News Erste Linux-Benchmarks von AMDs Kabini-APU aufgetaucht

Na irgendwelche Makrobenchmarks für Office gibts sicher auch unter Linux. Dachte das wäre sowas. Man lernt eben nie aus. Dafür läuft der Bench grade hier durch :)
 
So, hier meine Ergebnisse. System wie unter "mein System". RAM-Takt ist aktuell 1866. Kernelversion ist 3.6.11-ck.

Die Makeopts waren in dem Fall wie vorgegeben: -O2 -fomit-frame-pointer -fforce-addr -ffast-math -Wall
Ich werde heute Abend mal versuchen, ob ich da mit einem -march=native noch etwas rauskitzeln kann..

Code:
[FONT="System"]========================================================================
   BYTE UNIX Benchmarks (Version 5.1.3)

   System: icarus: GNU/Linux
   OS: GNU/Linux -- 3.6.11-ck -- #6 SMP PREEMPT Mon Apr 1 14:08:09 CEST 2013
   Machine: x86_64 (AuthenticAMD)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: AMD A10-5800K APU with Radeon(tm) HD Graphics  (7585.8 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 1: AMD A10-5800K APU with Radeon(tm) HD Graphics  (7585.8 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 2: AMD A10-5800K APU with Radeon(tm) HD Graphics  (7585.8 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 3: AMD A10-5800K APU with Radeon(tm) HD Graphics  (7585.8 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   12:55:32 up  2:06,  2 users,  load average: 1.06, 1.16, 1.02; runlevel 

------------------------------------------------------------------------
Benchmark Run: Do Apr 18 2013 12:55:32 - 13:23:49
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       32570772.7 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     4743.9 MWIPS (9.4 s, 7 samples)
Execl Throughput                               5255.2 lps   (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        645270.5 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          212749.6 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1367778.0 KBps  (30.0 s, 2 samples)
Pipe Throughput                             1549343.0 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 220222.9 lps   (10.0 s, 7 samples)
Process Creation                              13604.2 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   6700.5 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2227.7 lpm   (60.0 s, 2 samples)
System Call Overhead                        2088834.1 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   32570772.7   2791.0
Double-Precision Whetstone                       55.0       4743.9    862.5
Execl Throughput                                 43.0       5255.2   1222.1
File Copy 1024 bufsize 2000 maxblocks          3960.0     645270.5   1629.5
File Copy 256 bufsize 500 maxblocks            1655.0     212749.6   1285.5
File Copy 4096 bufsize 8000 maxblocks          5800.0    1367778.0   2358.2
Pipe Throughput                               12440.0    1549343.0   1245.5
Pipe-based Context Switching                   4000.0     220222.9    550.6
Process Creation                                126.0      13604.2   1079.7
Shell Scripts (1 concurrent)                     42.4       6700.5   1580.3
Shell Scripts (8 concurrent)                      6.0       2227.7   3712.8
System Call Overhead                          15000.0    2088834.1   1392.6
                                                                   ========
System Benchmarks Index Score                                        1452.1

------------------------------------------------------------------------
Benchmark Run: Do Apr 18 2013 13:23:49 - 13:53:49
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       84167775.0 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    14865.0 MWIPS (9.7 s, 7 samples)
Execl Throughput                              13714.3 lps   (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        471220.7 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          144849.1 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1419028.7 KBps  (30.0 s, 2 samples)
Pipe Throughput                             4145544.3 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 285309.0 lps   (10.0 s, 7 samples)
Process Creation                              40074.6 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  17532.2 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2387.9 lpm   (60.0 s, 2 samples)
System Call Overhead                        3068883.8 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   84167775.0   7212.3
Double-Precision Whetstone                       55.0      14865.0   2702.7
Execl Throughput                                 43.0      13714.3   3189.4
File Copy 1024 bufsize 2000 maxblocks          3960.0     471220.7   1190.0
File Copy 256 bufsize 500 maxblocks            1655.0     144849.1    875.2
File Copy 4096 bufsize 8000 maxblocks          5800.0    1419028.7   2446.6
Pipe Throughput                               12440.0    4145544.3   3332.4
Pipe-based Context Switching                   4000.0     285309.0    713.3
Process Creation                                126.0      40074.6   3180.5
Shell Scripts (1 concurrent)                     42.4      17532.2   4134.9
Shell Scripts (8 concurrent)                      6.0       2387.9   3979.9
System Call Overhead                          15000.0    3068883.8   2045.9
                                                                   ========
System Benchmarks Index Score                                        2419.9[/FONT]
 
Zuletzt bearbeitet:
Wie versprochen, hier der Durchlauf mit gesetztem -march.

Die Makeopts waren also: -O2 -march=native -fomit-frame-pointer -fforce-addr -ffast-math -Wall

Die größten Änderungen liegen dabei im Bereich der parallelen Durchläufe. Allerdings ändert sich nichts grundlegend. Dafür müsste wohl der Code entsprechend angepasst werden.

