Memory bandwidth
Using the program cachebench the results are:
# ./cachebench
Memory Set Library Cache Test
C Size Nanosec MB/sec % Chnge
------- ------- ------- -------
256 0.05 20205.70 1.00
256 0.05 20210.63 1.00
384 0.05 20445.46 0.99
384 0.05 20447.49 1.00
512 0.04 22799.44 0.90
512 0.04 22851.47 1.00
768 0.04 26378.00 0.87
768 0.04 26362.02 1.00
1024 0.03 27987.20 0.94
1024 0.03 27868.26 1.00
1536 0.03 30555.05 0.91
1536 0.03 30600.60 1.00
2048 0.03 31588.43 0.97
2048 0.03 31572.33 1.00
3072 0.03 31405.71 1.01
3072 0.03 31397.64 1.00
4096 0.03 33861.48 0.93
4096 0.03 33910.90 1.00
6144 0.03 34731.45 0.98
6144 0.03 34755.92 1.00
8192 0.03 35086.88 0.99
8192 0.03 35120.02 1.00
12288 0.03 34854.81 1.01
12288 0.03 34863.39 1.00
16384 0.03 35282.06 0.99
16384 0.03 35286.52 1.00
24576 0.03 35720.84 0.99
24576 0.03 35720.80 1.00
32768 0.04 26525.08 1.35
32768 0.04 26534.71 1.00
49152 0.06 15003.31 1.77
49152 0.06 15004.54 1.00
65536 0.06 15026.68 1.00
65536 0.06 15029.94 1.00
98304 0.06 15048.04 1.00
98304 0.06 15052.69 1.00
131072 0.06 15060.22 1.00
131072 0.06 15065.03 1.00
196608 0.06 15073.04 1.00
196608 0.06 15076.64 1.00
262144 0.06 15079.84 1.00
262144 0.06 15080.40 1.00
393216 0.06 15086.09 1.00
393216 0.06 15087.90 1.00
524288 0.06 15088.01 1.00
524288 0.06 15090.52 1.00
786432 0.06 15089.41 1.00
786432 0.06 15092.04 1.00
1048576 0.06 15009.05 1.01
1048576 0.06 15013.61 1.00
1572864 0.06 14675.01 1.02
1572864 0.06 14690.48 1.00
2097152 0.07 14507.26 1.01
2097152 0.07 14527.73 1.00
3145728 0.21 4640.97 3.13
3145728 0.21 4641.73 1.00
4194304 0.21 4640.42 1.00
4194304 0.21 4639.80 1.00
6291456 0.21 4638.57 1.00
6291456 0.21 4641.71 1.00
8388608 0.21 4641.45 1.00
8388608 0.21 4641.38 1.00
12582912 0.21 4639.89 1.00
12582912 0.21 4641.15 1.00
16777216 0.21 4635.95 1.00
16777216 0.21 4638.22 1.00
Memory Copy Library Cache Test
C Size Nanosec MB/sec % Chnge
------- ------- ------- -------
256 0.03 27257.58 1.00
256 0.04 27240.95 1.00
384 0.03 29720.92 0.92
384 0.03 28648.10 1.04
512 0.03 31178.50 0.92
512 0.03 31187.37 1.00
768 0.03 32859.87 0.95
768 0.03 32851.33 1.00
1024 0.03 32810.48 1.00
1024 0.03 32722.88 1.00
1536 0.04 25218.32 1.30
1536 0.04 25301.41 1.00
2048 0.04 26790.56 0.94
2048 0.04 26784.68 1.00
3072 0.03 28591.99 0.94
3072 0.03 28568.35 1.00
4096 0.03 29342.82 0.97
4096 0.03 29395.67 1.00
6144 0.03 30380.29 0.97
6144 0.03 30379.74 1.00
8192 0.03 30895.29 0.98
8192 0.03 30896.40 1.00
12288 0.03 31428.01 0.98
12288 0.03 31431.92 1.00
16384 0.03 28911.49 1.09
16384 0.03 28901.91 1.00
24576 0.06 15160.88 1.91
24576 0.06 15163.85 1.00
32768 0.06 15465.43 0.98
32768 0.06 15462.28 1.00
