,[ebp-$04]
 mov edx,[ebp-$08]
 sub edx,eax
 jl @LoopEnd
 //inc edx //NoName intialize
 //mov [ebp-$14],edx //NoName intialize
 mov [ebp-$10],eax
 //begin
 @LoopStart :
 //Result := Result + I;
 mov eax,[ebp-$10]
 add [ebp-$0c],eax
 inc dword ptr [ebp-$10]
 //end;
 //dec dword ptr [ebp-$14] //NoName decrement
 mov ebx, [ebp-$10] //New
 mov ecx, [ebp-$08] //New
 cmp ebx, ecx //New
 //jnz @LoopStart
 jbe @LoopStart //New
 @LoopEnd :
 mov eax,[ebp-$0c]
 mov esp,ebp
 pop ebx //New
 pop ebp
 //ret
 end;

 The lines marked "New" are introduced to make I the loop control variable. The mov ebx, [ebp-$10] line copies I into ebx The next line copies Stop into ecx. Then the line cmp ebx, ecx compare them and jbe @LoopStart transfer execution to the start of the loop if I is below Stop or equal to it. Because we use ebx and it is not free for use we remember to push and pop it.

 We expect a for-loop to evaluate the loop condition at the top. This test is split in two by the compiler implementation. Before entering the loop it is tested that it will execute at least once and the actual loop closing test is done at the bottom. This is an optimization technique called loop inversion. Now we change the loop such that this optimization is removed. Then we see what the benefit of the optimization was.

 function ForLoopBASM4(Start, Stop : Integer) : Integer;
 asm
 push ebp
 push ebx
 mov ebp,esp
 add esp,-$14
 mov [ebp-$08],edx
 mov [ebp-$04],eax
 //Result := 0;
 xor eax,eax
 mov [ebp-$0c],eax
 //for I := Start to Stop do
 mov eax,[ebp-$04]
 mov edx,[ebp-$08]
 //sub edx,eax
 //jl @LoopEnd
 mov [ebp-$10],eax
 //begin
 @LoopStart :
 mov ebx, [ebp-$10]
 mov ecx, [ebp-$08]
 cmp ebx, ecx
 ja @LoopEnd
 //Result := Result + I;
 mov eax,[ebp-$10]
 add [ebp-$0c],eax
 inc dword ptr [ebp-$10]
 //end;
 //mov ebx, [ebp-$10]
 //mov ecx, [ebp-$08]
 //cmp ebx, ecx
 //jbe @LoopStart
 jmp @LoopStart
 @LoopEnd :
 mov eax,[ebp-$0c]
 mov esp,ebp
 pop ebx
 pop ebp
 end;

 The loop closing test has been moved to the top and the test has been inverted. At the place of the test there now is an unconditional jump to the top. This jump is what the loop inversion optimization is all about. In the nonoptimized for loop there are two jumps and only one in the optimized one. The test at the top that tested whether Start was above Stop is now redundant and is removed. Before making any timing to evaluate the effect of the two optimizations it is a good idea to optimize away some or all if possible, of the stack to register and register to stack moves. This process is called register allocation and is one of the most important optimizations on all architectures, but it is even more important on the Intel architecture because of the low number of available registers. If there is not a register available to all variables it is crucial which variables get a register. The mov instructions inside the loop body are the most important ones to get rid of. They are executed as many times as the number of loop iterations. The instructions outside of the loop are only executed once. The variables used inside the loop should be allocated to registers first. This is I, Stop and Result. At this point we could be smart and take a look at the use of registers as temporaries. If a variable is always copied into the same temp register it would be smart to allocate this register for the variable. Stop is in the edx register when we enter the function and this register is also used as temporary register for it in all but two lines. This is the two lines of the loop test that we added. Let us change

 mov ecx, [ebp-$08]
 cmp ebx, ecx

 to

 mov edx, [ebp-$08]
 cmp ebx, edx

 Eax is used by Start in the top of the function and by Result in the rest of the function. If there is no overlap in usage we can allocate eax for Result as soon as Start has finished using it. After Start is assigned to I (mov [ebp-$10],eax) it is not used any more and eax is free to use by Result, if it was not for those lines where eax is used as temp for I.

 mov eax,[ebp-$10]
 add [ebp-$0c],eax
 inc dword ptr [ebp-$10]

 After ecx got out of use by the last change we can use that as temp for I instead of eax.

 mov ecx,[ebp-$10]
 add [ebp-$0c],ecx
 inc dword ptr [ebp-$10]

 The summary of the first part of the register allocation is: Result in eax, I in ecx and Stop in edx.

