The OFFICIAL FFXIV Benchmark & Beta Performance Thread

[xopher]

Emirii If you haven't noticed, AMD processors are on the low end of the scale according to this benchmark.

It is your CPU holding you back.

[Celeras]

Emirii If you haven't noticed, AMD processors are on the low end of the scale according to this benchmark.

It is your CPU holding you back.

Only because they're all mostly old ones thus far, there's a few of the newer ones that have been posted elsewhere or have been copied to this thread by comparison(and thus aren't on the board).

I have a 1055T in route, and plan on posting comparisons to my previous scores(page 1, and on scoreboard) when its arrived and installed. Ill overclock it to 3.0 for 1:1 comparison, then overclock to it's max to see what it can really do.

[ClassicRed]

I made a friend of mine test his computer ~ He just spent a boat load on it I expected it to do pretty decent but my high beat his high by 300ish pts ~

Mine
http://www.melodiesofdeath.com/classicred/high.jpg

His
http://www.melodiesofdeath.com/classicred/score.png

1 4850 vs 2 4870

I are confused ._.; ~ I made him update to the same drivers I have and directx9c D:

[SamanosukeShiva]

does not support xfire yet, you have to change some stuff for it to do xfire in benchmark

[Cjskater]

Emirii If you haven't noticed, AMD processors are on the low end of the scale according to this benchmark.

It is your CPU holding you back.

My 940 scores weren't updated but I'm right between Celeras and Piroth at stock (3Ghz) for low and between SnowKnight/Piroth for high. I'd say they are pretty similar at least for the Core2s.

[eskimobobz]

My 940 scores weren't updated but I'm right between Celeras and Piroth at stock (3Ghz) for low and between SnowKnight/Piroth for high. I'd say they are pretty similar at least for the Core2s.

I posted my scores earlier with a Phenom II X4 955, slightly higher clock than yours, and I get just slightly higher scores as well, on low. My scores were High = 2212 Low = 3946

I overclocked my 955 from 3.2 to 3.6, and both High and Low scores went up by maybe 50 max. So, either there's a cap somewhere, or GPU plays a huge role in both scores, even though Low is more CPU dependent.

*notice how all the best low setting scores still have the best GPUs*

I've been looking at a bunch of benchmarks around the web and I've seen a Intel E8400 @ 3.0 stock w/ a 5870 get almost 4k.
And, a Q6600 @2.4 stock w/ a 5850 get about 3700. With their low scores being about the same as their high.

[Cjskater]

I posted my scores earlier with a Phenom II X4 955, slightly higher clock than yours, and I get just slightly higher scores as well, on low. My scores were High = 2212 Low = 3946

I overclocked my 955 from 3.2 to 3.6, and both High and Low scores went up by maybe 50 max. So, either there's a cap somewhere, or GPU plays a huge role in both scores, even though Low is more CPU dependent.

*notice how all the best low setting scores still have the best GPUs*

I've been looking at a bunch of benchmarks around the web and I've seen a Intel E8400 @ 3.0 stock w/ a 5870 get almost 4k.
And, a Q6600 @2.4 stock w/ a 5850 get about 3700. With their low scores being about the same as their high.

I get pretty much the exact same scores with mine overclocked to 3.52Ghz (seriously, its kind of freaky). We don't cap out there though, there's another guy with a 9800GTX+ or something like that that hit 4.5k on low and I think 2.5k on high. Granted he overclocked the GPU a bit as well.

Also I just tested:

Bwahaha!

My puppy accidentally knocked my monitor power cord loose while I was running the benchmark.... That's on High btw. Not counting it as an actual score, just showing how unoptimized the benchmark is.

I saw no difference in my score, I didn't really expect to either. The GPU does the same amount of work regardless of whether the monitor is plugged in or not (or people with small monitors would rape the 1080p test). So maybe it has something to do with SLI?

[Emirii]

Yeah, definetely my processor. It's a 2.2GHz AMD Phenom 9500 which is about 3/4 years old (my computer it's self is 2.5 years old with a few upgrades here and there). Telling from the reaction/loading time in the benchmark (which is obviously not affected by the graphics card) it was over 30k in my test. My graphics card has no reason to be slow as it is brand new with plendy of graphics RAM, but it's true the processor is slowing it down. But like I said it is a BENCHMARK and probably won't do anything based on it, even though getting some new shiney toy for my computer doesn't sound all that bad.

Plus, my computer is what the average market has right now, if they don't make the game at least somewhat playable for the average market, they are going to have trouble selling this game to anyone, thus loose a LOT of player base. So it still needs to be optimized for older computers and processors alike, just like mine.

[Hyan]

Initially yeah. Does the game need 6 million subs right off the bat though? People who want to play will get a new rig (or PS3) sooner or later and join in.

