nofib-compare: Put relative change first, remove duplicate stderr column
I find it far more intuitive to have the relative changes appear as the first column. In this way, it also lines up with the geom. mean.
Example:
# bytes allocated
+-------------------------------++--+------------+-----------+------------+
| || | new/ (rel) | std. err. | old/ |
+===============================++==+============+===========+============+
| imaginary/bernouilli || | 0.00% | 0.0% | 2.823e9 |
| imaginary/digits-of-e1 || | 0.00% | 0.0% | 7.830e8 |
| imaginary/digits-of-e2 || | 0.00% | 0.0% | 2.131e9 |
| imaginary/exp3_8 || | 0.00% | 0.0% | 5.888e9 |
| imaginary/gen_regexps || | 0.00% | 0.0% | 9.101e8 |
| imaginary/integrate || | 0.00% | 0.0% | 3.424e9 |
| imaginary/kahan || | 0.00% | 0.0% | 49008.000 |
| imaginary/paraffins || | 0.00% | 0.0% | 3.821e9 |
| imaginary/primes || | 0.00% | 0.0% | 2.928e9 |
| imaginary/queens || | 0.00% | 0.0% | 6.317e8 |
| imaginary/rfib || | 0.00% | 0.0% | 139104.000 |
| imaginary/tak || | 0.00% | 0.0% | 96288.000 |
| imaginary/wheel-sieve1 || | -0.05% | 0.0% | 1.344e8 |
| imaginary/wheel-sieve2 || | 0.00% | 0.0% | 2.428e9 |
| imaginary/x2n1 || | 0.00% | 0.0% | 56928.000 |
| real/anna || | +0.03% | 0.0% | 2.062e9 |
| real/ben-raytrace || | 0.00% | 0.0% | 2.010e9 |
| real/bspt || | -0.00% | 0.0% | 3.736e9 |
| real/cacheprof || | +0.30% | 0.0% | 2.517e9 |
| real/compress || | 0.00% | 0.0% | 2.793e9 |
| real/compress2 || | -1.90% | 0.0% | 3.436e9 |
| real/eff/CS || | 0.00% | 0.0% | 1.601e8 |
| real/eff/CSD || | 0.00% | 0.0% | 1.600e9 |
| real/eff/FS || | 0.00% | 0.0% | 1.760e9 |
| real/eff/S || | 0.00% | 0.0% | 2.401e8 |
| real/eff/VS || | +49.68% | 0.0% | 4.831e8 |
| real/eff/VSD || | 0.00% | 0.0% | 50208.000 |
| real/eff/VSM || | 0.00% | 0.0% | 4.001e8 |
| real/fem || | 0.00% | 0.0% | 4.810e9 |
| real/fluid || | +0.00% | 0.0% | 2.137e9 |
| real/fulsom || | -0.00% | 0.0% | 1.959e9 |
| real/gamteb || | 0.00% | 0.0% | 4.309e9 |
| real/gg || | +0.00% | 0.0% | 2.697e9 |
| real/grep || | 0.00% | 0.0% | 4.622e9 |
| real/hidden || | 0.00% | 0.0% | 5.096e9 |
| real/hpg || | 0.00% | 0.0% | 3.180e9 |
| real/infer || | 0.00% | 0.0% | 1.730e9 |
| real/lift || | 0.00% | 0.0% | 2.597e9 |
| real/linear || | +0.06% | 0.0% | 5.445e9 |
| real/maillist || | 0.00% | 0.0% | 3.266e9 |
| real/mkhprog || | 0.00% | 0.0% | 1.194e10 |
| real/parser || | +0.71% | 0.0% | 2.668e9 |
| real/pic || | 0.00% | 0.0% | 1.463e9 |
| real/prolog || | +0.11% | 0.0% | 2.915e9 |
| real/reptile || | +0.06% | 0.0% | 4.050e9 |
| real/rsa || | 0.00% | 0.0% | 1.999e9 |
| real/scs || | 0.00% | 0.0% | 3.575e9 |
| real/smallpt || | -11.44% | 0.0% | 1.301e9 |
| real/symalg || | -0.00% | 0.0% | 3.198e8 |
| real/veritas || | 0.00% | 0.0% | 4.176e9 |
| shootout/binary-trees || | 0.00% | 0.0% | 4.335e9 |
| shootout/fannkuch-redux || | 0.00% | 0.0% | 68288.000 |
| shootout/fasta || | -3.60% | 0.0% | 7.676e8 |
| shootout/k-nucleotide || | -0.04% | 0.0% | 2.183e7 |
| shootout/n-body || | -0.04% | 0.0% | 162376.000 |
| shootout/pidigits || | 0.00% | 0.0% | 1.027e10 |
| shootout/reverse-complement || | +0.00% | 0.0% | 2.013e9 |
| shootout/spectral-norm || | -0.01% | 0.0% | 270256.000 |
