The OFFICIAL FFXIV Benchmark & Beta Performance Thread

[Questionable]

Update, got a 5830 card that was on sale + MIB for my second system. Tested it on my current E6550 system.

High res bench: 2586

This is pretty awesome. I only OC my E6550 to 2800 from 2333. This is why 2GB DDR2 Ram. I ran the beta and was way smoother @ full screen 1920x1080 with window size buffer.

So perhaps we don't really need such a high system to play. Once I upgrade to a Q8400/9400 or similar + 4GB ram and OC that and OC my 5830, I think I would bench close to 3700-4000. Which would be enough to play without issues.

This card runs cooler too, 63C at full load playing beta. My old 8800GT ran 98C at full load, 75 Idle

HIGH:
http://i515.photobucket.com/albums/t...high_small.jpg

LOW:
http://i515.photobucket.com/albums/t..._stock_low.jpg

[kelvinblade]

questionable, what settings are you running the beta with at 1920x1080 and does the game run completely smooth? thx

[Questionable]

From the 10mins I've played in beta it is smooth. At least in the field near the crystal where everyone is. I'd say with my current setup it's playable 80% good.

Settings:

FullScreen
1920x1080
2Sample
Window Size
Standard
NO AO or DOF as Alecrast pointed out they kill FPS even with a powerful setup.

EDIT: Actually with 30FPS cap it seems not that playable as I thought, it still lags a bit with lots of people around, drops to 20fps

If you turn down settings or play lower resolution, it might be okay with my setup. Still need the CPU upgrade and ram, hopefully that + OC everything can make it playable at all high settings

[Amele]

Huh... I'm getting lower scores than I should? weird... only thing that I think may cause this is I've set my ram single channel because My ram sticks come with a heatsink and are to tall, can't fit them under the CPU cooler... so they're at 2x2 GB single channel, I'll try to get a new cooler or RAM soon, but can only that make a score drop 400-800 points? :/


OC'd the AMD 1090T to 4 Ghz

Spoiler: show

http://img691.imageshack.us/img691/5964/benchoc.jpg

Last screen was dark because I snapped it as the bench was closing.

unless you're overclocking I'm pretty sure just the spreaders should be enough. - your ram did come with spreaders right?

what ram timings are you at and what's the JEDEC spec? (or xmp spec, I guess if you don't know your JEDEC)

single channel will absolutely hurt your scores at the speed your cpu is at.


to someone who was talking about ati v. nvidia - about the only thing the nvidia cards do straight up better right now is physx support (being an nvidia thing) and tessellation. as far as I know, unless you're in a game which *heavily* leverages tessellated rendering you won't notice the difference on the latter, and there is as yet no title that requires physx to play.

(full disclosure, my last 3 videocards were all nvidia. I like the company, but this generation they got beat hard in price/performance.)

[Celeras]

Since post-release driver updates the nvidia cards are faster across the board in most everything, but yeah way faster in heavy tessellation as well. There's a thread stickied over in tech that may be more appropiate for this kind of comparison though.

Tech thread: http://www.bluegartr.com/forum/showthread.php?t=95796

Truth be told you can't go wrong with either, everything has a selling point.

[Wulfgang]

10.6 ATI drivers
ATI 5870
i7 930 @ stock

Literally just finished building my comp and installing windows / new drivers three minutes ago. On high, got 4724. Unfortunately will probably need to reseat HSF (Scythe Mugen 2). At stock 2.8, I'm at high-60's / mid-70's on all cores ~10 mins into Prime 95. Probably used too much thermal paste, was my first build.

[Questionable]

10.6 ATI drivers
ATI 5870
i7 930 @ stock

Literally just finished building my comp and installing windows / new drivers three minutes ago. On high, got 4633. Unfortunately will probably need to reseat HSF (Scythe Mugen 2). At stock 2.8, I'm at high-60's / mid-70's on all cores ~10 mins into Prime 95. Probably used too much thermal paste, was my first build.

Are you using Arctic Silver 5?

I have the Mugen 2 rev-b, not sure if I'm going to use it though, still deciding. I have a Zalman CNSP9500 on my current system, I was going to grab the updated version of this one to use instead. Though people were giving good reviews on both.

[Wulfgang]

No I used Arctic MX-2:

http://www.newegg.com/Product/Produc...82E16835186020

The reviews are good, but I was fairly certain the one thing I may have botched was the compound application. Although, it felt good to POST after building for the first time. I ran into one issue because the CPU fan socket on my mobo was borked which was a pain in the ass troubleshooting. Ended up moving it to a system fan socket np and got a single beep on boot.

