[LLVMdev] More Encoding Ideas

[Reid Spencer]

On Mon, 2004-08-23 at 19:29, Robert Mykland wrote:
> >Seriously though, unless you want to do it, I think I'll probably do
> >this sometime before 1.4 is released. The savings are not huge (type
> >table is pretty small), but every little bit helps and this will halve
> >the number of entries.
> 
> Maybe a few hundred bytes for big files.  Go for it.  I'm still trying to 
> get my code generator working with 1.3.  With all the optimizations you 
> guys did since 0.9, the code is all different, in a good way.  But I have 
> to make my code generator smarter to do the simple examples it could do 
> before, just because different stuff I don't yet support got used.

This change won't get done any time soon because (as discussed on other
messages on this thread), it will actually cause the bytecode file to be
larger because slot values will be higher. What we're going to do in the
future is to sort the type and value lists by frequency of use. That way
the most frequently used items will have the smallest slot numbers and
therefore tend to save the most bytes because of VBR encoding. When I
get around to doing that, I'll reconsider this issue and see if there's
a way not to write the derived pointer types.

> >Could you please open a bug report and attach a test case to it? This is
> >a serious problem that I need to fix but in order to do that I need a
> >reliable test case for repeatability.

> Will do.
> 
> -- Robert.

Thanks!

Reid.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20040823/be5c7b6e/attachment.sig>