49152 0.06 15812.25 0.98
49152 0.06 15816.79 1.00
65536 0.06 16020.25 0.99
65536 0.06 16025.29 1.00
98304 0.06 16218.49 0.99
98304 0.06 16224.20 1.00
131072 0.06 16325.38 0.99
131072 0.06 16319.58 1.00
196608 0.06 16432.16 0.99
196608 0.06 16421.04 1.00
262144 0.06 16478.84 1.00
262144 0.06 16473.13 1.00
393216 0.06 16481.93 1.00
393216 0.06 16479.39 1.00
524288 0.06 16482.07 1.00
524288 0.06 16482.70 1.00
786432 0.06 16378.76 1.01
786432 0.06 16370.23 1.00
1048576 0.06 16384.89 1.00
1048576 0.06 16390.15 1.00
1572864 0.07 12763.70 1.28
1572864 0.07 12786.26 1.00
2097152 0.08 11335.55 1.13
2097152 0.08 11340.17 1.00
3145728 0.11 8450.05 1.34
3145728 0.11 8446.56 1.00
4194304 0.15 6548.52 1.29
4194304 0.15 6551.65 1.00
6291456 0.18 5372.22 1.22
6291456 0.18 5369.39 1.00
8388608 0.19 5088.09 1.06
8388608 0.19 5083.32 1.00
12582912 0.19 4985.97 1.02
12582912 0.19 4984.00 1.00
16777216 0.19 4982.10 1.00
16777216 0.19 4976.06 1.00
double Read Cache Test
C Size Nanosec MB/sec % Chnge
------- ------- ------- -------
256 3.75 2035.09 1.00
256 3.75 2034.92 1.00
384 3.75 2034.97 1.00
384 3.75 2035.07 1.00
512 3.75 2035.35 1.00
512 3.75 2035.06 1.00
768 3.79 2011.75 1.01
768 3.79 2011.86 1.00
1024 3.78 2017.48 1.00
1024 3.78 2017.71 1.00
1536 3.77 2023.52 1.00
1536 3.77 2023.31 1.00
2048 3.76 2026.62 1.00
2048 3.76 2026.46 1.00
3072 3.76 2029.42 1.00
3072 3.76 2029.07 1.00
4096 3.76 2027.86 1.00
4096 3.76 2027.50 1.00
6144 3.76 2030.16 1.00
6144 3.76 2030.24 1.00
8192 3.76 2029.64 1.00
8192 3.76 2029.94 1.00
12288 3.76 2030.85 1.00
12288 3.76 2030.39 1.00
16384 3.76 2030.98 1.00
16384 3.76 2030.92 1.00
24576 3.76 2031.55 1.00
24576 3.76 2031.32 1.00
32768 3.76 2031.11 1.00
32768 3.76 2030.82 1.00
49152 3.76 2031.34 1.00
49152 3.76 2031.40 1.00
65536 3.76 2031.45 1.00
65536 3.76 2031.40 1.00
98304 3.76 2031.61 1.00
98304 3.76 2031.50 1.00
131072 3.76 2031.78 1.00
131072 3.75 2031.83 1.00
196608 3.75 2031.93 1.00
196608 3.75 2031.96 1.00
262144 3.75 2032.15 1.00
262144 3.75 2031.82 1.00
393216 3.75 2032.01 1.00
393216 3.75 2032.15 1.00
524288 3.75 2032.03 1.00
524288 3.76 2031.63 1.00
786432 3.75 2031.84 1.00
786432 3.75 2031.91 1.00
1048576 3.76 2031.78 1.00
1048576 3.75 2031.83 1.00
1572864 3.76 2031.07 1.00
1572864 3.76 2031.14 1.00
2097152 3.76 2030.76 1.00
2097152 3.76 2030.54 1.00
3145728 3.76 2029.86 1.00
3145728 3.77 2023.50 1.00
4194304 3.78 2016.44 1.00
4194304 3.78 2016.54 1.00
6291456 3.83 1993.31 1.01
6291456 3.83 1994.22 1.00
8388608 3.84 1988.83 1.00
8388608 3.84 1988.37 1.00
12582912 3.84 1986.52 1.00
12582912 3.84 1986.86 1.00
16777216 3.85 1979.75 1.00
16777216 3.84 1986.33 1.00
double Write Cache Test
