WindowsXP x64 and Flight Simulator 9: UPDATE!
Posted: Tue Jan 10, 2006 9:42 pmI have posted the following in 2 areas of this forum in order to make sure everyone who may be interested reads it. Because this report is a bit long I broke it down to several sequential posts below.
I recently tested WindowsXP Pro x64 to see how it compared to the 32bit WindowsXP Pro. The results were interesting. The test bed was as follows:
Asus A8V Deluxe
A64 3700+ San Diego core overclocked = 2752 MHz = (11x 250) @ 1000x2 HT
ATI AIW x800 AGP @ 520 GPU - 520 MEM
2GB OCZ rated @ 2-3-2-5, running 2.5-3-3-5 with 5:3 divider (211.6 MHz)
WindowsXP Pro corporate 32bit - and - WindowsXP Pro corporate 64bit: in dual boot configuration.
What I found was that both operating systems have their advantages and disadvantages. Individual benchmarks of the memory and CPU remained the same for both operating systems however Windows x64 will enable the full abilities of the motherboard in the CPU-Memory/cache. A 9000+ point increase in the Sandra cache/memory benchmark was found. That advantage allows for a significant increase in throughput performance between the memory and CPU.
Total (final result) 3Dmark benchmark numbers remained the same for both operating systems, however, the x64 total result was comprised of LOWER scores for the 3D load tests and HIGHER scores for the CPU load tests. In other words, the x64 operating system increases the memory/CPU performance and lowers the 3D. The result is a total (final result) benchmark score which is equal for each operating system.
Benchmarks are good for gauging a system but are not real world performance tests. I installed FS9 and imported my original Flight Simulator 9 folder (and all configuration files) from Windows x32 into the Windows x64 install to assure both test beds were equal. (same textures, scenery, aircraft, add-on, etc). I then created and saved a flight that had the same aircraft, flight path, traffic, weather, time, etc, conditions and ran a 15 minute flight test on each operating system. All display settings were equal for each test and frames are locked @ 28:
PMDG 747 with Nicks AI Exhaust
KSEA @ 12noon (very busy) with 100% UT traffic running
WEATHER set to: REAL WORLD which produced very heavy cloud cover in both test flights
NOTE: My graphics card was set to 4x FSAA and 16x AF and 100% high quality for all other settings. I set up the test for ultra high quality and extreamly high 3D load in FS9 in order to force severe frame rate drops and note the differences.
Results:
Flight Simulator frame rates were equal in both Windows x32 and Windows x64 but there was a difference!! Windows x64 ran visually smoother and loaded textures visually faster than Windows x32. At equal selected points in the flights where maximum loads were achieved, each test resulted in frame rate drops of equal value (from 28fps to 15-19fps). The reduced frame rate loads in the Windows XP 32bit tests were distinctly choppy and had much more stutter than the same load points in Windows x64. The extreme high load points had
I recently tested WindowsXP Pro x64 to see how it compared to the 32bit WindowsXP Pro. The results were interesting. The test bed was as follows:
Asus A8V Deluxe
A64 3700+ San Diego core overclocked = 2752 MHz = (11x 250) @ 1000x2 HT
ATI AIW x800 AGP @ 520 GPU - 520 MEM
2GB OCZ rated @ 2-3-2-5, running 2.5-3-3-5 with 5:3 divider (211.6 MHz)
WindowsXP Pro corporate 32bit - and - WindowsXP Pro corporate 64bit: in dual boot configuration.
What I found was that both operating systems have their advantages and disadvantages. Individual benchmarks of the memory and CPU remained the same for both operating systems however Windows x64 will enable the full abilities of the motherboard in the CPU-Memory/cache. A 9000+ point increase in the Sandra cache/memory benchmark was found. That advantage allows for a significant increase in throughput performance between the memory and CPU.
Total (final result) 3Dmark benchmark numbers remained the same for both operating systems, however, the x64 total result was comprised of LOWER scores for the 3D load tests and HIGHER scores for the CPU load tests. In other words, the x64 operating system increases the memory/CPU performance and lowers the 3D. The result is a total (final result) benchmark score which is equal for each operating system.
Benchmarks are good for gauging a system but are not real world performance tests. I installed FS9 and imported my original Flight Simulator 9 folder (and all configuration files) from Windows x32 into the Windows x64 install to assure both test beds were equal. (same textures, scenery, aircraft, add-on, etc). I then created and saved a flight that had the same aircraft, flight path, traffic, weather, time, etc, conditions and ran a 15 minute flight test on each operating system. All display settings were equal for each test and frames are locked @ 28:
PMDG 747 with Nicks AI Exhaust
KSEA @ 12noon (very busy) with 100% UT traffic running
WEATHER set to: REAL WORLD which produced very heavy cloud cover in both test flights
NOTE: My graphics card was set to 4x FSAA and 16x AF and 100% high quality for all other settings. I set up the test for ultra high quality and extreamly high 3D load in FS9 in order to force severe frame rate drops and note the differences.
Results:
Flight Simulator frame rates were equal in both Windows x32 and Windows x64 but there was a difference!! Windows x64 ran visually smoother and loaded textures visually faster than Windows x32. At equal selected points in the flights where maximum loads were achieved, each test resulted in frame rate drops of equal value (from 28fps to 15-19fps). The reduced frame rate loads in the Windows XP 32bit tests were distinctly choppy and had much more stutter than the same load points in Windows x64. The extreme high load points had