 Let us first change the lines with Stop. [ebp-$08] is replaced by edx everywhere.

 function ForLoopBASM6(Start, Stop : Integer) : Integer;
 asm
 push ebp
 push ebx
 mov ebp,esp
 add esp,-$14
 //mov [ebp-$08],edx
 mov edx,edx
 mov [ebp-$04],eax
 //Result := 0;
 xor eax,eax
 mov [ebp-$0c],eax
 //for I := Start to Stop do
 mov eax,[ebp-$04]
 //mov edx,[ebp-$08]
 mov edx,edx
 mov [ebp-$10],eax
 //begin
 @LoopStart :
 mov ebx, [ebp-$10]
 //mov edx, [ebp-$08]
 mov edx, edx
 cmp ebx, edx
 ja @LoopEnd
 //Result := Result + I;
 mov ecx,[ebp-$10]
 add [ebp-$0c],ecx
 inc dword ptr [ebp-$10]
 //end;
 jmp @LoopStart
 @LoopEnd :
 mov eax,[ebp-$0c]
 mov esp,ebp
 pop ebx
 pop ebp
 end;

 Then allocate ecx for I by replacing [ebp-$10] by ecx.

 function ForLoopBASM7(Start, Stop : Integer) : Integer;
 asm
 push ebp
 push ebx
 mov ebp,esp
 add esp,-$14
 mov edx,edx
 mov [ebp-$04],eax
 //Result := 0;
 xor eax,eax
 mov [ebp-$0c],eax
 //for I := Start to Stop do
 mov eax,[ebp-$04]
 mov edx,edx
 //mov [ebp-$10],eax
 mov ecx,eax
 //begin
 @LoopStart :
 //mov ebx, [ebp-$10]
 mov ebx, ecx
 mov edx, edx
 cmp ebx, edx
 ja @LoopEnd
 //Result := Result + I;
 //mov ecx,[ebp-$10]
 mov ecx,ecx
 add [ebp-$0c],ecx
 //inc dword ptr [ebp-$10]
 inc ecx
 //end;
 jmp @LoopStart
 @LoopEnd :
 mov eax,[ebp-$0c]
 mov esp,ebp
 pop ebx
 pop ebp
 end;

 And last eax is allocated for use by Result. Because eax is used in the top of the function by Start and as temp in the lines that initializes Result to zero, we need to add an extra line to copy Result into eax after eax is not in use anymore for these other purposes.

 function ForLoopBASM8(Start, Stop : Integer) : Integer;
 asm
 push ebp
 push ebx
 mov ebp,esp
 add esp,-$14
 mov edx,edx
 mov [ebp-$04],eax
 //Result := 0;
 xor eax,eax
 mov [ebp-$0c],eax
 //for I := Start to Stop do
 mov eax,[ebp-$04]
 mov edx,edx
 mov ecx,eax
 mov eax, [ebp-$0c] //New
 //begin
 @LoopStart :
 mov ebx, ecx
 mov edx, edx
 cmp ebx, edx
 ja @LoopEnd
 //Result := Result + I;
 mov ecx,ecx
 //add [ebp-$0c],ecx
 add eax,ecx
 inc ecx
 //end;
 jmp @LoopStart
 @LoopEnd :
 //mov eax,[ebp-$0c]
 mov eax,eax
 mov esp,ebp
 pop ebx
 pop ebp
 end;

 Because we were pretty smart when we chose the registers there is a lot of lines like mov eax,eax. It is easy to see how redundant they are ;-). Let us remove them.

