The OFFICIAL FFXIV Benchmark & Beta Performance Thread

[avelle]

Hello, this is my first post here so please bear with me. I recently ran the ffxiv benchmark with the new nvidia drivers and renaming the .exe to crysis (instead of 3dmark06, which I was using to no avail) and I saw my score literally double. Before, I hovered around 2500 high, but now I'm sitting at 5300. Could the driver and the crysis.exe change really have increased my performance by 100%?


I am really quite baffled (and extremely pleased) that I got such a huge performance boost, but I can't seem to understand why.

Unfortunately, since I do not have the post count to post URLs, I linked the img to ZAM's FFXIV benchmakr thread as Coh.

Are you running SLI or have a 295gtx? If so I could see you getting that kind of an increase.

[Kumaeru]

Hello, this is my first post here so please bear with me. I recently ran the ffxiv benchmark with the new nvidia drivers and renaming the .exe to crysis (instead of 3dmark06, which I was using to no avail) and I saw my score literally double. Before, I hovered around 2500 high, but now I'm sitting at 5300. Could the driver and the crysis.exe change really have increased my performance by 100%?


I am really quite baffled (and extremely pleased) that I got such a huge performance boost, but I can't seem to understand why.

Unfortunately, since I do not have the post count to post URLs, I linked the img to ZAM's FFXIV benchmakr thread as Coh.

The fullscreen .dll and .exe name change only affects scores if you have more than 1 GPU in you're system. So either you are running SLI/Crossfire or you are using a single video card with 2 GPU's. That's the only way you can see a score difference. On a single 5770 I've tried both the windowed and fullscreen version and was around the same score.

The windowed benchmark does not utilize multi GPU.

[Tythera]

That has nothing to do with your GPU. SE is using a software mouse client instead of the typical hardware mouse. Yes, it's annoying, as it makes the mouse pretty much useless.

Yeah, I learned that after I had posted as I hadn't done the beta before. I don't have any idea how they thought it was a good idea. Think I've got things figured out with control but it definitely still runs pretty ragged as the original benchmark would indicate.

[Caldor]

The fullscreen .dll and .exe name change only affects scores if you have more than 1 GPU in you're system.

This doesn't seem to be entirely true. I got a 4655 on my high res run on the windowed benchmark while the non windowed one gave me a 4750(iCore 7 860 + single ATI 5870). This might just be about a 100 point increase but still. I'm not entirely sure what affected the increase though. I can only imagine that the benchmark is for some reason using my onboard sound in the windowed mode instead of the x-fi titanium and therefore I lose a bit of CPU performance but that's just a guess.

[Treijhen]

im upgrading my graphics card to a Nvidia 9800 most likely but my other specs are

AMD 64x2 Dual Core 4800
2048mb of system memory
320gb hard drive
Windows Vista 32bit
2gb of ram

will this run the game with the Nvidia 9800?

[Slyth]

Nice results for the G73. I had a question though, were you able to monitor your CPU/GPU temps? I'm interested to know what type of temps your seeing with the G73.

The Guy who does the Live stream at the FFXIVLounge.ning.com uses a G73. He said he is able to run the game at 1920x1080 30 fps every where except for limsa lominsa.

[FerBedolla]

Intel Core i7 920 D0 @ 4.00 Ghz
Kingston KHX1600C9D3K3 4GB X 1 DDR3 @ 1600 Mhz Timing 9-9-9-24 Single Channel
ATI Radeon XFX HD 5970 Black Edition 1 GPU Working @ Core 1000 Mhz Memory 1300 Mhz
Edit: I forgot to say it's High Settings (1080p)

http://www.enri.us/ATIPwnsYourLife.jpg

[shaddix]

jeebus

[FerBedolla]

Hehe

Even at 70c it should be no problem, but you are doing the right thing start slow.

On the other hand, GTX 480 on stock fan at 100% speed and 850 clock will hit around 75~84C. Max GPU Temperature is 105C... You should have seen people cooking on that thing rofl...

