• Robert Griesemer's avatar
    math/big: faster assembly kernels for AddVx/SubVx for 386. · 654a185f
    Robert Griesemer authored
    (analog to Change-Id: Ia473e9ab9c63a955c252426684176bca566645ae)
    
    Fixes #9243.
    
    benchmark              old ns/op     new ns/op     delta
    BenchmarkAddVV_1       5.76          5.60          -2.78%
    BenchmarkAddVV_2       7.17          6.98          -2.65%
    BenchmarkAddVV_3       8.69          8.57          -1.38%
    BenchmarkAddVV_4       10.5          10.5          +0.00%
    BenchmarkAddVV_5       13.3          11.6          -12.78%
    BenchmarkAddVV_1e1     20.4          19.3          -5.39%
    BenchmarkAddVV_1e2     166           140           -15.66%
    BenchmarkAddVV_1e3     1588          1278          -19.52%
    BenchmarkAddVV_1e4     16138         12657         -21.57%
    BenchmarkAddVV_1e5     167608        127836        -23.73%
    BenchmarkAddVW_1       4.87          4.76          -2.26%
    BenchmarkAddVW_2       6.10          6.07          -0.49%
    BenchmarkAddVW_3       7.75          7.65          -1.29%
    BenchmarkAddVW_4       9.30          9.39          +0.97%
    BenchmarkAddVW_5       10.8          10.9          +0.93%
    BenchmarkAddVW_1e1     18.8          18.8          +0.00%
    BenchmarkAddVW_1e2     143           134           -6.29%
    BenchmarkAddVW_1e3     1390          1266          -8.92%
    BenchmarkAddVW_1e4     13877         12545         -9.60%
    BenchmarkAddVW_1e5     155330        125432        -19.25%
    
    benchmark              old MB/s     new MB/s     speedup
    BenchmarkAddVV_1       5556.09      5715.12      1.03x
    BenchmarkAddVV_2       8926.55      9170.64      1.03x
    BenchmarkAddVV_3       11042.15     11201.77     1.01x
    BenchmarkAddVV_4       12168.21     12245.50     1.01x
    BenchmarkAddVV_5       12041.39     13805.73     1.15x
    BenchmarkAddVV_1e1     15659.65     16548.18     1.06x
    BenchmarkAddVV_1e2     19268.57     22728.64     1.18x
    BenchmarkAddVV_1e3     20141.45     25033.36     1.24x
    BenchmarkAddVV_1e4     19827.86     25281.92     1.28x
    BenchmarkAddVV_1e5     19092.06     25031.92     1.31x
    BenchmarkAddVW_1       822.12       840.92       1.02x
    BenchmarkAddVW_2       1310.89      1317.89      1.01x
    BenchmarkAddVW_3       1549.31      1568.26      1.01x
    BenchmarkAddVW_4       1720.45      1703.77      0.99x
    BenchmarkAddVW_5       1857.12      1828.66      0.98x
    BenchmarkAddVW_1e1     2126.39      2132.38      1.00x
    BenchmarkAddVW_1e2     2784.49      2969.21      1.07x
    BenchmarkAddVW_1e3     2876.89      3157.35      1.10x
    BenchmarkAddVW_1e4     2882.32      3188.51      1.11x
    BenchmarkAddVW_1e5     2575.16      3188.96      1.24x
    
    (measured on OS X 10.9.5, 2.3 GHz Intel Core i7, 8GB 1333 MHz DDR3)
    
    Change-Id: I46698729d5e0bc3e277aa0146a9d7a086c0c26f1
    Reviewed-on: https://go-review.googlesource.com/2560Reviewed-by: 's avatarKeith Randall <khr@golang.org>
    654a185f
Name
Last commit
Last update
..
arith.go Loading commit data...
arith_386.s Loading commit data...
arith_amd64.s Loading commit data...
arith_amd64p32.s Loading commit data...
arith_arm.s Loading commit data...
arith_decl.go Loading commit data...
arith_ppc64x.s Loading commit data...
arith_test.go Loading commit data...
calibrate_test.go Loading commit data...
example_test.go Loading commit data...
gcd_test.go Loading commit data...
hilbert_test.go Loading commit data...
int.go Loading commit data...
int_test.go Loading commit data...
nat.go Loading commit data...
nat_test.go Loading commit data...
rat.go Loading commit data...
rat_test.go Loading commit data...