Avalox
2004-03-24, 23:44:56
Hat jemand eine Idee, wie sehr sich die Sandra2004 Benchmarks der 32Bit Version mit den Ergebnissen der 64Bit Version vergleichen lassen?
Kommen ähnliche Algorithmen zum Einsatz? Die Ergebnisse gehen ganz schön auseinander.
http://www.planetamd64.com
------------------------------------
Now I'm not posting scores here to show off, or to compete with anyone...simply to show the difference between Win32 and Win64.
The setup is below.
Athlon 3000+ @ 2.10Mhz
512MB Kingston HyperX RAM @ 420Mhz @ 1:1 divider.
MSI NEO-6702 motherboard
WinXP32...all applications (AV, Firewall etc) disabled.
WinXP64...new release, straight from the MS website.
All tests were performed three times, and then an average from the scores was taken.
Filesystem and cache/memory benchmarks were not recorded, the same applies to LAN,internet, CD-ROM drives etc...as the Cache was a pain to record, and none of the others would make a difference to the scores.
The Win32 drivers were from MSI and completely up to date.
The Win64 drivers are all default ones from Microsoft
The results are below :
Win32 - CPU Arithmetric - Dhrystone ALU - 8764 MIPS
Win64 - CPU Arithmetric - Dhrystone ALU - 8941 MIPS
Win32 - CPU Arithmetric - Whetstone FPU/isse2 - 1324/4335 MFLOPS
Win64 - CPU Arithmetric - Whetstone FPU/isse2 - 3744/3979 MFLOPS
Win32 - CPU Multimedia - Integer isse2 - 6257
Win64 - CPU Multimedia - Integer isse2 - 15521
Win32 - CPU Multimedia - Floating point isse2 - 20696
Win64 - CPU Multimedia - Floating point isse2 - 22323
I considered including the memory benchmarks here...but the Win32 and the Win64 ones are identical, so there is no point.
Points to note.
Win64 is slightly faster at CPU Arithmetic ALU tests.
Win64 HAMMERS Win32 at CPU Arithmetric FPU tests
Win32 is a quite a bit faster at CPU arthimetric isse2 tests
Win64 is about two and a half times FASTER than Win32 in CPU multimedia Integer tests!
Win64 is slightly faster than Win32 in Floating point Multimedia tests.
Well, hope those results help some people....give me a shout if you want anymore tests done on this .
-BTJ
--------------------------------------------
Kommen ähnliche Algorithmen zum Einsatz? Die Ergebnisse gehen ganz schön auseinander.
http://www.planetamd64.com
------------------------------------
Now I'm not posting scores here to show off, or to compete with anyone...simply to show the difference between Win32 and Win64.
The setup is below.
Athlon 3000+ @ 2.10Mhz
512MB Kingston HyperX RAM @ 420Mhz @ 1:1 divider.
MSI NEO-6702 motherboard
WinXP32...all applications (AV, Firewall etc) disabled.
WinXP64...new release, straight from the MS website.
All tests were performed three times, and then an average from the scores was taken.
Filesystem and cache/memory benchmarks were not recorded, the same applies to LAN,internet, CD-ROM drives etc...as the Cache was a pain to record, and none of the others would make a difference to the scores.
The Win32 drivers were from MSI and completely up to date.
The Win64 drivers are all default ones from Microsoft
The results are below :
Win32 - CPU Arithmetric - Dhrystone ALU - 8764 MIPS
Win64 - CPU Arithmetric - Dhrystone ALU - 8941 MIPS
Win32 - CPU Arithmetric - Whetstone FPU/isse2 - 1324/4335 MFLOPS
Win64 - CPU Arithmetric - Whetstone FPU/isse2 - 3744/3979 MFLOPS
Win32 - CPU Multimedia - Integer isse2 - 6257
Win64 - CPU Multimedia - Integer isse2 - 15521
Win32 - CPU Multimedia - Floating point isse2 - 20696
Win64 - CPU Multimedia - Floating point isse2 - 22323
I considered including the memory benchmarks here...but the Win32 and the Win64 ones are identical, so there is no point.
Points to note.
Win64 is slightly faster at CPU Arithmetic ALU tests.
Win64 HAMMERS Win32 at CPU Arithmetric FPU tests
Win32 is a quite a bit faster at CPU arthimetric isse2 tests
Win64 is about two and a half times FASTER than Win32 in CPU multimedia Integer tests!
Win64 is slightly faster than Win32 in Floating point Multimedia tests.
Well, hope those results help some people....give me a shout if you want anymore tests done on this .
-BTJ
--------------------------------------------