| spectral/ansi || | 0.00% | 0.0% | 9.390e9 |
| spectral/atom || | 0.00% | 0.0% | 3.580e9 |
| spectral/awards || | 0.00% | 0.0% | 4.793e9 |
| spectral/banner || | 0.00% | 0.0% | 6.124e9 |
| spectral/boyer || | 0.00% | 0.0% | 4.665e9 |
| spectral/boyer2 || | 0.00% | 0.0% | 1.220e9 |
| spectral/calendar || | 0.00% | 0.0% | 7.144e9 |
| spectral/cichelli || | 0.00% | 0.0% | 1.941e9 |
| spectral/circsim || | 0.00% | 0.0% | 4.850e9 |
| spectral/clausify || | 0.00% | 0.0% | 2.095e9 |
| spectral/constraints || | 0.00% | 0.0% | 4.833e9 |
| spectral/cryptarithm1 || | 0.00% | 0.0% | 5.981e9 |
| spectral/cryptarithm2 || | 0.00% | 0.0% | 3.722e9 |
| spectral/cse || | 0.00% | 0.0% | 3.802e9 |
| spectral/dom-lt || | -0.05% | 0.0% | 3.945e9 |
| spectral/eliza || | 0.00% | 0.0% | 4.097e9 |
| spectral/exact-reals || | +0.69% | 0.0% | 1.116e9 |
| spectral/expert || | +0.08% | 0.0% | 2.114e9 |
| spectral/fft2 || | 0.00% | 0.0% | 1.420e9 |
| spectral/fibheaps || | 0.00% | 0.0% | 5.561e9 |
| spectral/fish || | -0.00% | 0.0% | 6.193e9 |
| spectral/gcd || | 0.00% | 0.0% | 2.057e9 |
| spectral/hartel/comp_lab_zift || | 0.00% | 0.0% | 4.753e9 |
| spectral/hartel/event || | +1.34% | 0.0% | 3.334e9 |
| spectral/hartel/fft || | +0.29% | 0.0% | 1.544e9 |
| spectral/hartel/genfft || | 0.00% | 0.0% | 8.157e9 |
| spectral/hartel/ida || | 0.00% | 0.0% | 3.501e9 |
| spectral/hartel/listcompr || | 0.00% | 0.0% | 6.164e9 |
| spectral/hartel/listcopy || | 0.00% | 0.0% | 6.765e9 |
| spectral/hartel/nucleic2 || | 0.00% | 0.0% | 2.817e9 |
| spectral/hartel/parstof || | 0.00% | 0.0% | 1.225e9 |
| spectral/hartel/sched || | +1.95% | 0.0% | 3.665e9 |
| spectral/hartel/solid || | 0.00% | 0.0% | 3.440e9 |
| spectral/hartel/transform || | +1.42% | 0.0% | 3.924e9 |
| spectral/hartel/typecheck || | 0.00% | 0.0% | 2.629e9 |
| spectral/hartel/wang || | +0.02% | 0.0% | 4.735e9 |
| spectral/hartel/wave4main || | -0.23% | 0.0% | 1.095e9 |
| spectral/integer || | 0.00% | 0.0% | 2.968e9 |
| spectral/knights || | 0.00% | 0.0% | 1.107e9 |
| spectral/lambda || | 0.00% | 0.0% | 2.947e9 |
| spectral/last-piece || | -1.16% | 0.0% | 7.425e8 |
| spectral/lcss || | 0.00% | 0.0% | 6.563e9 |
| spectral/life || | +0.00% | 0.0% | 6.805e9 |
| spectral/mandel || | 0.00% | 0.0% | 1.723e9 |
| spectral/mandel2 || | 0.00% | 0.0% | 2.258e7 |
| spectral/mate || | 0.00% | 0.0% | 3.465e9 |
| spectral/minimax || | +0.00% | 0.0% | 2.539e9 |
| spectral/multiplier || | 0.00% | 0.0% | 2.883e9 |
| spectral/para || | 0.00% | 0.0% | 3.962e9 |
| spectral/power || | -0.00% | 0.0% | 1.428e9 |
| spectral/pretty || | 0.00% | 0.0% | 136296.000 |
| spectral/primetest || | 0.00% | 0.0% | 8.427e8 |
| spectral/puzzle || | 0.00% | 0.0% | 1.530e9 |
| spectral/rewrite || | +0.00% | 0.0% | 1.716e9 |
| spectral/scc || | 0.00% | 0.0% | 57520.000 |
| spectral/simple || | 0.00% | 0.0% | 1.204e9 |
| spectral/sorting || | +0.00% | 0.0% | 3.076e9 |
| spectral/sphere || | 0.00% | 0.0% | 2.233e9 |
| spectral/treejoin || | 0.00% | 0.0% | 2.794e9 |
+===============================++==+============+===========+============+
| geom mean || | +0.24% | | |
+-------------------------------++--+------------+-----------+------------+
Edited by Sebastian Graf