dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
2207
share rss forum feed


Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12

My Crysis 3 benchmarks

I decided to do a couple different tests to see how my rig (with an aging i7 CPU) performed under Crysis 3 via FRAPS benchmarking. Feel free to do your own and post back!

This is not going to be up to professional benchmarking standards in terms of identically repeatable results, but I did my best to be methodical in my approach:

Instructions
• Post the settings you are using!
• Load the same save and follow the same path back and forth for the duration of the benchmarks.

My rig

Case: Cooler Master HAF 922
Mobo: EVGA x58 SLI Mobo
CPU: i7 920 quad-core w/HT @ 2.80 GHz
CPU cooler: Corsair Hydro-100 Integrated Water Loop
Memory: 24GB GSkill DDR3 1333
Boot Drive: Samsung 840 500GB SSD
Storage: 1.5 TB RAID 0 (2x WD Caviar Black 750GB 7200RPM), 1.5 TB Samsung Spinpoint F3
NAS: 1500GB Raid 1 (2x 1500GB WD Greens)
Optical Drive: LG DVDRW Litescribe
GPU: Sapphire Vapor-X Radeon 7970 3GB
PSU: Corsair Professional 750W PSU Gold, Modular
OS: Windows 7 Professional 64 bit
Mouse: Razer Naga Molten
Keyboard: CM Storm Trigger - Red Cherry MX
Monitor 1: AURIA EQ276W 27" IPS LED 2560 x 1440
Monitor 2: Apple Cinema Display 20" LCD 1680 x 1050

I ran two tests... "Max settings" (with everything in Advanced Graphics absolutely maxed, with 8x MSAA), and "Playable settings," which was FSAA + the following Advanced Graphics settings:









Takeaways: A single 7970 is certainly capable of driving the game at 2560x1440 if you tone down the goodies. Even with my i7-920, I was able to pull down 40.6 FPS average under what I called "Playable Settings." But my GPU choked under Max Settings, putting out an unplayable 17.4 FPS average. I'm certain a current generation CPU might help a tad, but I am already running under 8 threads, so I am not certain if it would be a night-and-day difference, or merely add a couple FPS to the average. A second 7970 is almost certain to be required to smoothly run this game at its max potential, regardless of your processor.

There's my piece. Have you done any similar Crysis 3 benchmarks? If so, please feel free to post them, but always remember to give us system specs, your testing methodology, and which settings you used!
--
Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety.


Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12

Tom's results for 16 cards: »www.tomshardware.com/reviews/cry···451.html



Ghastlyone
Premium
join:2009-01-07
Las Vegas, NV
kudos:5
reply to Krisnatharok

It's interesting that Tom's benchmarked an i5 3550 and not the 3570k instead.

I noticed that a single 680 at 1080P was averaging 30-40fps. Is this game that demanding, or just not optimized well yet?



Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12

It's that demanding, notice in Tom's reviews for the 16 cards, they don't even attempt a "max settings" test. They turn stuff down.
--
Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety.



Moos
Tequilablob
Premium
join:2008-12-11
Salt Lake City, UT
kudos:3
reply to Krisnatharok

Ive gathered all the info for my setup, i'll put it in some charts and post the results as soon as I have some time.



kvn864

join:2001-12-18
Sun City, AZ
kudos:1
reply to Krisnatharok

I didnt have enough time to run real benchies, but what I have seen just by running a game and glancing on FRAPS. At medium settings with AAx2, it was doing at fairly playable frames of 35-50. And the game looks absolutely beautiful even at those settings. I didnt see lots of a difference (lookwise) by going AAx2 to x4 but frame rates dropped by 5-10. However I can tell degradation in picture quality by disabling AA all together, to me even if x2 is enabled looked good.
This is on: Intel® Xeon® Processor E3-1240 (8M Cache, 3.30 GHz), Radeon 5870, 24G of RAM for my virtual stuff. All default clocks.
And last: great game. Thanks Kris for the effort.



Gordo74
Premium
join:2003-10-28
Monroeville, PA
reply to Krisnatharok

Do the same benchmark on your new machine. I want to see how it compares.



Dissembled

join:2008-01-23
Indianapolis, IN

said by Gordo74:

Do the same benchmark on your new machine. I want to see how it compares.

+1


Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12

Oh I plan to, so I have a benchmark to use 7970 CF against when my card comes.

Hopefully tonight or tomorrow.



Gordo74
Premium
join:2003-10-28
Monroeville, PA

said by Krisnatharok:

Oh I plan to, so I have a benchmark to use 7970 CF against when my card comes.

Hopefully tonight or tomorrow.

No no. I mean the exact benchmark with only one video card in your new machine. I want to see how big of a difference it was between 1366 and Ivy 1155.

CF will be interesting too, don't get me wrong, but at that point you're comparing apples to oranges.


Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12

said by Gordo74:

said by Krisnatharok:

Oh I plan to, so I have a benchmark to use 7970 CF against when my card comes.

Hopefully tonight or tomorrow.

No no. I mean the exact benchmark with only one video card in your new machine. I want to see how big of a difference it was between 1366 and Ivy 1155.

CF will be interesting too, don't get me wrong, but at that point you're comparing apples to oranges.