 function ForLoopBASM9(Start, Stop : Integer) : Integer;
 asm
 push ebp
 push ebx
 mov ebp,esp
 add esp,-$14
 //mov edx,edx
 mov [ebp-$04],eax
 //Result := 0;
 xor eax,eax
 mov [ebp-$0c],eax
 //for I := Start to Stop do
 mov eax,[ebp-$04]
 //mov edx,edx
 mov ecx,eax
 mov eax, [ebp-$0c]
 //begin
 @LoopStart :
 mov ebx, ecx
 //mov edx, edx
 cmp ebx, edx
 ja @LoopEnd
 //Result := Result + I;
 //mov ecx,ecx
 add eax,ecx
 inc ecx
 //end;
 jmp @LoopStart
 @LoopEnd :
 //mov eax,eax
 mov esp,ebp
 pop ebx
 pop ebp
 end;

 When optimizing ASM code there is generally two lines of thinking we can choose to follow. We can think as the human beings we are and try to be smart, use whatever information we can get from the code. We did some of this when we selected the registers here. Another line of thought is trying to do it systematically as an optimizer/compiler has to. This way we develop algorithms that can be coded in a tool. This tool can later take over the most boring of optimizations, these we do over and over again. Removal of the most obviously redundant line of code of all, the mov eax,eax, was an example of dead code removal, which is basic to any optimizer.
 In the top of the function we still have some references to the stack. To get rid of these we allocate registers for those variables too. This time we just pick edi and esi, which are not used elsewhere. Allocate esi for [ebp-$04] and edi for [ebp-$0c]. Because esi and edi must be preserved by the function we must push and pop them.

 function ForLoopBASM10(Start, Stop : Integer) : Integer;
 asm
 push ebp
 push ebx
 push esi
 push edi
 mov ebp,esp
 add esp,-$14
 //mov [ebp-$04],eax
 mov esi,eax
 //Result := 0;
 xor eax,eax
 //mov [ebp-$0c],eax
 mov edi,eax
 //for I := Start to Stop do
 //mov eax,[ebp-$04]
 mov eax,esi
 mov ecx,eax
 //mov eax, [ebp-$0c]
 mov eax, edi
 //begin
 @LoopStart :
 mov ebx, ecx
 cmp ebx, edx
 ja @LoopEnd
 //Result := Result + I;
 add eax,ecx
 inc ecx
 //end;
 jmp @LoopStart
 @LoopEnd :
 mov esp,ebp
 pop edi
 pop esi
 pop ebx
 pop ebp
 end;

 The stack frame is not used anymore and there is no need to set it up. This removes 4 instructions. Then we observe that the two lines

 mov eax,esi
 mov ecx,eax

 This can be replaced by one line.

 mov ecx, esi

 This is an example of a simple copy propagation followed by dead code removal. The value in eax is not used by any other lines than the next line that copies it back to ecx. It is in fact immediately overwritten by the line mov eax, edi. Therefore we can replace the second line by mov ecx, esi and remove the first one, which becomes dead.

 function ForLoopBASM11(Start, Stop : Integer) : Integer;
 asm
 //push ebp
 push ebx
 push esi
 push edi
 //mov ebp,esp
 //add esp,-$14
 mov esi,eax
 //Result := 0;
 xor eax,eax
 mov edi,eax
 //for I := Start to Stop do
 //mov eax,esi
 //mov ecx,eax
 mov ecx, esi
 mov eax, edi
 //begin
 @LoopStart :
 //mov ebx, ecx
 //cmp ebx, edx
 cmp ecx, edx
 ja @LoopEnd
 //Result := Result + I;
 add eax,ecx
 inc ecx
 //end;
 jmp @LoopStart
 @LoopEnd :
 //mov esp,ebp
 pop edi
 pop esi
 pop ebx
 //pop ebp
 end;

 The line xor eax,eax that initializes Result to zero can together with the line rigth after it be moved some lines down closer to the place where eax is used for the first time. It should not be moved into the loop that would change the logic of the function, but just before the line before loopStart. This removes the need for copying eax into edi and back into eax again in the line just before the comment line //for I := Start to Stop do, and in the line before the outcommented begin.


 function ForLoopBASM12(Start, Stop : Integer) : Integer;
 asm
 push ebx
 push esi
 push edi
 mov esi,eax
 //for I := Start to Stop do
 mov ecx, esi
 //Result := 0;
 xor eax,eax
 //mov edi,eax
 //mov eax, edi
 //begin
 @LoopStart :
 cmp ecx, edx
 ja @LoopEnd
 //Result := Result + I;
 add eax,ecx
 inc ecx
 //end;
 jmp @LoopStart
 @LoopEnd :
 pop edi
 pop esi
 pop ebx
 end;

