ARM Cortex-A57

AMD Opteron A1170 (ARM Cortex-A57), 2.0 GHz, 28 nm. RAM: 16 GB. (Probably it's SoftIron Overdrive 3000 server, DDR3 RDIMM).

L1 Data cache = 32 KB, 64 B/line, 2-WAY.
L1 Instruction cache = 48 KB, 64 B/line, 3-WAY.
L2 Cache = 1 MB (per 2 cores), 64 B/line, 16-WAY.
L3 Cache = 8 MB (per 8 cores), 64 B/line, ?-WAY.

L1 Data Cache Latency = 4 cycles for simple access via pointer
L1 Data Cache Latency = 5 cycles for access with complex address calculation (size_t n, *p; n = p[n]).
L2 Cache Latency = 18 cycles
L3 Cache Latency = 60 cycles
RAM Latency = 60 cycles + 124 ns

64 KB pages mode (64-bit Linux)

Data TLB L1: 32 items. full-assoc. Miss penalty = 7 cycles. Parallel miss: 8 cycles per access
Data TLB L2: 1024 items. 4-way. Miss penalty = 30-36. Parallel miss: 55-60 cycles per access (including L2 cache access latency)

  Size       Latency        Increase   Description

  32 K     4                           
  64 K    12               8           + 14 (L2)        
 128 K    15               3    
 256 K    16               1
 512 K    17               1   
   1 M    19               2
   2 M    39              20           + 42 (L3)
   4 M    51              12           + 7 (L1 TLB miss)        
   8 M    58 +  9 ns       7 +  9 ns   
  16 M    63 + 60 ns       5 + 51 ns   + 124 ns (RAM)
  32 M    66 + 92 ns       3 + 32 ns
  64 M    67 +108 ns       1 + 16 ns
 128 M    82 +116 ns      15 +  8 ns   + 30 (L2 TLB miss)
 256 M    89 +120 ns       7 +  4 ns
 512 M    94 +122 ns       5 +  2 ns
   1 G    98 +123 ns       4 +  1 ns
   2 G   102 +124 ns       4 +  1 ns
   4 G   109 +124 ns       7 
   8 G   123 +124 ns      14

MISC

Branch misprediction penalty = 16-19 cycles

64-bytes range cross penalty = 1 cycle
page cross penalty - ? cycles

L1 B/W (Parallel Random Read) = 1 cycle per one access

L2->L1 B/W (Parallel Random Read) = 6.3 cycles per cache line
L2->L1 B/W (Read, 64 bytes step) = 6.1 cycles per cache line
L2->L1 B/W (Read, 64 Bytes step - pointer chasing) = 8 cycles per cache line (HW prefetch)
L2 Write (Write, 64 bytes step) = 6 cycles per write

L3->L1 B/W (Parallel Random Read) = 13 cycles per cache line
L3->L1 B/W (Read, 64 bytes step) = 12 cycles per cache line
L3->L1 B/W (Read, 64 Bytes step - pointer chasing) = 12 cycles per cache line (HW prefetch)
L3 Write (Write, 64 bytes step) = 13 cycles per write

RAM Read B/W (Parallel Random Read) = 30 ns / access line
RAM Read B/W (Read, 8-64 Bytes step) = 4500-4700 MB/s
RAM Read B/W (Read, 8 Bytes step - pointer chasing) = 3700 MB/s (HW prefetch)
RAM Read B/W (Read, 16-64 Bytes step - pointer chasing) = 4300-4400 MB/s (HW prefetch)
RAM Write B/W (Write, 8 Bytes step) = 6500 MB/s

7-Zip Benchmark

7z b -mm=* -mmt1

Note: MIPS and Effectiveness values are normalized with AMD K8 cpu.

## Linux-4.1.12

32-bit, arm32


# gcc-6 -O2 -march=armv7-a


7-Zip (a) [32] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=ru_RU.UTF-8,Utf16=on,HugeFiles=on,32 bits,8 CPUs LE)

LE
CPU Freq:  1998  1997  1998  1999  1999  1999  1999  1999  1999

RAM size:   16252 MB,  # CPU hardware threads:   8
RAM usage:    225 MB,  # Benchmark threads:      1


Method           Speed Usage    R/U Rating   E/U Effec
                 KiB/s     %   MIPS   MIPS     %     %

CPU                      100   1999   1999
CPU                      100   1999   1999
CPU                      100   1999   1999   100   100