Code:
[FONT="System"]========================================================================
   BYTE UNIX Benchmarks (Version 5.1.3)

   System: icarus: GNU/Linux
   OS: GNU/Linux -- 3.6.11-ck -- #6 SMP PREEMPT Mon Apr 1 14:08:09 CEST 2013
   Machine: x86_64 (AuthenticAMD)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: AMD A10-5800K APU with Radeon(tm) HD Graphics  (7586.3 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 1: AMD A10-5800K APU with Radeon(tm) HD Graphics  (7586.3 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 2: AMD A10-5800K APU with Radeon(tm) HD Graphics  (7586.3 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   CPU 3: AMD A10-5800K APU with Radeon(tm) HD Graphics  (7586.3 bogomips)
          Hyper-Threading, x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
   17:56:33 up 41 min,  2 users,  load average: 0.15, 0.20, 0.37; runlevel 

------------------------------------------------------------------------
Benchmark Run: Do Apr 18 2013 17:56:33 - 18:25:03
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       31318804.8 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     4421.0 MWIPS (10.2 s, 7 samples)
Execl Throughput                               5334.9 lps   (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        660033.8 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          213574.4 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1395624.5 KBps  (30.0 s, 2 samples)
Pipe Throughput                             1556344.4 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 220896.4 lps   (10.0 s, 7 samples)
Process Creation                              13612.2 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   6729.1 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2229.3 lpm   (60.0 s, 2 samples)
System Call Overhead                        2080335.8 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   31318804.8   2683.7
Double-Precision Whetstone                       55.0       4421.0    803.8
Execl Throughput                                 43.0       5334.9   1240.7
File Copy 1024 bufsize 2000 maxblocks          3960.0     660033.8   1666.8
File Copy 256 bufsize 500 maxblocks            1655.0     213574.4   1290.5
File Copy 4096 bufsize 8000 maxblocks          5800.0    1395624.5   2406.2
Pipe Throughput                               12440.0    1556344.4   1251.1
Pipe-based Context Switching                   4000.0     220896.4    552.2
Process Creation                                126.0      13612.2   1080.3
Shell Scripts (1 concurrent)                     42.4       6729.1   1587.1
Shell Scripts (8 concurrent)                      6.0       2229.3   3715.5
System Call Overhead                          15000.0    2080335.8   1386.9
                                                                   ========
System Benchmarks Index Score                                        1447.4

------------------------------------------------------------------------
Benchmark Run: Do Apr 18 2013 18:25:03 - 18:54:13
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       84231856.0 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                    15826.1 MWIPS (9.4 s, 7 samples)
Execl Throughput                              13801.0 lps   (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        516573.5 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks          145881.1 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks       1474096.7 KBps  (30.0 s, 2 samples)
Pipe Throughput                             4133654.9 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                 284081.9 lps   (10.0 s, 7 samples)
Process Creation                              40093.5 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                  17557.4 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                   2329.6 lpm   (60.0 s, 2 samples)
System Call Overhead                        3155814.4 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   84231856.0   7217.8
Double-Precision Whetstone                       55.0      15826.1   2877.5
Execl Throughput                                 43.0      13801.0   3209.5
File Copy 1024 bufsize 2000 maxblocks          3960.0     516573.5   1304.5
File Copy 256 bufsize 500 maxblocks            1655.0     145881.1    881.5
File Copy 4096 bufsize 8000 maxblocks          5800.0    1474096.7   2541.5
Pipe Throughput                               12440.0    4133654.9   3322.9
Pipe-based Context Switching                   4000.0     284081.9    710.2
Process Creation                                126.0      40093.5   3182.0
Shell Scripts (1 concurrent)                     42.4      17557.4   4140.9
Shell Scripts (8 concurrent)                      6.0       2329.6   3882.7
System Call Overhead                          15000.0    3155814.4   2103.9
                                                                   ========
System Benchmarks Index Score                                        2461.5[/FONT]
 
Ich tippe mal darauf:
Mal ehrlich wenn sich das Bewahrheit mit der Leistung und die APU hat nur 25watt TDP, ist ja eigentlich bombastisch.
 
@all
Mal ehrlich wenn sich das bewahrheitet mit der Leistung, und die APU hat nur 25watt TDP ist ja eigentlich bombastisch.
das ist gemeint
 
Ich denke:
Mal ehrlich, wenn sich das bewahrheitet mit der Leistung und die APU hat nur 25watt TDP, ist ja eigentlich bombastisch.
:D
 
Zuletzt bearbeitet:
Wow....meine Rechtschreibung beweist eindeutig das ich um diese Uhrzeit nicht mehr schreiben sollte. *lol*
 
Zurück
Oben Unten