Fujitsu SPARC64 VII (Jupiter)

Configuration

Fujitsu SPARC64 VII (Jupiter), 2520 MHz, 4 cores, 65 nm, 600 MTr, 445 mm2, 8 MB L3, 135 W. SPARC Enterprise M3000. 32 GB DDR2 SDRAM

CPU / Core

L1 Data cache = 64 KB. 2-Way, 64-byte lines
L1 Instruction cache = 64 KB, 2-Way, 64-byte line.

L2 cache size = 5 MB, 10-WAY. 256-byte lines, (4x 64-byte sublines)

SMT2
Memory DIMMs are connected via Jupiter Memory Controller (JPC) chip

L1 Data Cache Latency = 4 cycles for simple access via pointer
L1 Data Cache Latency = 6 cycles for access with complex address calculation (size_t n, *p; n = p[n]).

L2 Cache Latency = 31.5 cycles

RAM Latency = 32 cycles + 207 ns

8 KB pages mode Solaris

MicroData TLB: 32 entries, Miss penalty = 9 cycles. Parallel miss: 1.3 cycles per access

TLB: 2048 entry, Miss penalty = 140 cycles. Parallel miss: 180 cycles per access

  Size        Latency       Increase   Description

  64 K    4                            
 128 K    18                      14   + 28 (L2)         
 256 K    25                       7   
 512 K    33                       8   + 9 (TLB L1 miss)
   1 M    37                       4   
   2 M    39                       2   
   4 M    40                       1	
   8 M    41 +  81 ns     1 +  81 ns   + 207 ns
  16 M    41 + 145 ns       +  64 ns   
  32 M   109 + 177 ns    68 +  32 ns   + 140 (TLB L2 miss)
  64 M   146 + 192 ns    37 +  15 ns
 128 M   164 + 200 ns    18 +   8 ns   
 256 M   176 + 203 ns    12 +   3 ns   
 512 M   194 + 205 ns    18 +   2 ns
1024 M   200 + 511 ns     6 + 306 ns   + 620 ns 
2048 M   200 + 715 ns     1 + 200 ns

4 MB pages mode Solaris

MicroData TLB: 32 entries, Miss penalty = 9 cycles.

  Size        Latency       Increase   Description

  64 K    4                            
 128 K    18                      14   + 28 (L2)         
 256 K    25                       7   
 512 K    28                       3   
   1 M    30                       2   
   2 M    31                       1   
   4 M    32                       1	
   8 M    32 +  81 ns       +  81 ns   + 207 ns
  16 M    32 + 145 ns       +  64 ns   
  32 M    32 + 177 ns       +  32 ns
  64 M    32 + 192 ns       +  15 ns
 128 M    32 + 200 ns       +   8 ns   
 256 M    37 + 203 ns     5 +   3 ns   + 9 (TLB miss)
 512 M    39 + 205 ns     2 +   2 ns
1024 M    40 + 206 ns     1 +   1 ns   
2048 M    41 + 207 ns     1 +   1 ns

MISC

Branch misprediction penalty = 10-14 cycles

L1 B/W (Parallel Random Read) = 0.71 cycles per one access

L2->L1 B/W (Parallel Random Read) = 9.0 cycles per cache line (64 bytes)
L2->L1 B/W (Read, 64 bytes step) = 10 cycles per cache line (64 bytes)
L2 Write (Write, 64 bytes step) = 12 per write

RAM Read B/W (Parallel Random Read) = 55 ns / read (64-256 bytes read ?)
RAM Read B/W (Read, 8-256 Bytes step) = 4.3-4.7 GB/s
RAM Write B/W (Write, 8 Bytes step, full) = 2.5 GB/s

7-Zip Benchmark

Notes:

7z b -mm=* : MIPS and Effectiveness values are normalized with AMD K8 cpu.