LZMA:x1           7811   100   2856   2856   143   143
                 30851   100   2512   2512   126   126
LZMA:x5:mt1       1845   100   2305   2305   115   115
                 30274   100   2554   2554   128   128
LZMA:x5:mt2       3595   183   2448   4492   122   225
                 30278   100   2554   2554   128   128
Deflate:x1       20144   100   2558   2558   128   128
                 73879   100   2296   2295   115   115
Deflate:x5        6023   100   2319   2319   116   116
                 74046   100   2299   2299   115   115
Deflate:x7        2318   100   2569   2569   128   128
                 74507   100   2312   2312   116   116
Deflate64:x5      5746   100   2484   2483   124   124
                 74535   100   2331   2331   117   117
BZip2:x1          3573   100   2159   2159   108   108
                 15642   100   1696   1696    85    85
BZip2:x5          3150   100   2629   2629   132   132
                 12890   100   2530   2530   127   127
BZip2:x5:mt2      5790   194   2489   4832   125   242
                 19642   155   2492   3855   125   193
BZip2:x7           913   100   2366   2366   118   118
                 12990   100   2548   2547   127   127
PPMD:x1           2176   100   2251   2251   113   113
                  1476   100   1739   1739    87    87
PPMD:x5           1636   100   2774   2774   139   139
                  1153   100   2162   2162   108   108
Delta:4         362427   100   2227   2227   111   111
                310341   100   1907   1907    95    95
BCJ             565826   100   2318   2318   116   116
                565895   100   2318   2318   116   116
AES256CBC:1      74489   100   1831   1831    92    92
                 71971   100   1769   1769    88    88
AES256CBC:2 

CRC32:1         244029   100   1777   1777    89    89
CRC32:4         698817   100   1560   1560    78    78
CRC32:8         978912   100   1327   1327    66    66
CRC64           509734   100   1044   1044    52    52
SHA256          106661   100   2176   2176   109   109
SHA1            232187   100   2173   2173   109   109
BLAKE2sp         87190   100   1918   1918    96    96

CPU                      100   1999   1999
------------------------------------------------------
Tot:                     111   2376   2654   119   133

64-bit

# gcc-6 -O2 -march=armv8-a

7-Zip (a) [64] 16.02 : Copyright (c) 1999-2016 Igor Pavlov : 2016-05-21
p7zip Version 16.02 (locale=ru_RU.UTF-8,Utf16=on,HugeFiles=on,64 bits,8 CPUs LE)

LE
CPU Freq:  1998  1998  1998  1998  1999  1999  1998  1999  1999

RAM size:   16252 MB,  # CPU hardware threads:   8
RAM usage:    225 MB,  # Benchmark threads:      1


Method           Speed Usage    R/U Rating   E/U Effec
                 KiB/s     %   MIPS   MIPS     %     %

CPU                      100   1999   1999
CPU                      100   1999   1999
CPU                      100   1999   1999   100   100

LZMA:x1           8380   100   3064   3064   153   153
                 23488   100   1912   1912    96    96
LZMA:x5:mt1       1964   100   2455   2455   123   123
                 23370   100   1971   1971    99    99
LZMA:x5:mt2       3688   175   2638   4608   132   230
                 23411   100   1975   1975    99    99
Deflate:x1       20483   100   2601   2601   130   130
                 73839   100   2294   2294   115   115
Deflate:x5        6874   100   2647   2647   132   132
                 74132   100   2301   2301   115   115
Deflate:x7        2395   100   2654   2654   133   133
                 74614   100   2315   2315   116   116
Deflate64:x5      6495   100   2807   2807   140   140
                 74748   100   2338   2338   117   117
BZip2:x1          3531   100   2134   2134   107   107
                 18741   100   2032   2032   102   102
BZip2:x5          3105   100   2591   2591   130   130
                 15074   100   2959   2959   148   148
BZip2:x5:mt2      5753   194   2473   4801   124   240
                 24853   168   2912   4878   146   244
BZip2:x7           870   100   2256   2256   113   113
                 15177   100   2976   2976   149   149
PPMD:x1           2496   100   2582   2582   129   129
                  1723   100   2030   2030   102   102
PPMD:x5           1818   100   3081   3081   154   154
                  1394   100   2613   2613   131   131
Delta:4         371710   100   2284   2284   114   114
                354386   100   2177   2177   109   109
BCJ             570403   100   2336   2336   117   117
                570263   100   2336   2336   117   117
AES256CBC:1      72218   100   1775   1775    89    89
                 73995   100   1819   1819    91    91
AES256CBC:2 

CRC32:1         244024   100   1777   1776    89    89
CRC32:4         663477   100   1481   1481    74    74
CRC32:8        1039715   100   1410   1410    71    71
CRC64           678760   100   1390   1390    70    70
SHA256          132415   100   2701   2701   135   135
SHA1            215661   100   2019   2019   101   101
BLAKE2sp        132259   100   2910   2910   146   146

CPU                      100   1999   1999
------------------------------------------------------
Tot:                     111   2337   2615   118   131