Edit · semi tagging · Wiki · Glasgow Haskell Compiler / GHC

Title

Tip: You can move this page by adding the path to the beginning of the title. More information.

Format

Content

**NB.** This is an old design document.  The current implementation of Pointer Tagging is described in [Commentary/Rts/HaskellExecution/PointerTagging](commentary/rts/haskell-execution/pointer-tagging).

---

# OLD DESIGN DOCUMENT: The semi-tagging optimisation

Here I describe the design of the semi-tagging optimisation. Originally the text comes from [http://hackage.haskell.org/trac/summer-of-code/ticket/48](http://hackage.haskell.org/trac/summer-of-code/ticket/48)

This page reflects my current understanding on the compiler and the RTS, so if there is something wrong, just yell!

## The starting point

Currently when evaluating an expression that is the scrutinee of a case:

```wiki
case x of { ... }
```

GHC jumps to the code for (i.e. "enters") the x closure, which returns when x is evaluated. Commonly, x is already evaluated, and the code for an evaluated constructor just (vector) returns immediately.

The code for the `not` function:

```wiki
not x = case x of
        False -> True
        True  -> False
```

jumps to the boolean argument, passed in `R2`, after pushing a case frame (the continuation of the function):

```wiki
        <stack check omitted>
        R1 = R2;
        I64[Sp + (-8)] = notcont_info;
          -- push not continuation on the stack
        Sp = Sp + (-8);
        jump I64[R1];
          -- enter x
```

(Note that the examples in this page assume a 64-bit architecture, therefore a word consists of 8 bytes.)

Before looking at the rest of the `not` function, let's look at the code for the `True` and `False` closures

```wiki
True_info:
        jump [[Sp]] --address to True alternative;

False_info:
        jump [[Sp]+8] --address to False alternative;
```

they just jump to the appropriate case alternative that is evaluating the closure. These addresses are calculated from the case frame that is on the top of the stack. In this case they select the alternatives from the jump table that is referred to by the `not` case frame. Below you see the `True` alternative

```wiki
notcont_info {
  notcont_0_alt,
  notcont_1_alt
};

notcont_0_alt() {
        R1 = False_closure;
        Sp = Sp + 8;
        jump [[Sp]+8] --address to False alternative;
}
```

and the `False` alternative of the `not` function.

```wiki
notcont_1_alt() {
        R1 = True_closure;
        Sp = Sp + 8;
        jump [[Sp]] --address to True alternative;
}
```

Just like the constructor closures, they jump to the appropriate branch of the case expression that is evaluating the `not` function.

## Testing before jumping

The simplest optimisation is this.  Instead of entering the closure, grab its info pointer, and follow the info pointer to get the tag.  Now test the tag; if it's evaluated, don't enter the closure.

The benefit is that processors are typically faster at "test-and-jump to known location" than they are at "jump to this pointer".

Under this scheme, the entry code for the `not` function would look as follows:

```wiki
        <stack check omitted>
        if([[R2]+type_offset] == 0) goto unevaluated
        if([[R2]+type_offset] > CONSTR_NOCAF_STATIC) goto unevaluated
        goto evaluated
          -- If the closure type is a constructor we can extract
          -- the tag directly in label evaluated.
          -- Otherwise we enter it as usual:
unevaluated:
        R1 = R2;
        I64[Sp + (-8)] = notcont_info;
        Sp = Sp + (-8);
        jump I64[R1];
evaluated:
        R2=[[R2]+type_offset];
          -- extract constructor tag from pointer
        if(R2==0) goto notcont_0_alt
        goto notcont_1_alt
```

## Tagging the LSB of an evaluated closure

>
>
> The idea is to encode the fact that a pointer points to an evaluated object by setting the LSB of the pointer. If the case expression detects that the closure is evaluated, it can avoid the jump and return, which are expensive on modern processors (indirect jumps).
>
>

<table><tr><th>  </th>
<th> bits 31..2 </th>
<th> bits 1 0 
</th></tr>
<tr><th> unevaluated closure </th>
<th> ptr </th>
<th> 00 
</th></tr>
<tr><th> evaluated constructor closure </th>
<th> ptr </th>
<th> 01 
</th></tr></table>

