[Yhc] Slower than nhc?

[Neil Mitchell]

Hi,

The benchmarks suggest Yhc is slower than nhc, is that really true?

dons>	Test: recursive (naive recursive functions)
	<dons>	http://www.cse.unsw.edu.au/~dons/code/nobench/imaginary/recursive
	<dons>	ghc 2.334 seconds (1.0 x)
	<dons>	ghc-old 2.353 seconds (1.0 x)
	<dons>	ghci 99.580 seconds (42.7 x)
	<lambdabot>	Title: Index of /~dons/code/nobench/imaginary/recursive
	<dons>	nhc98 111.865 seconds (47.9 x)
	<dons>	yhc 143.878 seconds (61.6 x)
	<dons>	for example, just ran
	<dons>	whereas it was hugs 208.266 seconds (72.7 x)
	<dons>	yhc 242.875 seconds (84.8 x)
	<dons>	before
	<ndm>	i find it hard to believe that Yhc is slower than nhc...
	<ndm>	i'll check with Tom
	<Igloo>	Does nhc compile to bytecode or machinecode?
	<dons>	sorry, before I did a 'release' build, yhc ran recursive at
yhc 172.517 seconds (73.6 x)

Could someone who has both Yhc and nhc try and replicate these
benchmarks for the above program? I remember Tom giving out numbers
than Yhc was "around 20% faster" than nhc, it would be a shame if
we've broken that since...

Thanks

Neil