Gigabyte Windforce OC HD 7970, my review & benchmarks

A few Tek episodes ago Logan said you guys will be reviewing an Asus 7970 card. I just bought the Gigabyte version of it. The specific model is: GV-R797TO. I had to review and benchmark it! But, I wanted to benchmark as many of today's games as possible, not just Battlefield and Crysis like every other review.

<!--break-->

I just finished writing my initial review, you can check it out here: http://www.gtmsanctuary.com/2013/02/review-gigabyte-gv-r797to-amd-radeon.html I will be adding more games to the review, but here's what I have so far. The hardware, noise, cooling & temps, Kombustor benchmarks, all that is in the initial review. I used Fraps for benchmarking.

Card specs:

Interface: PCI-E 3.0 X16

Core clock: 1100 MHz

Memory clock: 1500 MHz

Memory: 3 GB, 384-bit, DDR5

Stream processors: 2048

Output ports: 1 DVI, 1 HDMI, 2 Mini-Display

Card dimensions: Length = 11.22 in / 28.5 cm | Height = 4.96 in / 12.6 cm | Thickness = 1.5 in / 3.8 cm

 

My PC specs:

Motherboard: ASRock Z75 Pro3

Processor: Core i3 2100 (Sandy Bridge) @ 3.1 GHz

RAM: 8 GB DDR3 1333

Power Supply: Thermaltake Toughpower 675 Watts

Storage: one Samsung 840 Series 120 GB SSD (system drive), one Samsung HD103SI 1 TB HDD (1st media drive), one Seagate Barracuda 250 GB HDD (2nd media drive). NOTE: all games are stored on the 1st media drive

Case: Cooler Master HAF 912

Monitor: Samsung SyncMaster 27" @ 1080p

 

Gaming Benchmarks

 

Anno 2070

Played for 15 minutes, doing whatever. I have 2 big islands, one of which is completely covered with cities and factories.

Settings: everything maxed out, V-Sync off

Results:
Minimum frames per second = 23 fps

Maximum frames per second = 88 fps

Average frames per second = 50 fps

Notes:
Game is very playable at max settings. I did not notice any screen tearing with V-Sync off.


Batman: Arkham City

I played outside the whole time, flying around and beating down some prisoners and did one side mission (last Deadshot mission).

Settings: Anti-Aliasing at 2X MSAA, V-Sync off, DX11 features - MVSS and HBAO, everything else maxed/turned on

Results:
Minimum frames per second = 18 fps

Maximum frames per second = 64 fps

Average frames per second = 53 fps

Notes:
I did experience slight stutter a few times during my 15-minute test, but overall, the game is very playable at those settings.


Battlefield 3

I benchmarked about 20 minutes of Conquest Large on Operation Firestorm.

Settings: V-Sync off, graphics set to Ultra

Results:

Minimum frames per second = 33 fps

Maximum frames per second = 175 fps

Average frames per second = 79 fps

Notes:
Gameplay is very smooth. No lag anywhere, no stutter, no screen tearing.


Borderlands 2

I played for around 15 minutes at Caustic Caverns level, mostly just shooting and looting (what else is there to do?).

Settings: everything maxed out, V-Sync off

Results:

Minimum frames per second = 21 fps

Maximum frames per second = 67 fps

Average frames per second = 36 fps

Notes:
Even though the average fps is at a playable level, the game had quite a few major slow-downs where it would drop to about 22 fps for about 30 seconds or so. Most of the slowdown came from the extra particle effects of PhysX, but sometimes it happened at random, like right at the start of the level.


Civilization 5

Benchmarked for a while, maybe half an hour.

Settings: V-Sync off, everything at max

Results:

Minimum frames per second = 17 fps

Maximum frames per second = 225 fps

Average frames per second = 132 fps

Notes:
I haven't noticed when the framerate dropped to that crawl of 17 fps, the gameplay was pretty smooth. The only thing that definitely stood out was screen tearing. It was horrible in this game! Probably the worst I've seen in any game. It only happened when scrolling through the map while completely zoomed out. Bottom line is, play Civ 5 with V-Sync on!


Dishonored

Benchmarked around 10-15 minutes during the mission "Lady Boyle's Last Party.

Settings: V-Sync off, everything at max.

Results:

Minimum frames per second = 53 fps

Maximum frames per second = 132 fps

Average frames per second = 122 fps

Notes:
The game was very smooth during actual missions, but in the Hound Pits, the game stuttered quite a few times.


DMC: Devil May Cry

Not sure how long I actually played for benchmarking, but I fought the Succubus boss and that's it. 10 minutes maybe?

