The OFFICIAL FFXIV Benchmark & Beta Performance Thread

[Knyghtmayre]

How much memory do you have and what processor are you running? If you run the benchmark twice in a row, does it improve your score? It should, provided you have enough available memory since the entire benchmark fits comfortably within the OS buffer cache. If I recall correctly, on a cold boot I got 27000ms from a 500GB Samsung Spinpoint 7200RPM drive. My SSDs all get around 12-16k depending upon what my processor overclock is. I can test that old 500GB drive and give you a definitive answer later if you're really interested.

AMD Phenom II X4 955 BE OC'd @ 3798Mhz
Kingston 4GB (2x2) DDR3 PC3-10600 1333Mhz
Sapphire Radeon 5750 1GB GDDR5 running Catalyst 10.6 & OC'd to 825/1250
Hitachi Deskstar 1TB 7200RPM (wasn't WD, my bad, my external is a WD)
Asus M4A78T-E Board

Just ran the benchmark 3 more times (looped):

• Card Settings: Max Quality
  Bench Settings: Low (720p)
  Score: 4000
  Load Time: 21918

• Card Settings: Max Quality
  Bench Settings: Low (720p)
  Score: 3953
  Load Time: 16494

• Card Settings: Application Controlled/High Performance
  Bench Settings: Low (720p)
  Score: 4298
  Load Time: 17030

But as you can see from my previous posting, these load times are a far cry better than:
http://i50.tinypic.com/2ur6r10.png

I don't quite understand what causes a 12k+ change D; Nothing else was running in any of these.

[Kirbyprime]

http://jonchau.com/ffxiv/images/FFXI...k_1280x720.jpg

Well, I'm planning on building a new PC sometime closer to Christmas anyways, hopefully this will at least "run" for the meantime.

[NeoForte]

Anyone with a iMac 27" i7 with a Radeon HD 4850 on Win7 64bit bootcamp got theirs to work? What drivers were you using? I got the lastest ones from ATI's site and the latest DirectX from MS and yet nothing...

[Araelus]

A reminder to take these benchmarks with a grain of salt. Ran the test just now for fun and got:

http://farm5.static.flickr.com/4116/...35ba0cf3_o.png

http://www.playonline.com/ff11eu/dow.../imgs/tx06.gif
We assume that your computer can run FINAL FANTASY XI for Windows enjoyably with the default settings.
If your video card exceeds the recommended system requirements, it may be possible to run "FINAL FANTASY XI for Windows" easily even in high resolution mode.

That computer runs FFXI at 29fps everywhere, main window 1680x1050 with maxed settings, forced supersampling, etc. etc. It should score elevafiveblaminallion points.

Now a score of n from FFXI is not the same as a score of n from FFXIV.

tier  FF14   FF11
 1    8000   7000
 2    5500   4000
 3    4500   3000
 4    3000   2500
 5    2500   2000
 6    2000   1500
 7    1500   1000
lower? fail! pay $$$ or wait until March

I'm satisfied with the following result on
LOW (1280x720) with Core 2 Quad Q6600 and ATI 4870

http://farm5.static.flickr.com/4079/...c40daba6_o.png


More grist for the mill:

Low-test FFXIV ran one core at 75%, the other three at 25%, and the GPU at 60%. However the rain scene pegged GPU activity at 100%.

High-test FFXI ran all four cores at 80%, and the GPU at 100%.

[Celeras]

You're always going to want to take benches with a grain of salt, but comparing it to the ffxi bench isn't such a good idea for a lot of reasons.. particularly the 30fps cap. Most people would consider that to be the bare minimum for comfortable gameplay in other games(such as this one), whereas in ffxi its 'maxed'. Completely different scale alltogether.

[Araelus]

You're always going to want to take benches with a grain of salt, but comparing it to the ffxi bench isn't such a good idea for a lot of reasons.. particularly the 30fps cap. Most people would consider that to be the bare minimum for comfortable gameplay in other games(such as this one), whereas in ffxi its 'maxed'. Completely different scale alltogether.