C Size Nanosec MB/sec % Chnge
------- ------- ------- -------
256 1.26 6042.55 1.00
256 1.26 6040.44 1.00
384 1.26 6063.37 1.00
384 1.26 6063.59 1.00
512 1.26 6073.30 1.00
512 1.26 6074.22 1.00
768 0.96 7920.17 0.77
768 0.96 7921.69 1.00
1024 0.93 8197.47 0.97
1024 0.93 8197.68 1.00
1536 0.90 8493.55 0.97
1536 0.90 8493.27 1.00
2048 0.88 8651.66 0.98
2048 0.88 8651.29 1.00
3072 0.87 8813.72 0.98
3072 0.87 8813.70 1.00
4096 0.86 8896.61 0.99
4096 0.86 8897.41 1.00
6144 0.85 8982.53 0.99
6144 0.85 8984.13 1.00
8192 0.85 8988.53 1.00
8192 0.85 9026.08 1.00
12288 0.84 9069.26 1.00
12288 0.84 9070.03 1.00
16384 0.84 9091.72 1.00
16384 0.84 9092.04 1.00
24576 0.84 9086.20 1.00
24576 0.84 9113.88 1.00
32768 0.84 9124.38 1.00
32768 0.84 9124.01 1.00
49152 0.84 9135.65 1.00
49152 0.83 9137.08 1.00
65536 0.83 9141.01 1.00
65536 0.83 9141.54 1.00
98304 0.83 9146.49 1.00
98304 0.84 9114.36 1.00
131072 0.83 9149.46 1.00
131072 0.83 9150.82 1.00
196608 0.83 9152.17 1.00
196608 0.83 9150.87 1.00
262144 0.83 9153.54 1.00
262144 0.83 9153.42 1.00
393216 0.83 9154.30 1.00
393216 0.83 9155.94 1.00
524288 0.83 9154.90 1.00
524288 0.83 9155.41 1.00
786432 0.83 9154.75 1.00
786432 0.83 9156.89 1.00
1048576 0.84 9135.19 1.00
1048576 0.84 9135.87 1.00
1572864 0.85 9027.86 1.01
1572864 0.85 9027.23 1.00
2097152 0.85 8989.68 1.00
2097152 0.85 9022.11 1.00
3145728 0.86 8882.21 1.02
3145728 0.86 8861.15 1.00
4194304 1.32 5798.98 1.53
4194304 1.31 5805.26 1.00
6291456 2.36 3239.06 1.79
6291456 2.35 3242.85 1.00
8388608 2.83 2698.66 1.20
8388608 2.82 2704.82 1.00
12582912 3.06 2496.80 1.08
12582912 3.05 2500.96 1.00
16777216 3.09 2465.28 1.01
16777216 3.09 2466.29 1.00
double RMW Cache Test
C Size Nanosec MB/sec % Chnge
------- ------- ------- -------
256 0.62 12209.84 1.00
256 0.62 12212.67 1.00
384 0.62 12211.04 1.00
384 0.62 12208.95 1.00
512 0.62 12211.87 1.00
512 0.62 12210.71 1.00
768 0.48 15825.75 0.77
768 0.50 15349.91 1.03
1024 0.48 15759.09 0.97
1024 0.51 14929.36 1.06
1536 0.47 16300.53 0.92
1536 0.45 17021.14 0.96
2048 0.44 17331.27 0.98
2048 0.48 15900.51 1.09
3072 0.46 16591.49 0.96
3072 0.44 17326.56 0.96
4096 0.43 17811.72 0.97
4096 0.43 17812.72 1.00
6144 0.42 17976.43 0.99
6144 0.42 17977.75 1.00
8192 0.47 16065.23 1.12
8192 0.43 17915.50 0.90
12288 0.43 17551.04 1.02
12288 0.44 17420.38 1.01
16384 0.42 18185.73 0.96
16384 0.42 18142.81 1.00
24576 0.47 16246.48 1.12
24576 0.42 18210.98 0.89
32768 0.46 16532.44 1.10
32768 0.42 18166.19 0.91
49152 0.43 17571.23 1.03
49152 0.44 17201.28 1.02
65536 0.43 17799.25 0.97
65536 0.44 17513.32 1.02
98304 0.46 16514.74 1.06
98304 0.42 18140.35 0.91
131072 0.42 18162.94 1.00
