Some of the parallel benchmarks aren't very stable with a limited number
of iterations so don't run them by default. They used to be not run at
all! So not a huge change imo.

Also bump the base bounds to make newer ghcs work again.

Increased input size and added argument for iteration count.
Run-times were: 0.095s/0.651s/3.257s
Are now: 0,145s/1,551s/8,064s

Times on Ryzen 4700 laptop (16GB):
From 0.038s/0.310s/1.474s
To   0.112s/1.222s/7.051s

Input generation with the shake-based nofib-run didn't work for
shootout/reverse-complement, so the test was never run.
Created a boot.sh script analogous to shootout/k-nucleotide.
Compilation of fasta-c.c and revcomp-c.c gave warnings for
ignored return values of 'write' - suppressed them by explicitly
ignoring with '(void)!' (negation and void cast).

real/smallpt's .slowstdout file was missing.
The value in spectral/dom-lt's was wrong, also runtime was lower
than 5 sek on my machine (should be between 5-10s),
so I increased the argument and updated dom-lt.slowstdout.

bc. program invocation on shake based nofib-run was broken.

Also added a few arguments to input to make single iteration longer
and adjusted iteration counts to get desired execution times.

ghc-9 introduces a regression(?) in FP precision (See ghc ticket #19569).
While this is unfortunate we want nofib to be useable with ghc 8.10 and
9.0 as far as possible.

So I added an additional argument to the benchmark which controls if we
output the benchmark result or not, which is off by default.

This allows easy checking of the output manually without causing
benchmark errors because the output differs.

I've also added the benchmark to the ghc testsuite to avoid future
regressions.

Alex D authored 3 years ago and Matthew Pickering committed 3 years ago

Current test does some floating-point manipulations which can be non-deterministic
due to different platforms/architectures. This change truncates the result to
avoid potential stdout mismatch.

Ben Gamari authored 3 years ago and Andreas Klebinger committed 3 years ago

By reducing the size of the raytraced image.

This can be used via passing --head to the nofib runner.

It will build benchmark dependencies using the given benchmark compiler
and hackage.head.

* We adjust for the specialisation of Data.List by hiding the generic
  version from the Prelude.

This should make this change backwards compatible with older base
versions.

* Adjust the gc benchmarks. I've mostly made them the same as their
  non-gc parts now.