[gambit-list] Guile's new compiler
Bradley Lucier
lucier at math.purdue.edu
Sat Nov 7 18:41:24 EST 2015
Speaking of benchmarking ...
I just ran the tests in misc/bench.tgz on 4.8.1 and with your latest
changes. The results are all over the place. Perhaps I'm doing
something wrong. Perhaps you might check it yourself.
The times for dderiv, lattice, conform, nd scheme roughly doubled with
the new compiler; takl, tfib, graphs, matrix, maze, mazefun, peval, pi
ray are up 50%.
Other things have improved, too, of course.
I'll attach the summaries.
Brad
-------------- next part --------------
(succeed 0 0)
(fail )
(crash )
(browse 1008 1005)
(cpstak 284 279)
(ctak 356 353)
(dderiv 840 840)
(deriv 288 288)
(destruc 288 286)
(diviter 308 307)
(divrec 456 455)
(puzzle 196 195)
(takl 332 330)
(trav1 224 223)
(trav2 708 710)
(triangl 308 306)
(fft 76 76)
(fib 236 237)
(fibfp 192 191)
(mbrot 168 166)
(nucleic 80 80)
(pnpoly 80 81)
(sum 72 73)
(sumfp 684 684)
(tak 272 272)
(tfib 640 638)
(ack 164 164)
(array1 68 64)
(cat 332 330)
(string 288 288)
(sum1 548 546)
(sumloop 1608 1609)
(tail 260 258)
(wc 164 163)
(conform 576 574)
(dynamic 260 254)
(earley 236 235)
(fibc 292 291)
(graphs 560 560)
(lattice 1144 1143)
(matrix 640 638)
(maze 340 338)
(mazefun 476 471)
(nqueens 300 297)
(paraffins 724 724)
(peval 476 476)
(pi 648 646)
(primes 484 485)
(ray 140 139)
(scheme 1036 1034)
(simplex 104 103)
(slatex 184 184)
(perm9 316 314)
(nboyer 368 364)
(sboyer 268 265)
(gcbench 732 728)
-------------- next part --------------
(succeed 0 0)
(fail )
(crash )
(browse 928 930)
(cpstak 328 328)
(ctak 404 406)
(dderiv 436 437)
(deriv 300 303)
(destruc 320 321)
(diviter 352 352)
(divrec 428 429)
(puzzle 172 171)
(takl 216 216)
(trav1 192 192)
(trav2 716 715)
(triangl 320 321)
(fft 80 82)
(fib 276 275)
(fibfp 216 217)
(mbrot 156 158)
(nucleic 72 66)
(pnpoly 76 75)
(sum 88 89)
(sumfp 704 705)
(tak 284 285)
(tfib 444 443)
(ack 168 166)
(array1 72 75)
(cat 360 358)
(string 300 301)
(sum1 404 405)
(sumloop 1576 1578)
(tail 308 309)
(wc 184 186)
(conform 316 315)
(dynamic 232 232)
(earley 260 260)
(fibc 288 289)
(graphs 356 358)
(lattice 652 652)
(matrix 420 421)
(maze 208 208)
(mazefun 300 301)
(nqueens 384 384)
(paraffins 808 809)
(peval 308 311)
(pi 480 482)
(primes 536 535)
(ray 100 101)
(scheme 428 428)
(simplex 136 136)
(slatex 168 168)
(perm9 324 325)
(nboyer 388 391)
(sboyer 280 279)
(gcbench 800 798)
More information about the Gambit-list
mailing list