Aye, and nice to see FFXIV running at 60-70 fps. Just thought it was interesting.

EDIT: just checked with FRAPS, the in-game fps cap does not exist for the 11 bench, it runs up to 60fps... somehow.

Just wanted to show how this comp grinds out 11 bench at five fps but runs the game flawlessly. Since the 14 bench is better crafted, guess the failures of the past are irrelevant?

[alecrast]

AMD Phenom II X4 955 BE OC'd @ 3798Mhz
Kingston 4GB (2x2) DDR3 PC3-10600 1333Mhz
Sapphire Radeon 5750 1GB GDDR5 running Catalyst 10.6 & OC'd to 825/1250
Hitachi Deskstar 1TB 7200RPM (wasn't WD, my bad, my external is a WD)
Asus M4A78T-E Board

Just ran the benchmark 3 more times (looped):

• Card Settings: Max Quality
  Bench Settings: Low (720p)
  Score: 4000
  Load Time: 21918

• Card Settings: Max Quality
  Bench Settings: Low (720p)
  Score: 3953
  Load Time: 16494

• Card Settings: Application Controlled/High Performance
  Bench Settings: Low (720p)
  Score: 4298
  Load Time: 17030

But as you can see from my previous posting, these load times are a far cry better than:
Load Time: 28843

I don't quite understand what causes a 12k+ change D; Nothing else was running in any of these.

Subsequent runs of the benchmark "load" faster because they're not being loaded at all. Your load time of 28843ms is from disk. The 16k and 17k load times are because Windows still has the contents of the benchmark sitting in RAM, cached for better performance. If you disable SuperFetch and reboot, your first run of the benchmark from your mechanical hard drive will always be around 30k if nothing else is going on on your system. Subsequent runs (provided you don't do something else in between the runs that forces the benchmark to be purged from memory) will always be faster because Windows has already loaded the benchmark into RAM. Even if you exit the benchmark and then immediately restart the benchmark, Windows will still have a copy of it sitting in physical memory so that it doesn't have to go fetch it from (slow) disk. It will keep a copy in memory until something else needs that memory.

[Celeras]

Rewording my last post with a little more detail hoping for some input. Just reran the benchmark on my laptop and hers and here's the scores + specs listed. Both laptops were ran on low:

Her's

Spoiler: show

Windows 7 Premium 64-bit (6.1 build 7600)(7600.win7_gdr.100226-1909)
Intel Core i7 CPU 720QM @ 1.60GHz (8GPUs) ~1.6GHz
Nvidia GeForce 310M

SCORE:571
Load Time: 42750ms

Mine

Spoiler: show

Windows 7 Premium 64-bit (6.1 build 7600)(7600.win7_gdr.100226-1909)
Intel Core2 Duo CPU T9600 @ @2.8GHz (2CPUs) ~2.8GHz
Nvidia GeForce GTX 260M

SCORE:1619
Load Time: 13586ms

Now I know my gfx card is better than hers, but her processor should be better than mine yes? I thought running the low setting pulled mostly from the processer meaning she should at least be getting a score around me? Like I said last post I know dick about computers, but I thought her score would be at last comparable to mine.

If everything is correct and as it should be, is there anything we could do to better her laptop? I've always heard laptops are for all intents and purposes non-upgradable but is that still true today?

*EDIT* Ok, both scores are now there.

If I have to venture a guess, it's her CPU clock. You could confirm this by letting me know both of your GPU usage percentages while the bench is running.. I'm assuming they're both under 100% on low. 1.6 is an awful clock, even if it is a mobile i7. Low settings are going to scale a lot with CPU clock, it's the nature of computing. High res=GPU, Low res=CPU

(edit: it's a standard rule not hard fact, it would vary between systems and how stressed the hardware is on each setting. With a balanced system like mine I could lower my GPU clocks in half and get basically the same score on Low since my GPU isn't running anywhere near full load(small frames, easier to render)... whereas this would cripple my High score since it's at full load the majority of the time(larger frame, takes more muscle to render every frame the CPU throws out). And vice versa with CPU).

