The OFFICIAL FFXIV Benchmark & Beta Performance Thread

[Cylo]

From my understanding, and someone can correct me if I'm wrong, but there's a "skipping" issue with ATI cards. I have an even better processor and twice the ram you have and I had the same issue - fine on the boat and jumped frames in town.

The skipping problem was supposedly happening on both ATI and NVIDIA's high end cards.

I have a 470. It doesn't "skip" for me except rarely when loading a huge object.
I had really bad jittery / skipping performance after I installed the the DirectX 11 files that came with Kombustor, but I was able to fix it by installing the June DirectX Redist I posted earlier in this thread.

Did you try that DirectX 11 update I posted earlier?

[alecrast]

I agree, it seems like since the ATI 59NN series is the speed leader according to the benchmarks I've seen, there is no immediate pressure to lower the price.

If you rate them if terms of performance from fastest to slowest and use the cheapest newegg prices for comparison:

1. GTX 480 SLI (2-way) - ~$920
2. HD 5870 Crossfire (2-way) - $760
3. HD 5970 - $679
4. GTX 480 - $460
5. HD 5870 - $380

At the high end of the spectrum, there's almost no pressure to change prices. Any need for change will come in the form of a refresh from either company.

[Celeras]

You still running on the GTX 470?

Yep.

Couldn't get an 'off' screenshot for AO in the same location before the test ended, I'll do the comparison on the next test.

[volcanicweather]

The skipping problem was supposedly happening on both ATI and NVIDIA's high end cards.

I have a 470. It doesn't "skip" for me except rarely when loading a huge object.
I had really bad jittery / skipping performance after I installed the the DirectX 11 files that came with Kombustor, but I was able to fix it by installing the June DirectX Redist I posted earlier in this thread.

Did you try that DirectX 11 update I posted earlier?

I did not... I'll install it when I get home this evening. Thanks Cylo.

[BlackheartX]

If you rate them if terms of performance from fastest to slowest and use the cheapest newegg prices for comparison:
1. HD 5970 Crossfire (2-way) -~$1200-1400
2. GTX 480 SLI (2-way) - ~$920
3. HD 5870 Crossfire (2-way) - $760
4. HD 5970 - $679
5. GTX 480 - $460
6. HD 5870 - $380

At the high end of the spectrum, there's almost no pressure to change prices. Any need for change will come in the form of a refresh from either company.

fix'd (what chip can even run 2 5970s right now without gimping them?)

[Amele]

Thanks for posting about the 460. I guess the other FFXIV benchmark out there was actually reflecting the card pretty well. Compared to official reviews, you already have your core/shader clocked higher for the 1GB than any of the other ones (using gpureview.com) leaving only memory a bit slower.. considering the other reviewed 460 got ~2700 at stock speeds, I wonder if pushing it further will do too much. I know it's nothing but a midrange card, but considering how it compared to the 5850 in pretty much every other game out there, I'm a bit disappointed I suppose.

I pulled out over 1k extra points by OC'ing my 9800GX2, the 460 does seem to be underperforming on this bench compared to other game real world tests though.

[Katlan]

Yep.

Couldn't get an 'off' screenshot for AO in the same location before the test ended, I'll do the comparison on the next test.

I didn't notice input lag w/ AO on, but then again, I didn't get a chance to test it w/o AO because when I dc from warping I couldnt get back in game 9I also use a gamepad, dunno if that makes a diff)

[Amele]

fix'd (what chip can even run 2 5970s right now without gimping them?)

I would think the X58 chipset could?

[Zellot_OG]

Well, my benchmark scores were far lower than I expected:

2400 Hi Res
3300 Low Res

System specs are as follow:

AMD Phenom II X3 710 Processor (2.6 GHz)
8 GB Kingston HyperX @ 1066
Ati Radeon 4890 (1gb)
1 Terabit Hardrive
Windows 7 (64 bit)


I plan on playing on my 32" HD 1080p TV. Both my computer and my PS3 are connected to the TV. Since PS3 won't be available for at least 6 months after PC release, I'm going to have to go the PC route for now.

Initially, I wanted to get a Phenom X4 965 BE and a ATI HD 5850 to get ~3800-4000 Benchmark score on hi res. But, I read that the 6000 series cards are due to be released in October of 2010, which is a few weeks after the game's release.

Getting a Phenom X4 965 BE/5850 will cost me around $500. I'm wondering whether the lower end 6000 series card will perform better than the 5850, at potentially the same cost.

On the flip side... If my eventual goal is to play on the PS3, maybe I shouldn't upgrade at all. But looking at the benchmark requirements, I have no idea how well the game will play on the PC with my current setup (Phenom X3 710/4890).

So....

(1) Should I upgrade Now?
(2) Should I wait until October 2010?
(3) Should I not upgrade at all and wait for PS3?

Any suggestions? And to be honest at the outset, the smartest decision would be to wait and see how well/poor my current system performs before making a decision. Waiting will also cause current prices to drop on CPU's and GPU's, but maybe not that much.