[Madmartigan]

I am currently running:

Win7 64-bit Ultimate
Asus P6T v2 Deluxe motherboard
IntelCore i7 [email protected] GHz
6 GB Corsair Dominator (12800) DDR3 Ram
1 TB Western Digital Caviar Black HD
Sapphire ATI Radeon HD4890 Vapor-X 2GB Video Card

My low benchmark tested at 4256, but my high benchmark test was only 2476.

I understand that my GPU is what is lowering my scores. My specific question is this...

Is it better to purchase and run a single HD5870 ($430) or two HD4890's (since I have one already, only $200 more)

[Questionable]

Yeah your high scores are lower than mine and I'm using E6550.

I'd think with your setup I would get a 5870. You can always CF the 5870 too.

Also I'd like to point out, the benchmark is misleading. Even though it felt smooth 100% with my 2500 score, in game it's a different story. There's still the usual lag you feel, just like in XI @ whitegate or besieged when there's a lot of characters around. I guess they haven't fix their issue with how stuff is being rendered in the engine.

I just hope the 30FPS cap is removed...

[Melee101]

FFXIVBenchmark / PC question?
Right off the bat, so you know, I don't know much about the workings of PCs. From the top. My fiance and I went to Best Buy with the specs needed to play FFXIV and asked which gear we could take home and play on when the game is released. We were sold (2) Alienware m17x, window 7 home premium, Intel Core2 quad cpu q9000 @ 2.00GHz, 4GB, NVidia GeForce 9400M. Yesterday FFXIV released the "benchmark" software. After some figuring we got DirectX updated and got the program working. I now have more questions than before benchmark.

A) I cannot see my HD score due to the fact that the window is larger than my screen will show and won't let me drag it high enough. How do i fix this problem?

B) On low res. I hit 2500-3000 on a few different tries. Why are such variances possible and how do I get this # higher?

C) Any concern with that benchmark program making my fans kick on super high mode?

D) Her PC, which is EXACTLY the same as mine is only hitting 300s after a painfully lagged video. When I try to close out all of the boxes I get a smaller box warning me that a file is corrupt. I've tried removing and reinstalling benchmark software a few times but, nothing changes. This is the question I need answered the most. Matching PCs on same network should run near the same #s, right? What have we done wrong here?

[Strife]

Specs:
ASUS P6X58D Premium
i7 930 @ 4.0ghz
EVGA Geforce GTX 480 SuperClocked
6gb DDR3 1600
Intel X25-M 80gb SSD

Low:
http://img266.imageshack.us/img266/641/fflow.jpg

High:
http://img340.imageshack.us/img340/7240/ffhi.jpg

[Sephiroth]

From the 10mins I've played in beta it is smooth. At least in the field near the crystal where everyone is. I'd say with my current setup it's playable 80% good.

Settings:

FullScreen
1920x1080
2Sample
Window Size
Standard
NO AO or DOF as Alecrast pointed out they kill FPS even with a powerful setup.

EDIT: Actually with 30FPS cap it seems not that playable as I thought, it still lags a bit with lots of people around, drops to 20fps

If you turn down settings or play lower resolution, it might be okay with my setup. Still need the CPU upgrade and ram, hopefully that + OC everything can make it playable at all high settings

Interesting. You've got almost the same setup as me. I have a 5830 with a 2.66ghz dual core and 3gb RAM, Vista. I score roughly 2570 on high.

I hope to run the game in 1920x1200 or 1080p, looking like I'll need a CPU upgrade to do that based on your experiences. I don't have beta access yet so can't tell you how it runs for me. I don't plan to upgrade CPU until I play it and find it's worth doing.

[Icarian]

unless you're overclocking I'm pretty sure just the spreaders should be enough. - your ram did come with spreaders right?

what ram timings are you at and what's the JEDEC spec? (or xmp spec, I guess if you don't know your JEDEC)

single channel will absolutely hurt your scores at the speed your cpu is at.

I have these http://www.ocztechnology.com/product...666_reaper_hpc

Now, what's weird is, my comp detects them as PC3-8500E 1066 Mhz, my old one did the same thing, timings are set to Auto (8-8-8-20).