131072 0.45 16904.79 1.07
196608 0.42 18172.95 0.93
196608 0.43 17702.74 1.03
262144 0.46 16730.77 1.06
262144 0.44 17254.95 0.97
393216 0.42 18181.15 0.95
393216 0.42 18129.23 1.00
524288 0.45 17125.48 1.06
524288 0.43 17891.14 0.96
786432 0.42 18181.77 0.98
786432 0.43 17932.65 1.01
1048576 0.44 17260.01 1.04
1048576 0.43 17873.87 0.97
1572864 0.46 16603.70 1.08
1572864 0.45 16825.95 0.99
2097152 0.46 16720.90 1.01
2097152 0.46 16440.88 1.02
3145728 0.46 16414.83 1.00
3145728 0.44 17182.97 0.96
4194304 0.68 11185.53 1.54
4194304 0.67 11419.15 0.98
6291456 1.20 6360.79 1.80
6291456 1.20 6377.23 1.00
8388608 1.45 5274.40 1.21
8388608 1.45 5272.67 1.00
12582912 1.56 4877.82 1.08
12582912 1.56 4879.69 1.00
16777216 1.58 4821.30 1.01
16777216 1.58 4816.21 1.00
Using the program bandwidth, the measurements are:
# ./bandwidth64 This is bandwidth version 0.16. Copyright (C) 2005,2007-2009 by Zack T Smith. This software is covered by the GNU Public License. It is provided AS-IS, use at your own risk. See the file COPYING for more information. Using 64- and 128-bit data transfers. CPU speed is 2403 MHz. L1 theoretical max is 19224 MB/second (one 64-bit access/cycle). Sequential read (64-bit, size=256, loops = 4194304) 18281.1 MB/sec Sequential read (64-bit, size=512, loops = 2097152) 18303.4 MB/sec Sequential read (64-bit, size=1024, loops = 1048576) 18302.7 MB/sec Sequential read (64-bit, size=2048, loops = 524288) 18304.7 MB/sec Sequential read (64-bit, size=4096, loops = 262144) 18297.8 MB/sec Sequential read (64-bit, size=8192, loops = 131072) 18300.4 MB/sec Sequential read (64-bit, size=16384, loops = 65536) 18300.7 MB/sec Sequential read (64-bit, size=32768, loops = 32768) 17858.1 MB/sec Sequential read (64-bit, size=65536, loops = 16384) 13575.7 MB/sec Sequential read (64-bit, size=131072, loops = 8192) 13600.7 MB/sec Sequential read (64-bit, size=262144, loops = 4096) 13593 MB/sec Sequential read (64-bit, size=524288, loops = 2048) 13590.5 MB/sec Sequential read (64-bit, size=1048576, loops = 1024) 13599.5 MB/sec Sequential read (64-bit, size=2097152, loops = 512) 13474.9 MB/sec Sequential read (64-bit, size=4194304, loops = 256) 8104.54 MB/sec Sequential read (64-bit, size=8388608, loops = 128) 5359.15 MB/sec Sequential read (64-bit, size=16777216, loops = 64) 5037.86 MB/sec Sequential write (64-bit, size=256, loops = 4194304) 17747.6 MB/sec Sequential write (64-bit, size=512, loops = 2097152) 17747 MB/sec Sequential write (64-bit, size=1024, loops = 1048576) 17693 MB/sec Sequential write (64-bit, size=2048, loops = 524288) 17738.4 MB/sec Sequential write (64-bit, size=4096, loops = 262144) 17747.9 MB/sec