I plan to do both.
--
Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety.


Ghastlyone
Premium
join:2009-01-07
Las Vegas, NV
kudos:5
reply to Gordo74

said by Gordo74:

said by Krisnatharok:

Oh I plan to, so I have a benchmark to use 7970 CF against when my card comes.

Hopefully tonight or tomorrow.

No no. I mean the exact benchmark with only one video card in your new machine. I want to see how big of a difference it was between 1366 and Ivy 1155.

CF will be interesting too, don't get me wrong, but at that point you're comparing apples to oranges.

That's what I was thinking also. I want to see a single card performance.

After that, plug that Crossfire setup in, and go to town.


Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12
reply to Gordo74

In the meantime, consider the Unigine Valley difference--16.41% between the two with the same GPU:

x58 (use the second one from the top--score is 1792):
»/r0/download/2···2450_881

Z77:
»/r0/download/2···core.jpg
--
Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety.



Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12
reply to Gordo74

I did some comparisons for ya'll:

»As requested, Crysis 3 benchmarks



Gordo74
Premium
join:2003-10-28
Monroeville, PA
reply to Krisnatharok

Thanks for the follow up Kris.

It confirms what I thought: If you aren't going to use dual GPUs, there is no point in going form x58 to z77 at this point as far as gaming performance is concerned.

If you feel like being REALLY ambitious, and if your X58 mobo supports it, I'd love to see the dual GPU config in the i7 920 to see what it shows.



Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12

The x58 would perform at x16/x16 under PCIe 2.0, which is the same bandwidth as x8/x8 under PCIe 3.0, so I doubt you'd see any difference, except maybe from the CPU throttle the cards (I could try forcing the 3770K to downclock to 2.8 GHz and seeing what I lose under Crossfire).

I won't be ripping the build apart, the x58 now has a 600w PSU and a 560 Ti and is functioning as the wife's VM machine (24 GB ram + 8 threads make it well suited for that), so she wouldn't appreciate me destroying both machines just to test a hypothetical.

That said, I am extremely impressed by by the scaling of the 7970s--considering the margin for human error in the tests (did I look out the window too long or not long enough?) I'd say it's exactly 100%, which is extremely hard to do.
--
Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety.



Gordo74
Premium
join:2003-10-28
Monroeville, PA

said by Krisnatharok:

The x58 would perform at x16/x16 under PCIe 2.0, which is the same bandwidth as x8/x8 under PCIe 3.0, so I doubt you'd see any difference, except maybe from the CPU throttle the cards (I could try forcing the 3770K to downclock to 2.8 GHz and seeing what I lose under Crossfire).

So... then why did you build a new machine?


Krisnatharok
Caveat Emptor
Premium
join:2009-02-11
Earth Orbit
kudos:12

A bunch of reasons: I wanted to (one of the biggest), I couldn't Crossfire this card on my old board, native USB3/SATA3 (for the SSD), I was tired of looking at my case, dealing with the very spartan EVGA BIOS (and lack of features), and I wanted an unlocked chip to push. My goal is to eventually watercool this rig (probably a 2014 project) and push the CPU to 5 GHz and the GPUs each to 1250mhz.

Granted, no one "needs" to rebuild if all they do is game and they have a first gen i7. But very few of us in this game are driven by need, more by want.
--
Those who would give up Essential Liberty to purchase a little Temporary Safety, deserve neither Liberty nor Safety.



Ghastlyone
Premium
join:2009-01-07
Las Vegas, NV
kudos:5
reply to Gordo74

said by Gordo74:

said by Krisnatharok:

The x58 would perform at x16/x16 under PCIe 2.0, which is the same bandwidth as x8/x8 under PCIe 3.0, so I doubt you'd see any difference, except maybe from the CPU throttle the cards (I could try forcing the 3770K to downclock to 2.8 GHz and seeing what I lose under Crossfire).

So... then why did you build a new machine?

He should be getting about a 15% performance improvement from that 1st Gen Intel to the new Ivy Bridge.


Moos
Tequilablob
Premium
join:2008-12-11
Salt Lake City, UT
kudos:3
Reviews:
·Comcast
reply to Krisnatharok

Here are my results using various settings.




I started each test from the same game save and attempted to do the exact same things during the game play which included alot of combat and running around. I averaged the data over approximately 5 minutes of gameplay for each test. Data was collected with MSI afterburner logs. These results were using AMD catalyst 13.1 beta6.

My specs are as follows:
Case: Corsair Carbide 500R
CPU: I5-2500k @ 4.4 Ghz
Mainboard: Gigabyte z68XP-UD3
Memory: 16GB corsair vengeance @ 1600
SSD: 120GB OCZ-Vertex3 (Main Boot Drive with Windows 7)
HDD: 2TB Seagate Barracuda 7200RPM. (Games are stored here including Crysis 3)
GPU: Crossfire 7950's. Gigabyte windforce @ 1125Mhz and 1400Mhz memory.
PSU: Seasonic X-series 850w.
Mouse: Steelseries Sensei
Keyboard: Logitech G15
Gamepad: Razer Nostromo
Monitors 1-2: Asus VH238H 1920x1080
Monitor 3: Crossover 27Q-P 2560x1440 (Main gaming monitor)