Re: [Lse-tech] [PATCH] New dcache / inode hash tuning patch

["Martin J. Bligh" <mbligh@aracnet.com>]

It actually seems a fraction slower (see systimes for Kernbench-16,
for instance).

Kernbench-2: (make -j N vmlinux, where N = 2 x num_cpus)
                              Elapsed        User      System         CPU
              2.5.62-mjb3       43.92      557.65       94.12     1483.50
         2.5.62-mjb3-andi       44.28      557.90       95.79     1475.67

Kernbench-16: (make -j N vmlinux, where N = 16 x num_cpus)
                              Elapsed        User      System         CPU
              2.5.62-mjb3       45.21      560.46      114.58     1492.67
         2.5.62-mjb3-andi       45.39      561.29      117.73     1495.67

DISCLAIMER: SPEC(tm) and the benchmark name SDET(tm) are registered
trademarks of the Standard Performance Evaluation Corporation. This 
benchmarking was performed for research purposes only, and the run results
are non-compliant and not-comparable with any published results.

Results are shown as percentages of the first set displayed

SDET 1  (see disclaimer)
                                Throughput    Std. Dev
                   2.5.62-mjb3       100.0%         3.1%
              2.5.62-mjb3-andi       101.8%         6.6%

SDET 2  (see disclaimer)
                                Throughput    Std. Dev
                   2.5.62-mjb3       100.0%         4.0%
              2.5.62-mjb3-andi        96.4%         4.8%

SDET 4  (see disclaimer)
                                Throughput    Std. Dev
                   2.5.62-mjb3       100.0%         2.0%
              2.5.62-mjb3-andi       100.4%         2.4%

SDET 8  (see disclaimer)
                                Throughput    Std. Dev
                   2.5.62-mjb3       100.0%         4.6%
              2.5.62-mjb3-andi       100.3%         2.3%

SDET 16  (see disclaimer)
                                Throughput    Std. Dev
                   2.5.62-mjb3       100.0%         2.7%
              2.5.62-mjb3-andi        96.6%         4.3%

SDET 32  (see disclaimer)
                                Throughput    Std. Dev
                   2.5.62-mjb3       100.0%         0.9%
              2.5.62-mjb3-andi        97.0%         2.1%

SDET 64  (see disclaimer)
                                Throughput    Std. Dev
                   2.5.62-mjb3       100.0%         1.1%
              2.5.62-mjb3-andi       100.2%         0.6%


Diffprofile (+ worse with your patch, - better)

       500     3.1% total
       138     6.9% .text.lock.file_table
       103     2.2% default_idle
        73     9.3% d_lookup
        50    11.0% __copy_to_user_ll
        45    27.4% file_move
        40    28.2% path_lookup
...        
	-7    -3.9% do_schedule
        -9    -2.0% get_empty_filp
       -10    -8.6% dput
       -10    -9.8% link_path_walk

That text.lock.file_table has been bugging me for a bit, and I need to
drill down into it some more.

M.