The HD 5970 Clock @ 1Ghz Stock Fan Speed at 100% temperature is 65°C @100% GPU Load

[ScootSeraph]

I am still in awe that I can't play this game on 2009 Mac OSX, it's ridiculous. Nobody should have to upgrade their computer for a video game unless it is like 10 years old.

Like usual, SE is digging themselves a hole and hanging on to their small niche following.

[Jianfre]

Lol Mac.

This is just about in tune with people trying to run it on netbooks.

[natethegreat]

im upgrading my graphics card to a Nvidia 9800 most likely but my other specs are

AMD 64x2 Dual Core 4800
2048mb of system memory
320gb hard drive
Windows Vista 32bit
2gb of ram

will this run the game with the Nvidia 9800?

With my 9800 GTX+ I was getting about 1700ish on high and 2200ish on low. So if the benchmark is any indication if a particular hardware set up will run then game then I would assume, yes. That was paired with a Intel E6300 at 2.4 Ghz. If you goal is just to get the game to run then perhaps the 9800 is a good choice. However, I have not played the game so I can not be certain that my assertions are true.

[BlackheartX]

Here's a new one for 5850 fans 4.15ghz Cool n Quite makes it 2.8ghz on the bench. (Low)
http://img641.imageshack.us/img641/6...relow6160x.jpg

High Huge!!!

Spoiler: show

http://img10.imageshack.us/img10/2011/capturehi5095.jpg

[Mrmartin]

i7 930 OC @ 3.8 Ghz
2 x HIS HD 5870 Turbo OC @ 900/1300 Mhz (Wish both worked for the benchmark)

High Setting Score -
http://img820.imageshack.us/img820/1...mark201007.png


Low Setting Score -
http://img231.imageshack.us/img231/1...mark201007.png

[Sanhedrin]

This doesn't seem to be entirely true. I got a 4655 on my high res run on the windowed benchmark while the non windowed one gave me a 4750(iCore 7 860 + single ATI 5870). This might just be about a 100 point increase but still. I'm not entirely sure what affected the increase though. I can only imagine that the benchmark is for some reason using my onboard sound in the windowed mode instead of the x-fi titanium and therefore I lose a bit of CPU performance but that's just a guess.

When an application runs in fullscreen mode (real fullscreen, not maximized), it gains ownership of the render surface (screen area) and so the application frame buffer (area of memory with image data) gets to be displayed directly. In windowed mode, some element of the OS' display model has ownership of your screen's real estate and has to do the work of compositing multiple elements onto the screen (all of the data from your desktop/other windows). The former, as you can imagine, is more efficient, as there really isn't any middleman or other data needing to be displayed. Your sound card is going to work, or not work, the same in both instances.

[Olo401]

real benchmark doesnt allow sli/cf so rly those results dont mean anything

The 5970 isn't Xfire - it's a straight-up dual GPU card.

[Celeras]

The 5970 isn't Xfire - it's a straight-up dual GPU card.

It's still crossfire. If the program doesnt allow it, the 5970 will only have one active gpu.

[bonji]

"postponed until FPS is no longer capped at 30"

is that cap going to be removed?

[Celeras]

"postponed until FPS is no longer capped at 30"

is that cap going to be removed?

Maybe

[Evanos]

Intel Core i7 920 D0 @ 4.00 Ghz
Kingston KHX1600C9D3K3 4GB X 1 DDR3 @ 1600 Mhz Timing 9-9-9-24 Single Channel
ATI Radeon XFX HD 5970 Black Edition 1 GPU Working @ Core 1000 Mhz Memory 1300 Mhz
Edit: I forgot to say it's High Settings (1080p)

Which catalyst drivers are you using for that? Also, single channel memory? Why did you set it to that? I have an ATI 5970 OCed to 950 and 1275 and am hardly breaking 5.8k with single GPU test, and when I use the SLI workaround easily getting almost to 8k.