PerfectDisk 7 testing

[Speedo]

I took sitboy up on his offer for a copy of PD7 and have been doing some testing with it, figured I'd share the results.

I picked out 4 games to do testing with: Battlefield 2, F.E.A.R., Elder Scrolls 4: Oblivion, and Falcon 4.0: Allied Force.

I started out by analyzing my drive (array) with both the windows defragmenter and PD7 to see what kind of shape it was it. Windows defragmenter reported 6% total fragmentation, 13% file fragmentation, and said that no action was necessary. PD7 didn't have a "total fragmentation" stat that I could find, but reported only 2% file fragmentation, and recommended defragmenting. It's only been 2-3 weeks since I last defragmented, so it makes sense that my array is in fairly good shape.

I rebooted and ran the first batch of tests with that level of fragmentation. Then, I ran the standard windows defragmenter, rebooted and ran a second series of benchmarks. After that I opened up PD7 and did it's "Smart placement" defrag, follow by a reboot and benchmarks. After reading through the PD7 documentation, I discovered that the program actually runs two different defrag modes, "online" defrag, done normally while windows is running, and "offline" defrag, which usually must be scheduled to be done on reboot - it will run before windows starts, similar to boot-up scandisk checks, which allows it to defrag the windows pagefile and system files that can't be optimized by an online scan. So, I performed an offline scan and a final series of benchmarks.

All the testing was done via a 2 minute benchmark using FRAPS 2.6.4 build 4979. This isn't meant to be ultra precise or scientific testing, just to get a general idea of the gaming performance benefits of PD7.

First the load time results. In each case this is the time from when the "load" button is pressed until the 3D environment is displayed. See the performance results below for exactly which levels are being loaded in each game.

Battlefield 2 Performance results:

I loaded the the Strike a Karkand map in single player, and ran the benchmark while fighting in the main street as you enter the city from the US spawnpoints. One of the more intense areas in BF2 single player, with tons of effects, bullets ricocheting off of walls, explosions, vehicles popping smoke, etc. In case you're wondering why my average FPS is relatively low, it's because in BF2 single player I run a large number of bots - 75 in this case.

F.E.A.R. Performance results:

I loaded a savegame that I had near in the "Flight" level of the game. While the level itself isn't overly demanding graphics-wise (in the parking garage of a building), it involves some of the largest and most intense firefights in the game.

TES4: Oblivion Performance results:

Again I loaded a savegame that I had handy. Starting on horseback outside the Imperial City, I headed out from the city and south along the main road for the duration of the benchmark, which takes you through both fairly open areas and the more demanding heavily forested areas.

Falcon 4.0: Allied Force Performance results:

I opened a saved Korea 2010 campaign, Day 1 0945, and joined a flight which was approaching the FLOT. I engaged the combat autopilot and let it fly the mission for the duration of the benchmark, which involved a bombing run against ground troops and a brief engagement with some nearby Migs.

My system specs:
Windows XP Pro SP2
Athlon64x2 4400+ (2.2Ghz, 2x 1MB L2 cache)
Asus A8N-SLI Premium (BIOS ver 1009, nvidia chipset drivers 6.70)
2x 1GB Corsair XMS DDR400 RAM at 3-3-3-8-1T
eVGA GF7800GTX (Forceware 84.21)
Creative X-fi XtremeMusic (Creative 2.07.0004 drivers)
2 Samsung HD080HJ (80GB, SATA 3.0 Gb/s, 8MB cache) configured in nForce4 RAID0
Sony DRU-720A DVD-R/RW, DDU1621 DVD-ROM, CDU5215 CD-ROM
Seasonic S12-600W PSU

 

[Genocidicbunny]

so what does this show overall? Falcon had reduced performance after the defrags, Bf and oblivion both had increased performance with the PD and FEAR had increased performance with the win defrag. So, how much did it cost? And how long did the PD defrags take? The results overall seem somewhat vague, as the performance increases and decreases are only slight. Maybe i should do this on my good machine? The test i mean

I just realised, im gonna need PD if i want to do these tests

 

[Speedo]

Check the "better defrag" thread. I think sitboy was offering the NFR (not for resale) version, which is what I was using. The retail version is $40.

Pay more attention to the average FPS than the min or max. The FPS can spike to 200 for an instant and make that middle bar look impressive, but the average FPS tells the main story - I just included the min and max for completeness.

Overall there's not a lot of improvement. The performance changes are pretty much with the margin of error I'd expect, though ideally I would have liked to run each test multiple times and under more controlled conditions. I'm not sure sure exactly what happened with Falcon, but overall it's basically what I expected.

I do think PD may have decreased my boot time somewhat, but that's a bit hard to quantify on my system with the various password prompts and menus between power-on and the desktop.

The PD online scan took about an hour, the offline scan maybe 5 min.

 

[Mise]

I'm not sure if that's a fair test, since once you've defragged with one program, defragging with another program won't make any difference, surely?

That said, it's probably safe to say that there is little correlation between frame rates and hard drive fragmentation. I am rather surprised about the load times though.

 

[Speedo]

I'm not sure if that's a fair test, since once you've defragged with one program, defragging with another program won't make any difference, surely?

The claim is that PD will optimize the disk and various other things that make it better than Windows - it certainly spent an hour rearranging stuff on my disk. If it didn't do a better job of defragging than windows, what would be the point of buying it?

 

[Genocidicbunny]

I did the testing, with a wider variety of games, and might post results later tonite

I did the testing on two machines, an old p3 867mhz, mx 4400 gpu, 768 ram.
and on a much newer comp-amd athalon 64 fx 57, Dual 7800GTX, 2 gigs ram
Both had a raptor hd(10000rpm), although the older one had the 30+ one and the newer one the 70+

 

[Volum]

Speedo: This is a little OT but can you please tell me how i can get so many bots in single BF2 player? Or maybe a little linky?

 

[Speedo]

Look in the folder <BF2 install path>\Battlefield 2\mods\bf2\AI
Open the the file AIDefault.ai in notepad
You're basically editing lines 8-12, it has instructions in the file.

Mine is set to this:
aiSettings.overrideMenuSettings 1
aiSettings.setMaxNBots 75
aiSettings.maxBotsIncludeHumans 0
aiSettings.setBotSkill 0.7

You'll have to just play around to find the point where you get decent performance and stability. I can run 96 or so bots with the game relatively stable, but performance starts to go downhill. I think when I was playing around with it I tried as many as 130 bots.