Case in point, just ran low for you:

http://img149.imageshack.us/img149/3729/cpuscaling.jpg

Note the CPU clocks. I went up instead of down for ya because going down would require me to change multipliers to keep ram/NB/HT frequencies the same... and I'm lazy (edit: was bored.. went down for ya) ;p Anddd to further hammer home the point in my technical jargon edit, here's low with my GPU at the lowest clock I can set it to (all other runs I've posted were 850/1000):

http://img816.imageshack.us/img816/1050/gpulow.jpg

Notice how the score actually fluctuated UP slightly compared to the above 2.8 ghz run despite my GPU clock being cut in half.

(I'm done editing now.)

[Emirii]

Sorry if this is the wrong place... someone mentioned load time having to do with your hard drives. My hard drives are very slow, so my load time is in the 35k+ I just bought a new motherboard that supports 6GB/s and came with the cables to do so for 3 hard drives. I have 3 hard drives, but I have them connected to the SATA right now. Do you think plugging them into the 6GB/s cable port give them an increase in speed at all? I'm not sure it will even work, I don't know the speeds of my HDDs aside from they are 7200 RPM (and probably 1.5GB/s), Thanks if anyone know!

[Kurokikaze]

Hmm.. I can't quite figure out what I did different this time around. I'm actually clocked .3GHz lower than my previous bench even though the benchmark and windows says I'm running at stock. Recently formatted but I'm sure that's got nothing to do with it... My low score jumped ~1300 while my high score went up like.. 20-30 points? Shit makes no sense.

Spoilered for huge:

Spoiler: show

http://img.photobucket.com/albums/v2...e01/newlow.png
http://img.photobucket.com/albums/v2...01/newhigh.png

Fuck it, I'll take it.

Edit: Holy fuck... While the new high changes my score but not my placement; the new low throws me from 8th to 2nd in with the 5870 crowd. This makes absolutely no sense. Wait no.. That hasn't been updated in a few weeks. lol

[Mife]

Sorry if this is the wrong place... someone mentioned load time having to do with your hard drives. My hard drives are very slow, so my load time is in the 35k+ I just bought a new motherboard that supports 6GB/s and came with the cables to do so for 3 hard drives. I have 3 hard drives, but I have them connected to the SATA right now. Do you think plugging them into the 6GB/s cable port give them an increase in speed at all? I'm not sure it will even work, I don't know the speeds of my HDDs aside from they are 7200 RPM (and probably 1.5GB/s), Thanks if anyone know!

What kinda HDD"S are you using your new mobo shouldnt speed it up in anyway and if your connected threw Sata now i doubt you will get a performance increase unless your running your HDD"s in raid 0 which is performance mode for HDD"s .. Now for your problem if you could list the maker of your HDD"S and what type CPU and GPU your using would greatly help us out in figuring out your problem or what is holding you back hope to here from you soon :D


Plus we are talking about the (MS) load time after the benchmark is done correct?

[Celeras]

Hmm.. I can't quite figure out what I did different this time around. I'm actually clocked .3GHz lower than my previous bench even though the benchmark and windows says I'm running at stock. Recently formatted but I'm sure that's got nothing to do with it... My low score jumped ~1300 while my high score went up like.. 20-30 points? Shit makes no sense.

Spoilered for huge:

Spoiler: show

http://img.photobucket.com/albums/v2...e01/newlow.png
http://img.photobucket.com/albums/v2...01/newhigh.png

Fuck it, I'll take it.

Edit: Holy fuck... While the new high changes my score but not my placement; the new low throws me from 8th to 2nd in with the 5870 crowd. This makes absolutely no sense. Wait no.. That hasn't been updated in a few weeks. lol

Did you really not notice that you didnt have Hyperthreading enabled on your original run? 4CPUs vs 8 dude lol. Though thanks for the perfect example of the point I was making on the last page ;p Helped your low score tons, but no effect on your high

And I stopped updating because people were trying to update their damn scores over and over. Only adding new hardware now.. we have a decent database for people.

