The OFFICIAL FFXIV Benchmark & Beta Performance Thread

[Celeras]

Celeres, if you want you can update my scores from my previous post, also my CPU is actually running at 3.2, not 3.6 (Windows 7 misreports it for some reason)

Ill update again later, but I'm just sticking to what the benchmark reports for consistency.

[arus2001]

http://i561.photobucket.com/albums/s...s2001/14bm.gif

Feel like I should be higher, but maybe that's the bottleneck thing some have mentioned.

Recently upgraded to Windows 7, so I guess I'm curious about how I might be able to optimize further, too.

[Charismatic]

5870
http://img682.imageshack.us/img682/2...chmarkhigh.png


5870

[Cjskater]

Another comparison for 2 vs 4 cores:

4: (60+FPS)
http://i50.tinypic.com/30sd1ua.jpg

2: (30-50FPS)
http://i48.tinypic.com/358b04g.jpg

[Celeras]

Thats a decent increase Domon, you turned things up in driver settings? Weird, I dont get why it would behave like that unless the default settings for the benchmark are actually higher? Which would essentially be disabling portions of the bench to increase your score, which you shouldnt be doing.

Still not sure how Kryss doubled his score though.. reluctant to include them as they're not in line with other scores of similar hardware.

[shidobu]

No registry tweaks/ATT tweaks, everything from within CCC this time. Maybe registry tweaks later once the game is more finalized. Not worth it at the moment.

Differences between first and second:

Standard Cataylst AI -> AI Disabled.
AA Filter Box -> AA Filter Edge Detect.
AAA Disabled -> AAA Enabled, Supersampling, Highest Quality.
Anisotropic Filtering Normal -> Anisotropic Filtering High Quality AF Enabled.
Mipmapping Quality -> Mipmapping High Quality.

No other software settings changed. Catalyst AI is horribly clunky though for the performance change you're supposed to get, as mentioned previously by others.




Oh yeah, and to highlight on another old point Celeras showed from another thread... I'm using a single 7200RPM 320GB OS drive with a single 5600 RPM 1.5TB green data drive.

Just did all these Tweeks and mind you i updated to 10.5
i got 4806...
not rly that much of a fall ima run it 3 times tho

[Kryssan]

Krysan if u shut off 1 card. You will prob get higher than mine. Should try it my 5870 is XFX. I was gonna get a sapphire but couldnt wait for Newegg to restock. ~.~

I tried it. All settings the same with only one card installed.

Spoiler: show

http://www.aethernavale.net/Media/im...4-high-cfx.jpg

I didn't really think it would change that much. Interesting. I wonder what you could do with registry tweaks.

[Kryssan]

The tweaked scores are the second options? I fail to see why that would make your score double, seems like you're turning things up rather than down o.O Or are you saying you were forcing those things via driver for everything, and then turned them off and reran?

The first pair of low/high are the left of the ->, the second pair are the right of the ->.

[Cjskater]

What happens if you rename "FFXivWinBenchmark.exe" to something like "bioshock.exe" (with working crossfire/sli profiles)? Don't have 2 cards so can't really test. I do know though you can force SLI/Crossfire that way, not sure if the engine matters though.

[Akiyama]

I've actually never touched anything within my CCC because I didn't know what a lot of that shit did. D: And I'm on 10.3 Drivers still, Lol

[Celeras]

Just gonna go ahead and stick with untweaked scores. Overclocking is fine but anything else gets dicey.. especially when turning up settings on the driver level seemingly results in higher scores for undermined reasons.

Leaderboard is more for comparison sake anyway, rather than big numbers for people.

[mastamind]

e6600 oc'd to 3.0 and 3.4
single 5870 (shows 2 because i have my tv hooked up to the card as well, maybe)
both on high quality

Spoiler: show

http://img195.imageshack.us/img195/8...mark201006.jpg


http://img252.imageshack.us/img252/8...mark201006.jpg

now i'm slightly confused, reran it on low this time and got a 3238, which was lower than my 3263 on high

[LD]

I just wish I knew why I got an 'unplayable' score while it seemed playable.

[Jinko]

Tried those settings and it turned our worse for me, noticed a lot more jaggies with edge detect aswell.

[Insanecyclone]

Low settings. High doesn't fit the screen but I'm going to do it again and see for high.

[Ferien]

Tells me exactly what I expected: Time for an upgrade.

Spoiler: show

Hi: http://i170.photobucket.com/albums/u...ino/ff14hi.png
Lo: http://i170.photobucket.com/albums/u...no/ff14low.png

[Isla]

I just wish I knew why I got an 'unplayable' score while it seemed playable.

Well keep in mind it's just a rough benchmark. It doesn't have the dynamic multi-player environments that will actually be in the game. It's just giving you a rough estimate of what to expect.

[Snowknight26]

Mirror cause nvidia.com was sucking for me: http://stfcc.org/misc/FFXIVBenchmark.zip

Edit: RAM disks don't speed up the loading times it seems.

[Alderaan]

CPU is obviously on the lower end of the spectrum - GPU is decent.
Almost identical numbers.

High:
http://img.photobucket.com/albums/v6...1520-39-57.jpg

Low:
http://img.photobucket.com/albums/v6...1520-47-07.jpg

[shidobu]

I tried it. All settings the same with only one card installed.

Spoiler: show

http://www.aethernavale.net/Media/im...4-high-cfx.jpg

I didn't really think it would change that much. Interesting. I wonder what you could do with registry tweaks.

Blew my score outa the fucking water, knew your rig could.

Atleast i bested my first.

4813

Spoiler: show

http://i3.photobucket.com/albums/y97...obuxx/4813.jpg

my poor alienware laptop is fucking getting its ass kicked
Alienware laptop
Windows Vista ultimate 32bit SP2
Core 2 duo p8600 @ 2.4
ATI mobility radeon hd 3870 (gonna buy another later)
Low score 1152
load time 31011 ms

Lol @ me tryin 2 play this shit away from my main rig (not alienware before people jump my ass)