### Solaris 11 Zone
## gcc-5 -m32 -O3

7-Zip (a) 16.04 : Copyright (c) 1999-2016 Igor Pavlov : 2016-10-04
p7zip Version 16.04 (locale=en_US.UTF-8,Utf16=on,HugeFiles=on,32 bits,8 CPUs BE)

freq= 2525
BE
CPU Freq:  2524  2525  2525  2525  2524  2525  2525  2525  2525

RAM size:   32768 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   2529   2526   100   100
CPU                      100   2523   2526   100   100
CPU                      100   2529   2526   100   100

LZMA:x1           7342   100   2686   2684   107   107
                 25302   100   2061   2060    82    82
LZMA:x5:mt1       1418   100   1774   1772    70    70
                 24900   100   2098   2100    83    83
LZMA:x5:mt2       2918   201   1816   3645    72   145
                 24916   100   2098   2102    83    83
Deflate:x1       16980   100   2161   2156    86    86
                 70798   100   2198   2200    87    87
Deflate:x5        6148   100   2372   2367    94    94
                 70965   100   2201   2203    87    87
Deflate:x7        2073   100   2298   2297    91    91
                 71417   100   2216   2216    88    88
Deflate64:x5      5894   100   2547   2547   101   101
                 70821   100   2213   2215    88    88
BZip2:x1          3234   100   1956   1954    78    78
                 13389   100   1451   1451    58    58
BZip2:x5          2937   100   2455   2451    97    97
                 12259   100   2406   2406    95    95
BZip2:x5:mt2      4366   195   1871   3644    74   145
                 17015   137   2436   3339    97   133
BZip2:x7           820   100   2127   2126    84    84
                 13210   100   2591   2591   103   103
PPMD:x1           1983   100   2051   2051    81    81
                  1103   100   1300   1300    52    52
PPMD:x5           1411   100   2392   2392    95    95
                   813   100   1524   1525    60    61
Delta:4         142253   100    874    874    35    35
                152312   100    936    936    37    37
BCJ             508657   100   2083   2083    83    83
                525018   100   2149   2150    85    85
AES256CBC:1      54355   100   1337   1336    53    53
                 54454   100   1339   1338    53    53
AES256CBC:2 

CRC32:1         241702   100   1759   1760    70    70
CRC32:4         515457   100   1151   1151    46    46
CRC32:8         678754   100    921    920    37    37
CRC64           469569   100    962    962    38    38
SHA256           49818   100   1016   1016    40    40
SHA1            127905   100   1197   1197    48    48
BLAKE2sp         82086   100   1806   1806    72    72

CPU                      100   2529   2525   100   100
------------------------------------------------------
Tot:                     113   1985   2226    78    88



RAM usage:   1802 MB,  # Benchmark threads:      8


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

CPU                      796   1263  10060    50   399
CPU                      798   1263  10081    50   400
CPU                      799   1263  10085    50   400

LZMA:x1          16906   790    782   6181    31   245
                132825   791   1368  10818    54   429
LZMA:x5:mt1       5416   785    862   6767    34   269
                122611   774   1336  10340    53   410
LZMA:x5:mt2       5544   787    880   6927    35   275
                126046   794   1338  10630    53   422
Deflate:x1       75806   786   1225   9626    49   382
                350593   792   1375  10894    55   432
Deflate:x5       30284   784   1487  11660    59   463
                353500   795   1380  10975    55   436
Deflate:x7       10776   794   1504  11940    60   474
                353225   791   1386  10962    55   435
Deflate64:x5     22396   775   1249   9678    50   384
                352499   794   1388  11028    55   438
BZip2:x1         16490   799   1248   9963    50   395
                 67251   740    985   7290    39   289
BZip2:x5          6195   784    660   5170    26   205
                 32265   771    821   6333    33   251
BZip2:x5:mt2      6703   792    707   5595    28   222
                 34975   795    864   6865    34   272
BZip2:x7          3084   781   1023   7991    41   317
                 31718   765    814   6220    32   247
PPMD:x1           9618   788   1263   9948    50   395
                  6185   797    915   7284    36   289
PPMD:x5           4438   792    950   7523    38   299
                  3623   789    861   6790    34   269
Delta:4         661471   794    512   4064    20   161
                669236   798    516   4112    20   163
BCJ            2648304   780   1391  10847    55   430
               2766781   793   1428  11333    57   450
AES256CBC:1     221428   797    683   5442    27   216
                215577   799    663   5298    26   210
AES256CBC:2 

CRC32:1        1585645   795   1452  11544    58   458
CRC32:4        2664468   798    745   5947    30   236
CRC32:8        3035398   799    515   4116    20   163
CRC64          2038794   798    523   4175    21   166
SHA256          206986   798    529   4223    21   168
SHA1            529258   797    622   4954    25   197
BLAKE2sp        348175   796    963   7660    38   304

CPU                      796   1263  10061    50   399
------------------------------------------------------
Tot:                     787   1051   8268    42   328

Links

SPARC64 V at Wikipedia