[Nightofday]

My set up
AMD 955be @3.8
MSI 790FX-GD70
8g of OCZ Gold AMD Edition ram
CORSAIR 750w

Ran 3 diffrent cards each time I used driver sweeper and the latest drivers.

ati 5830
ati 4980
9800 gx2

5830
low 4759
high 3133

4890
low 4260
high 2619

GX2
low 4234
high 2664


I like to note the gx2 was choppy (msi ab showed both cores to be actived) even tho it scored about the same as 4890. The 4890 and 5830 ran smoth to me.

[alecrast]

For those people interested in performance, I wrote down the steps to create a very basic performance counter for CPU usage.

1. Start -> Control Panel -> Administrative Tools -> Performance Monitor

2. Expand Data Collection Sets

3. Right click User Defined

4. New -> Data Collection Set

5. Call it "All CPU"

6. Check "Create Manually (Advanced)"

7. Click "Next"

8. Make sure "Create data logs" is checked

9. Then check "Performance Counter" beneath it

10. Click "Next"

11. Set "Sample Interval" to 1 (make Units is in Seconds)

12. Click "Add"

13. Find "Processor" and click the little down arrow next to it to expand the list

14. Find and select "% Processor Time"

15. In the box below, select "<All instances>"

16. Click "Add" (this will add it to the window on the right)

17. Click "OK"

18. Click "Next"

19. Save the data wherever you want

20. Click "Next"

21. Make sure "Run as:" is set to "<Default>" and "Save and close" is checked.

22. Click "Finish"

Now you should see "All CPU" under the User Defined list

23. Click it and something like "DataCollector01" should show up in the right window pane

24. Right-click DataCollector01 and choose "Properties"

25. Set "Log Format" to "Comma Separated"

26. Click "Apply" and then click "OK"

27. Now you can right-click "All CPU" and choose "Start"

Remember to click "Stop" when you're done collecting data.

I recommend the above in tandem with MSI Afterburner which you can set to log data, as well. Then you can take the two datasets and graph them together in Excel or GNUplot for some useful overlays of CPU vs GPU vs Framerate.


I'm logged in doing performance analysis this morning because there's an area that is interesting to me in the city where framerate is 21fps but GPU usage is only 45% so I'm trying to figure out what is going on there.

Also, interesting observation out at Bearded Rock this morning. With my preferred settings (2560x1440, MS 8x, Shadows Highest, FoV ON, Buffer size: Window), when it was raining, GPU usage was up to ~93% with approximately 25 people standing around. When the rain stopped, GPU usage dropped to ~78%.

[Reality]

Please tell me this isn't only happening to me, and it's not my TV dying

Does anyone else notice half of the screen (while playing) is slightly darker than the other half

The darker half goes diagonally from the upper left corner to the lower right, so the screen is split diagonally with the upper section being slightly darker...

I don't notice this on my windows desktop, and I don't have any other games installed to test it on...

edit: I had supcom: FA installed and I don't notice it... so hopefully it's just FF14 beta

[Cooleko]

to someone who was talking about ati v. nvidia - about the only thing the nvidia cards do straight up better right now is physx support (being an nvidia thing) and tessellation. as far as I know, unless you're in a game which *heavily* leverages tessellated rendering you won't notice the difference on the latter, and there is as yet no title that requires physx to play.

(full disclosure, my last 3 videocards were all nvidia. I like the company, but this generation they got beat hard in price/performance.)

I heard metroid 2033 (may have butchered the name) is really physx heavy. but I doubt it is "required" to play.

[shaddix]

Please tell me this isn't only happening to me, and it's not my TV dying

Does anyone else notice half of the screen (while playing) is slightly darker than the other half

The darker half goes diagonally from the upper left corner to the lower right, so the screen is split diagonally with the upper section being slightly darker...

I don't notice this on my windows desktop, and I don't have any other games installed to test it on...

edit: I had supcom: FA installed and I don't notice it... so hopefully it's just FF14 beta

malfunctioning tv will usually screw up with the defect affecting the top or bottom half split horizontally down the middle. Or it will manifest as a vertical line of any width going the full vertical length of the screen or starting/stopping at the midpoint of the screen.

[Ashtray]

http://img.photobucket.com/albums/v7...52_Cropped.jpg

http://img.photobucket.com/albums/v7...a/High4788.jpg

http://img.photobucket.com/albums/v7.../CPUID_4_2.jpg

Rated 2000 Mhz 6GB G.Skill Trident Memory Settings.

Didn't touch the GTX 470.



Shame SLI isn't supported on this benchmark.