Dans la série des c’est moi qui ai la plus grosse…
13 sujets de 1 à 13 (sur un total de 13)
-
Mod La société qui distribue la carte samantha a publié un bench de performance mémoire (le test stream, qui d’aprés l’université est le standard… pourquoi pas, ca reste du benchmark synthétique pas orienté application mais ca donne un aperçu quand même…).
Je vous propose de réaliser le test sur tout les différents matériels dont nous disposons actuellement en partant de nos vieux coucous jusqu’aux dernières machines succeptibles d’accueillir nos OS préférés.
Pegasos I et II, Amiga 060@50, PPC604e, Aone et uAOne manquent encore à l’appel. Pour ceux qui se sentent le courage de faire le test, voici le code source du site de l’université de virginie :
Un simple gcc -O stream.c donnera un fantatisque a.out prêt pour une exécution.
Qui aura la plus grosse ?
Voila pour l’Efika, qui en a une petite :
STREAM version $Revision: 5.7 $
This system uses 8 bytes per DOUBLE PRECISION word.
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
Printing one line per active thread….
Your clock granularity/precision appears to be 2 microseconds.
Each test below will take on the order of 108855 microseconds.
(= 54427 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
WARNING — The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
Function Rate (MB/s) Avg time Min time Max time
Copy: 197.2363 0.1623 0.1622 0.1625
Scale: 188.5170 0.1704 0.1697 0.1755
Add: 185.8031 0.2612 0.2583 0.2831
Triad: 183.5045 0.2616 0.2616 0.2618
Solution Validates
Pegasos II G4 :
STREAM version $Revision: 5.7 $
This system uses 8 bytes per DOUBLE PRECISION word.
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
Printing one line per active thread….
Your clock granularity/precision appears to be 2 microseconds.
Each test below will take on the order of 88895 microseconds.
(= 44447 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
WARNING — The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
Function Rate (MB/s) Avg time Min time Max time
Copy: 353.7398 0.0929 0.0905 0.0971
Scale: 281.0148 0.1162 0.1139 0.1284
Add: 260.4888 0.1872 0.1843 0.1935
Triad: 263.7623 0.1846 0.1820 0.1942
Solution Validates
—
/me se demande si ça vaut vraiment quelque chose, mais bon.
Only Amiga makes it possible !
Peg II G4 avec – O
STREAM version $Revision: 5.7 $
This system uses 8 bytes per DOUBLE PRECISION word.
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
Printing one line per active thread….
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 76645 microseconds.
(= 76645 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
WARNING — The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
Function Rate (MB/s) Avg time Min time Max time
Copy: 369.4639 0.0867 0.0866 0.0868
Scale: 362.9517 0.0882 0.0882 0.0883
Add: 364.7142 0.1317 0.1316 0.1317
Triad: 367.3964 0.1357 0.1306 0.1754
Solution Validates
Pareil avec -mcpu=7450 -O3
STREAM version $Revision: 5.7 $
This system uses 8 bytes per DOUBLE PRECISION word.
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
Printing one line per active thread….
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 75796 microseconds.
(= 75796 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
WARNING — The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
Function Rate (MB/s) Avg time Min time Max time
Copy: 371.9241 0.0861 0.0860 0.0862
Scale: 364.6478 0.0878 0.0878 0.0879
Add: 367.9907 0.1465 0.1304 0.2736
Triad: 370.7624 0.1296 0.1295 0.1303
Solution Validates
25cm au repos.
SSQ> nan,ca CMB
Mod Powerbook G4 1,5Ghz avec un gcc -O3 -mcpu=7450…
Si l’on prend les chiffres du PEGII que l’on multiplie par la différence de fréquence (x1,5) entre les deux machines, le résultat est cohérent avec l’augmentation de fréquence.
370 x 1,5 = 555.
STREAM version $Revision: 5.7 $
This system uses 8 bytes per DOUBLE PRECISION word.
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
Printing one line per active thread….
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 53287 microseconds.
(= 53287 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
WARNING — The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
Function Rate (MB/s) Avg time Min time Max time
Copy: 498.3946 0.0644 0.0642 0.0649
Scale: 498.9095 0.0645 0.0641 0.0653
Add: 561.1345 0.0865 0.0855 0.0894
Triad: 562.4924 0.0858 0.0853 0.0864
Solution Validates
AmigaOneG4
8.RAM Disk:> stream.os4
STREAM version $Revision: 5.7 $
This system uses 8 bytes per DOUBLE PRECISION word.
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
Printing one line per active thread….
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 132620 microseconds.
(= 132620 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
WARNING — The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
Function Rate (MB/s) Avg time Min time Max time
Copy: 173.6893 0.1845 0.1842 0.1852
Scale: 216.4544 0.1481 0.1478 0.1487
Add: 188.9569 0.2545 0.2540 0.2561
Triad: 189.5681 0.2536 0.2532 0.2543
Solution Validates
Efika : MPC5200B / 396Mhz.
STREAM version $Revision: 5.7 $
This system uses 8 bytes per DOUBLE PRECISION word.
Array size = 2000000, Offset = 0
Total memory required = 45.8 MB.
Each test is run 10 times, but only
the *best* time for each is used.
Printing one line per active thread….
Your clock granularity/precision appears to be 2 microseconds.
Each test below will take on the order of 108983 microseconds.
(= 54491 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
WARNING — The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
Function Rate (MB/s) Avg time Min time Max time
Copy: 197.3349 0.1622 0.1622 0.1623
Scale: 192.4533 0.1664 0.1663 0.1665
Add: 194.3288 0.2795 0.2470 0.3628
Triad: 191.7560 0.2633 0.2503 0.3118
Solution Validates
13 sujets de 1 à 13 (sur un total de 13)
- Vous devez être connecté pour répondre à ce sujet.
› Forums › AmigaOS, MorphOS et AROS › Guéguerres › Dans la série des c’est moi qui ai la plus grosse…