[actarus]

Interesting tab

Personal considerations ( just my 2 cents ):

High
Soap : i7 980X @ 3,3 Ghz / HD5850 -> 4248
Gulkeeva : i7 930 @ 3,2 Ghz / HD5870 -> 4911

Low
Soap : i7 980X @ 3,3 Ghz / HD5850 -> 8062
Gulkeeva : i7 930 @ 3,2 Ghz / HD5870 -> 6118

So GPU for 1080p and CPU for 720p ?

[alecrast]

Interesting tab

Personal considerations ( just my 2 cents ):

High
Soap : i7 980X @ 3,3 Ghz / HD5850 -> 4248
Gulkeeva : i7 930 @ 3,2 Ghz / HD5870 -> 4911

Low
Soap : i7 980X @ 3,3 Ghz / HD5850 -> 8062
Gulkeeva : i7 930 @ 3,2 Ghz / HD5870 -> 6118

So GPU for 1080p and CPU for 720p ?

Yes, with the understanding that the "Low" benchmark is only CPU bound provided that you have a sufficiently fast GPU to allow it.

Soap's "Low" score is ~30% faster than Gulkeeva's due to his Gulftown having 12MB of L3 cache as compared to the 8 that Gulkeeva's Bloomfield has, which gives a boost when running single applications that don't make use of all 6 physical cores. Were it not for that additional L3, the 980X and 930 would score pretty much identical when both clocked at 3.2GHz (with minor variance for "uncore" clock differences, if any, and the higher latency L3 of the 980X).

[ufo]

Low benchmark, netting only 70 points going from 3.0 to 3.3; score seems limited by the HD 4850 GPU.
img340.imageshack.us/img340/9316/low30.jpg
img62.imageshack.us/img62/8003/low33.jpg

[actarus]

Low benchmark, netting only 70 points going from 3.0 to 3.3; score seems limited by the HD 4850 GPU.

I've 4850 too and I'm scoring 3500 in low with i5 @ 3,58 ghz ... seems like gpu act as a cap anyway

I'll start save some cash for a 6850 8)

[ufo]

I've 4850 too and I'm scoring 3500 in low with i5 @ 3,58 ghz ... seems like gpu act as a cap anyway

I'll start save some cash for a 6850 8)

3500 is damn good for a 4850... but if we both had a better graphics card the difference between our scores would be much larger.

[Katlan]

My new tech arrived today, but wont be installed until saturday. Will post my scores then!

just arrived:

Nvidia GTX 480 SuperClocked
Excellent aftermarket fan for CPU OC'ing

Currently using an i7 965 (Extreme) 3.2 ghz. I am going to OC it to at least 3.7/8, but I won't be able to OC until a week after I install the new fan since I want the thermal compound to fully set in before pushing the CPU. I know I can push the CPU higher then 3.7/8, but am lazy and that would be good enough

[ranmas]

I just love how close these scores are. I've done a bit of testing at 3.6 and 3.8ghz on the cpu, and it does not make a large difference in the score for the Phenom II 955.

@3.8 the Low score hit between 4800-4900 and the High score was between 4500-4600

720p
img690.imageshack.us/img690/7366/ffxiv720p.jpg
1080p
img705.imageshack.us/img705/3253/ffxiv1080p.jpg


I haven't seen anyone really talk about this on other forums, or this one. The scores that are listed for the benchmark are for low scores, always. Every FF benchmark that has been released was based upon the low score. The high scores were never really given a rating by SE.

That and the game is still in development. They still have quite a bit of time for improvements to the engine and we don't know what we may yet still see from SE before the game comes out. Its possible they will release another benchmark closer to the games release, or a updated one at least.

This is about the most annoying board I've ever posted on as well... This 10 posts before being able to link/post images is crap. This is the only board I've ever posted on that has had anything like that.

[Gulkeeva]

This is about the most annoying board I've ever posted on as well... This 10 posts before being able to link/post images is crap. This is the only board I've ever posted on that has had anything like that.

I know alot boards that do, because spam bots tend to like to hop on a board, post a image that "gets people cookie info" and or other relevant info to spam you even more, in the long run its more of a protection for you, not against you ;/

[naninani]

FINALLY got my HD5870 from NewEgg.

Machine:
[email protected]
ATI Radeon HD 5870

Low: 7115
http://i49.tinypic.com/2jaf9zb.png

High: 4899
http://i50.tinypic.com/2j0gr5u.png

FEELS GOOD MAN

[Cayos]

FINALLY got my HD5870 from NewEgg.

Machine:
[email protected]
ATI Radeon HD 5870

Low: 7115
http://i49.tinypic.com/2jaf9zb.png

High: 4899
http://i50.tinypic.com/2j0gr5u.png

FEELS GOOD MAN

For that high of an overclock, are you using liquid? Or is that just a normal overclock for that CPU?

[bluefan]

920s can overclock quite well on air. Granted, a good aftermarket HSF will be needed.