 After having cleaned up we spot two lines of mov which together copy eax to ecx via esi. This leaves a copy of eax in esi which is not used. Therefore these two lines can be replaced by one that moves eax directly into ecx. This is also copy propagation + dead code removal.

 function ForLoopBASM13(Start, Stop : Integer) : Integer;
 asm
 push ebx
 //push esi
 push edi
 //mov esi,eax
 //for I := Start to Stop do
 //mov ecx, esi
 mov ecx, eax
 //Result := 0;
 xor eax,eax
 //begin
 @LoopStart :
 cmp ecx, edx
 ja @LoopEnd
 //Result := Result + I;
 add eax,ecx
 inc ecx
 //end;
 jmp @LoopStart
 @LoopEnd :
 pop edi
 //pop esi
 pop ebx
 end;

 After having removed the only use of esi there is no need to push and pop it.

 function ForLoopBASM14(Start, Stop : Integer) : Integer;
 asm
 //push ebx
 //push edi
 //for I := Start to Stop do
 mov ecx, eax
 //Result := 0;
 xor eax,eax
 //begin
 @LoopStart :
 cmp ecx, edx
 ja @LoopEnd
 //Result := Result + I;
 add eax,ecx
 inc ecx
 //end;
 jmp @LoopStart
 @LoopEnd :
 //pop edi
 //pop ebx
 end;

 We also, a little late perhaps, observe that ebx and edi is not used either. After having cleaned up and relocated the comments a little, there is a nice clean function as a result.

 function ForLoopBASM15(Start, Stop : Integer) : Integer;
 asm
 mov ecx, eax
 //Result := 0;
 xor eax,eax
 //for I := Start to Stop do
 @LoopStart :
 cmp ecx, edx
 ja @LoopEnd
 //Result := Result + I;
 add eax,ecx
 inc ecx
 jmp @LoopStart
 @LoopEnd :
 end;

 It took a long time and a lot of optimizations to get here because we started with the non optimized output from the compiler. This long process illustrated the amount of work the compiler leaves to the optimizer. Sometimes we did not use the algorithmic approach to optimization, but we could have achieved the same result by doing it.

 Instead of taking the same long road again with the function with loop inversion present, we can cheat and compile the Pascal function with optimizations on. The compiler might do all the optimization we did.

 function ForLoopOpt(Start, Stop : Integer) : Integer;
 var
 I : Integer;

 begin
 {
 }
 Result := 0;
 {
 xor ecx,ecx
 }
 for I := Start to Stop do
 {
 sub edx,eax
 jl +$08
 inc edx
 xchg eax,edx
 }
 begin
 Result := Result + I;
 {
 add ecx,edx
 inc edx
 }
 end;
 {
 dec eax
 jnz -$06
 }
 {
 mov eax,ecx
 }
 end;

 This time Delphi did a really nice job. Only two lines jump into our eyes as possibly redundant. xchg eax,edx simply exchange the values in eax and edx and mov eax,ecx copy the Result into eax. Both lines are outside the loop and make little harm. We now have two functions - one with no loop optimizations and one with two. To make things complete we need two more functions, one with loop inversion only and one with the NoName variable optimization only. In the begining of the lesson we saw how to remove the two optimizations and this is what I have done to get to the last 2 functions. In the Delphi optimized function above, I optimized away the xchg instruction by swapping the use of the two registers it exchanged.

 Because we want to se the maximum effect of the loop optimizations I have removed the loop body code doing the operation Result := Result + I;

 Here are the four final functions

 function ForLoopNoLoopInverNoNoName(Start, Stop : Integer) : Integer;
 asm
 mov ecx, eax
 //Result := 0;
 xor eax,eax
 //for I := Start to Stop do
 @LoopStart :
 cmp ecx, edx
 ja @LoopEnd
 inc ecx
 jmp @LoopStart
 @LoopEnd :
 end;