I guess.. what I want to know is what is the highest settings I can play FFXIV on and still average 30 FPS with my current setup.

IF I can do default settings with 30 FPS average... i'll just wait 6 months before upgrading since that will probably save me ~200 bucks.

[Purrrfect]

stuff

I don't think those scores are too far off for the hardware. Surely there might be a couple hundred to be found by tweaking and all, but I don't think I'd expect to crack 3k on high with a 4890, just based on my experience with a 4870. You can definitely play the game as-is np though. i'd guess you can do 1080p at full 30 fps a good portion of the time with the base settings, no dof/ao.

[Vuitton]

Why do the 58/59s have to get cheaper when they can just charge more for the 60?

They do it to take advantage of the market. The 60xx would cost as much as the 5970 (~$700), and everything else will be lowered in price by ~$50 - ~$100, which would put nVidia at a big price/performance ratio disadvantage on all of their cards compared to ATI. ATI would rule the entire graphics market until nVidia releases their next line putting us back in the same situation we are in today (ATI having a slight lead) but with better cards.

[Jinko]

Guess I will hold out for a Radeon 6870.

[alecrast]

fix'd (what chip can even run 2 5970s right now without gimping them?)

The 5970 is a dual-GPU single-PCB part - it is two underclocked 5870s on the same PCB. Using your "fixed" chart, it becomes:

1. GTX 480 4-way SLI -- $1840
2. 2x 5970 Crossfire (this is 4-way Crossfire, btw) -- $1360
3. GTX 480 SLI (2-way) -- $920
4. HD 5870 Crossfire (2-way) -- $760
5. HD 5970 -- $679
6. GTX 480 -- $460
7. HD 5870 -- $380

3-way GTX 480 SLI (3 GPUs) is faster than 2x 5970 Crossfire (4 GPUs). SLI scales better than Crossfire. That's probably part of the reason that 3x 5870s is faster than 2x 5970s with inefficiency offsetting performance.

http://www.hardware.info/nl-NL/artic..._GTX_480_test/

They do it to take advantage of the market. The 60xx would cost as much as the 5970 (~$700), and everything else will be lowered in price by ~$50 - ~$100, which would put nVidia at a big price/performance ratio disadvantage on all of their cards compared to ATI. ATI would rule the entire graphics market until nVidia releases their next line putting us back in the same situation we are in today (ATI having a slight lead) but with better cards.

ATI is not going to introduce a single-GPU solution for $700. The market will not bear it. More likely what you'll see is ATI selling a reference design for ~$500 - $529 with OEMs charging up to $600 for non-reference designs. ATI doesn't want to create a niche market for their top GPU - they want to keep pressure on NVIDIA so the likelihood of a reference HD6000 at $600 is almost nil. Note that I am making the differentiation between ATI suggesting the MSRP of $500 and online vendors taking advantage of product launch and product shortages to jack the cards up $200 over their MSRP.

Don't expect to see too much movement in the 5870 price. The reason for that is that AMD meets with its partners and informs them of its plans (hence the HIS leak earlier this week) so that the OEMs aren't left with huge volumes of old product. AMD could potentially hold off on launching its new product until existing stock of 5870s declines.

[Vuitton]

Has anyone run the benchmark with the 460? I don't see one on the charts.

[Amele]

Has anyone run the benchmark with the 460? I don't see one on the charts.

Shaddix just did recently, look through the last two pages or so.

[avelle]

Are ATI cards still experiencing issues with FF14? I couldn't help myself and ended up purchasing a Sapphire 5870 Vapor-X card. :3

[Khalil]

Are ATI cards still experiencing issues with FF14? I couldn't help myself and ended up purchasing a Sapphire 5870 Vapor-X card. :3

I have the same card although mine's factory OC'd but I gotta say I'm entirely happy with Sapphire.

[Kawfee]

Are ATI cards still experiencing issues with FF14? I couldn't help myself and ended up purchasing a Sapphire 5870 Vapor-X card. :3

Yes, still stuttering, but "better"(?) is what people have said in various places recently.

[Kyrra]

Has anyone run the benchmark with the 460? I don't see one on the charts.

Yes: http://www.bluegartr.com/forum/showp...postcount=1917
Problem is, he's using a Core 2 Duo (dual core) CPU that's OCed. I'd like to see the 460 running on an i5 or i7 CPU.

[Katlan]

general tech question. We all know the gtx 480 runs hot as fuck at full load... about 89-92c. So... if your gonna be playing a game that's gonna keep the gpu at full lood for 80%+ of the time.. would it be better to manually turn the fan off of auto and keep it at 100% (5k rpm) to keep the temp down, or does it not matter since the card is meant to run this hot? Obviously lower heat is better, but don't feel like hearing the fan run at 5k rpm if it doesn't need to be (I think the fan only runs at 60-70% auto on full load)