[Kurokikaze]

....I actually didn't notice that. That explains a lot. Glad I helped prove a point though. lol

And yeah I figured that would be why.

[Gokulo]

Sorry if this is the wrong place... someone mentioned load time having to do with your hard drives. My hard drives are very slow, so my load time is in the 35k+ I just bought a new motherboard that supports 6GB/s and came with the cables to do so for 3 hard drives. I have 3 hard drives, but I have them connected to the SATA right now. Do you think plugging them into the 6GB/s cable port give them an increase in speed at all? I'm not sure it will even work, I don't know the speeds of my HDDs aside from they are 7200 RPM (and probably 1.5GB/s), Thanks if anyone know!

7200RMP are in no way "very slow" O.o

[Jojimbo]

Just got 6991 on Low using Xfire (x2 HD 5770), was hoping to break 7k ;;
Nice to see an actual performance boost from using the mod instead of it lowering the score lol (up from 4.4~4.5k using it normally)

[Jinfire]

Sorry in advance Im sure you get a lot of these posts.

I have tried running the benchmark program and I get an error, so anyone with similar specs tell me what can be my estimated score? Also I dont understand the score... >.> I mean whats it mean? is a 4000 good? lol.

Anyway my specs are
CPU Core i7 2.80ghz (4 cores)
Ram 6GB
OS Win7 64
Graphics ATI Radeon HD 5700 series (fairly sure its a 5770)

Now I understand I prolly wont be able to run at max settings perfectly but will it run at decent - good settings perfectly? Also if I have to upgrade anything Im assuming the best thing to upgrade would be the gpu, and I heard crossfiring 2 5700's would be cheaper / better then getting a 5800 series but I never did crossfire or anything >.> but I think upgrading to a whole new card would be easiest. (remove, put new, install drivers)

Idk I just want some advice for my rig.

[Kurokikaze]

Best place to get an idea would be the charts on the first page...

[sruon]

Sorry in advance Im sure you get a lot of these posts.

I have tried running the benchmark program and I get an error, so anyone with similar specs tell me what can be my estimated score? Also I dont understand the score... >.> I mean whats it mean? is a 4000 good? lol.

Anyway my specs are
CPU Core i7 2.80ghz (4 cores)
Ram 6GB
OS Win7 64
Graphics ATI Radeon HD 5700 series (fairly sure its a 5770)

Now I understand I prolly wont be able to run at max settings perfectly but will it run at decent - good settings perfectly? Also if I have to upgrade anything Im assuming the best thing to upgrade would be the gpu, and I heard crossfiring 2 5700's would be cheaper / better then getting a 5800 series but I never did crossfire or anything >.> but I think upgrading to a whole new card would be easiest. (remove, put new, install drivers)

Idk I just want some advice for my rig.

You should be fine, hard to give specifics before we can even play beta.

[Emirii]

What kinda HDD"S are you using your new mobo shouldnt speed it up in anyway and if your connected threw Sata now i doubt you will get a performance increase unless your running your HDD"s in raid 0 which is performance mode for HDD"s .. Now for your problem if you could list the maker of your HDD"S and what type CPU and GPU your using would greatly help us out in figuring out your problem or what is holding you back hope to here from you soon :D


Plus we are talking about the (MS) load time after the benchmark is done correct?

Yeah. All three of my HDDs are Seagate 7200 RPMs (1 500GB for backup, 1 320GB for OS, 1 320GB for extra). My CPU is an AMD Athlon II x2 260 Regor clocked at 3.4GHz, and my GPU is an ATI Radion HD 5670 1GB GDDR5. Sometimes my HDDs are slow as heck and I'm not sure if I should just save 100$ and buy a 6GB/s 1TB harddrive and use that for everything.

[Lovian]

Has there only been the 2 laptops that have got decent results? Would have thought the newer Asus Republic of Gamer ones would score quite decently.