 The loop consists of 4 instructions. 1 cmp, 1 ja, 1 inc and 1 jmp. Latency and throughput for these instructions on P4 are: cmp 0.5/0.5, ja X/0.5, inc 0.5/1 and jmp X/0.5 The X means "latency is not applicable to this instruction". Adding latencies we get: 0.5 + X + 0.5 + X = ? clock cycles

 function ForLoopLoopInverNoNoName(Start, Stop : Integer) : Integer;
 asm
 mov ecx, eax
 //Result := 0;
 xor eax,eax
 //for I := Start to Stop do
 cmp ecx, edx
 ja @LoopEnd
 @LoopStart :
 inc ecx
 cmp ecx, edx
 jbe @LoopStart
 @LoopEnd :
 end;

 This loop consists of 3 instructions also with unknown sum of latency.

 function ForLoopNoLoopInverNoName(Start, Stop : Integer) : Integer;
 asm
 //Result := 0;
 xor ecx,ecx
 //for I := Start to Stop do
 sub edx,eax
 cmp edx, 0
 @LoopStart :
 jz @LoopEnd
 inc eax
 dec edx
 jmp @LoopStart
 @LoopEnd :
 mov eax,ecx
 end;

 This loop consists of 4 instructions also with unknown sum of latency. We observe that the two inc/dec instructions are able to execute in parallel. Because the dec NoName instruction is not followed by the conditional jmp it would look like we throw away the benefit of not needing a cmp or test instruction to set the flags, but the jmp instruction does not change the flags and they are valid when we reach the jz instruction at the top of the loop. Only at the first iteration a cmp edx,0 instruction is needed.

 function ForLoopLoopInverNoName(Start, Stop : Integer) : Integer;
 asm
 //Result := 0;
 xor ecx,ecx
 //for I := Start to Stop do
 sub edx,eax
 jl @LoopEnd
 inc edx
 @LoopStart :
 inc eax
 dec edx
 jnz @LoopStart
 @LoopEnd :
 mov eax,ecx
 end;

 This loop consists of 3 instructions also with unknown sum of latency. Here there also is an independent inc/dec pair.

 This is the simple benchmark I have used to find the performance of the 4 functions

 var
 Starttime, Endtime, Runtime : TDateTime;
 I, LoopResult : Integer;
 RunTimeSec, NoOfLoopsPerSec, NoOfLoops, ClockCount, LoopEnd2Float, LoopEndFloat, LoopStartFloat : Double;

 begin
 Starttime := Time;
 for I := 1 to LOOPEND2 do
 begin
 LoopResult := ForLoopNoLoopInverNoName(LOOPSTART, LOOPEND);
 end;
 Endtime := Time;
 Runtime := Endtime - Starttime;
 CEdit.Text := IntToStr(LoopResult);
 RuntimeEdit4.Text := TimeToStr(Runtime);
 RunTimeSec := RunTime*24*60*60;
 LoopEnd2Float := LOOPEND2;
 LoopEndFloat := LOOPEND;
 LoopStartFloat := LOOPSTART;
 NoOfLoops := LoopEnd2Float * (LoopEndFloat - LoopStartFloat);
 NoOfLoopsPerSec := NoOfLoops / RunTimeSec;
 ClockCount := CLOCK / NoOfLoopsPerSec;
 ClockCountEdit4.Text := FloatToStrf(ClockCount, ffFixed, 9, 1);
 end;

 Results on P4 1920 are

 No Loop Inversion and No NoName variable 00:00:55 (2.7 Clock cycles)
 Loop Inversion but No NoName variable 00:00:39 (1.9 Clock cycles)
 No Loop Inversion but NoName variable 00:01:02 (3.0 Clock cycles)
 Loop Inversion + NoName 00:00:41 (2.0 Clock cycles)

 Results on P3 1400 are

 No Loop Inversion and No NoName variable 00:01:26 (3.0 Clock cycles)
 Loop Inversion but No NoName variable 00:01:26 (3.0 Clock cycles)
 No Loop Inversion but NoName variable 00:01:55 (4.0 Clock cycles)
 Loop Inversion + NoName 00:01:26 (3.0 Clock cycles)

 Of course the clock count numbers should be integers. On P4 half cycles are possible due to the double clocked alu's. Our timings are not as good as we could wish, but for comparing the performance of the loops they are OK.

 The conclusion on P4 is; apply loop inversion only or loop inversion with NoName variable optimization.

 The conclusion on P3 is; Do not apply the NoName variable optimization alone.

 The conclusion on both targets is; apply both optimizations as Delphi does.

 Also observe how efficient P4 is on this code.