Sequential write (64-bit, size=8192, loops = 131072) 17741.4 MB/sec Sequential write (64-bit, size=16384, loops = 65536) 17739.3 MB/sec Sequential write (64-bit, size=32768, loops = 32768) 17678.3 MB/sec Sequential write (64-bit, size=65536, loops = 16384) 11491.5 MB/sec Sequential write (64-bit, size=131072, loops = 8192) 11499.9 MB/sec Sequential write (64-bit, size=262144, loops = 4096) 11494.9 MB/sec Sequential write (64-bit, size=524288, loops = 2048) 11505.1 MB/sec Sequential write (64-bit, size=1048576, loops = 1024) 11913.1 MB/sec Sequential write (64-bit, size=2097152, loops = 512) 10521.3 MB/sec Sequential write (64-bit, size=4194304, loops = 256) 5929.15 MB/sec Sequential write (64-bit, size=8388608, loops = 128) 2696.65 MB/sec Sequential write (64-bit, size=16777216, loops = 64) 2457.62 MB/sec Sequential read (128-bit, size=256, loops = 4194304) 36562.3 MB/sec Sequential read (128-bit, size=512, loops = 2097152) 36606.7 MB/sec Sequential read (128-bit, size=1024, loops = 1048576) 36605.4 MB/sec Sequential read (128-bit, size=2048, loops = 524288) 36606.7 MB/sec Sequential read (128-bit, size=4096, loops = 262144) 36604.1 MB/sec Sequential read (128-bit, size=8192, loops = 131072) 36606.7 MB/sec Sequential read (128-bit, size=16384, loops = 65536) 36605.4 MB/sec Sequential read (128-bit, size=32768, loops = 32768) 34464.2 MB/sec Sequential read (128-bit, size=65536, loops = 16384) 16964.6 MB/sec Sequential read (128-bit, size=131072, loops = 8192) 16974.7 MB/sec Sequential read (128-bit, size=262144, loops = 4096) 16950.8 MB/sec Sequential read (128-bit, size=524288, loops = 2048) 16916.4 MB/sec Sequential read (128-bit, size=1048576, loops = 1024) 16886 MB/sec Sequential read (128-bit, size=2097152, loops = 512) 17404.3 MB/sec Sequential read (128-bit, size=4194304, loops = 256) 10495.9 MB/sec Sequential read (128-bit, size=8388608, loops = 128) 5865.47 MB/sec Sequential read (128-bit, size=16777216, loops = 64) 5442.99 MB/sec Sequential write (128-bit (bypasses caches), size=256, loops = 4194304) 3745.58 MB/sec Sequential write (128-bit (bypasses caches), size=4096, loops = 262144) 4645.15 MB/sec Sequential write (128-bit (bypasses caches), size=65536, loops = 16384) 4644.14 MB/sec Sequential write (128-bit (bypasses caches), size=1048576, loops = 1024) 4640.56 MB/sec Sequential write (128-bit (bypasses caches), size=16777216, loops = 64) 4638.88 MB/sec Framebuffer resolution: 1024x768, 16bpp Framebuffer memory sequential read 11.0282 MB/sec Framebuffer memory sequential write 271.02 MB/sec Library: memset 4510.88 MB/sec Library: memcpy 2481.12 MB/sec Performance of my_bzeroSSE2: 4635.55 MB/sec