This would require modifying

- the code generation so that when allocating a constructor, the pointer to it has the appropriate bits set (just a matter of adjusting the offset from Hp)
- perhaps, make the entry code for a constructor return a pointer with the appropriate bits set. That way, a function like `f xs = head xs` would enter the first element of the list, and return to the caller with appropriate tag bits set.
- the GC to set the LSB bit of constructor closure pointers,
- the GC and the RTS code to mask out the LSB pointer when dereferencing it,
- the code generation to test the LSB bit and case expressions and avoid the indirect jump.

```wiki
        <stack check omitted>
        if(R2 & 1 == 1) goto tagged
        R1 = R2;
        I64[Sp + (-8)] = sej_info;
        Sp = Sp + (-8);
        jump I64[R1];
tagged:
        R1 = R2 & ~1;  // mask pointer tag out
        R2 = [[R1]+type_offset];
          -- extract constructor tag from pointer
        if(R2==0) goto notcont_0_alt
        goto notcont_1_alt
```

## Using more than one bit

We can go a bit further than this, too. Since there are 2 spare bits (4 on a 64-bit machine), we can encode 4 (16) states. Taking 0 to mean "unevaluted", that leaves 3 (15) states to encode the values for the "tag" of the constructor. eg. an evaluated Bool would use 1 to indicate False and 2 to indicate True. An evaluated list cell would use 1 to indicate \[\] and 2 to indicate (:).

<table><tr><th>  </th>
<th> bits 31..2 </th>
<th> bits 1 0 
</th></tr>
<tr><th> unevaluated closure </th>
<th> ptr </th>
<th> 00 
</th></tr>
<tr><th> cons. no. 1    </th>
<th> ptr </th>
<th> 01 
</th></tr>
<tr><th> cons. no. 2    </th>
<th> ptr </th>
<th> 10 
</th></tr>
<tr><th> cons. no. 3    </th>
<th> ptr </th>
<th> 11 
</th></tr></table>

The nice thing about the current approach is that code size is small; implementing the test and jump will certainly add extra code to compiled case expressions. But the gains might be worth it. Complexity-wise this means masking out these bits when following any pointer to a heap object, which means carefully checking most of the runtime.

This would require modifying all of the above plus modifying

- the code generator so that it checks whether the number of constructors is smaller or equal than 3/15.

## Using a tag directly in the pointer

Constructors without children (such as `False` and `True`) only need their tag to be represented. Hence we can drop the pointer altogether as follows:

<table><tr><th>  </th>
<th> bits 31..2 </th>
<th> bits 1 0 
</th></tr>
<tr><th> unevaluated closure </th>
<th> ptr </th>
<th> 00 
</th></tr>
<tr><th> cons. w/no children </th>
<th> tag </th>
<th> 01 
</th></tr>
<tr><th> cons. w/children no. 1    </th>
<th> ptr </th>
<th> 10 
</th></tr>
<tr><th> cons. w/children no. 2    </th>
<th> ptr </th>
<th> 11 
</th></tr></table>

This still limits us to data types that have no more than two constructors with children. We can improve on this by noting that pointers will not point to low addresses. So we can make a simple test to distinguish between tags and pointers:

<table><tr><th>  </th>
<th> bits 31..2 </th>
<th> bits 1 0 </th>
<th> 
</th></tr>
<tr><th> unevaluated closure </th>
<th> ptr </th>
<th> 00 </th>
<th> 
</th></tr>
<tr><th> cons. w/no children </th>
<th> tag </th>
<th> 01 </th>
<th> tag &lt; no. of  constructors
</th></tr>
<tr><th> cons. w/ children </th>
<th> ptr </th>
<th> 01 </th>
<th> ptr &gt;= no. of constructors 
</th></tr>
<tr><th> cons. w/children no. 1    </th>
<th> ptr </th>
<th> 10 </th>
<th> 
</th></tr>
<tr><th> cons. w/children no. 2    </th>
<th> ptr </th>
<th> 11 </th>
<th> 
</th></tr></table>

Of course this assumes that we don't have data types with too many thousands of constructors.

It might be possible that the case code for the alternatives above is becoming too complex. We might settle for the following "simple" option:

To link to a (new) page, simply type [Link Title](page-slug). More examples are in the documentation.

Commit message

Cancel

GitLab

semi tagging · Edit Page