Settings: everything maxed, V-Sync off

Results:

Minimum frames per second = 26 fps

Maximum frames per second = 199 fps

Average frames per second = 102 fps

Notes:
Even though the game looks amazing, it's not a very graphically intensive game, as you can see from the average. I had the game maxed out even on my old video card - Saphire Radeon HD 6770, no lag, no stutter, some minor screen tearing with V-Sync off.


Driver: San Francisco

Benchmarked a bit over 15 minutes. Did a couple of police chases, and drove around Downtown area - drifting, jumping.

Settings: settings set to High (no individual settings in this game), V-Sync off

Results:

Minimum frames per second = 53 fps

Maximum frames per second = 121 fps

Average frames per second = 94 fps

Notes:
I benchmarked a little over 15 minutes and never saw any lag, stutter, or screen tearing.


Far Cry 3

I played for a couple of hours on these settings but benchmarked about 20 minutes of it. Ran around the island exploring, taking over outposts, hunting, and flew the hang glider once.

Settings: V-Sync off, Alpha to coverage Enhanced, SSAO = HDAO, video quality set to Ultra

Results (read notes!):

Minimum frames per second = 19 fps

Maximum frames per second = 63 fps

Average frames per second = 38 fps

Notes:
Even though the gameplay is pretty smooth for the most part, some areas present continuous lag. Such as the starting town of Amanaki. The framerate there drops to about 17 fps (MSI Afterburner showed 17, not sure why Fraps got 19). Another "issue" is that in the first hour or so of gameplay, the GPU was only utilized at around 40%. I think that was the reason for the lag, because once it got to around 80%, the lag was gone. Maybe somebody here can explain why that is. Is it like a "warmup period" for the GPU...?


Guild Wars 2

I benchmarked way over 15 minutes, probably an hour. Spent the whole time doing local missions and killing things.

Settings: V-Sync off, all settings at max

Results:

Minimum frames per second = 21 fps

Maximum frames per second = 81 fps

Average frames per second = 50 fps

Notes:
Despite the very acceptable average framerate, the gameplay was laggy from time to time.


Hawken

Played one whole match of Missile Assault. 15-20 minutes.

Settings: V-Sync off, everything maxed out, PhysX particles on. NOTE: there doesn't seem to be any AA settings

Results:

Minimum frames per second = 11 fps

Maximum frames per second = 60 fps

Average frames per second = 32 fps

Notes:
The game felt laggy all throughout. The significant drop in framerate was always right after respawning.


Saints Row 3

I played for just over 15 minutes, mostly blowing stuff up with a tank. After dying, I drove around the city in a fast car. Then again started blowing stuff up with a tank.

Settings: AA at 2X, everything else is maxed

Results:

Minimum frames per second = 26 fps

Maximum frames per second = 124 fps

Average frames per second = 45 fps

Notes:
Gameplay is very smooth, no lag, no stutter, no screen tearing.


Sleeping Dogs

Played for just under 15 minutes. Drove around the city in a fast car (class A). Did a bike race, which I failed 3 times in a row, which caused me to go on a rampage shooting everything that moves.

Settings: V-Sync off, Frame limiter off, HD textures on, everything else is maxed.

Results:

Minimum frames per second = 21 fps

Maximum frames per second = 48 fps

Average frames per second = 33 fps

Notes:
Game played very smoothly, no issues of any kind.


Skyrim

Wanted to benchmark the standard 15 minutes, got sucked into the game, ended up benchmarking about an hour. I think... Just did the usual Skyrim stuff, exploring, looting, killing, admiring the scenery, rearranging dead bodies in sexual positions...

Settings: V-Sync off, all other settings maxed out, official HD texture used, Dynamic & realistic lighting mod used,

Results:

Minimum frames per second = 16 fps

Maximum frames per second = 65 fps

Average frames per second = 58 fps

Notes:
Despite the high average fps, Skyrim always seems to have micro-stutter. It's faint, but still noticeable.

 

Would appreciate any kind of input on the card, the review, the results, whatever. I've only recently discovered your guys' YouTube channel and quickly became a huge fan of it. The Tek episodes are awesome!

I'd also love to hear what others think about just how much the framerate bottlenecked due to my Core i3 2100 and 8 gigs of 1333 RAM. I asked this question on Tom's Hardware but didn't get any solid answers. But now I have the results...

A very well done review! I like how you included your thoughts, very professional and clean. As soon as I saw your CPU I immediately though "Bottleneck!" Your RAM doesn't matter though since you can easily OC that. I know some of these games can be very intensive on your CPU, GW2 for example, so you could of seen different results if you were using a i5-3570k or FX-8350.

Yeah, I know about the bottleneck. I figured I'd wait to see what Intel has with their Haswell line before upgrading. I know they said they'll mostly focus on Ultrabooks and laptops with that line, but by the time they are released, I'm sure Ivy Bridge CPU will get cheaper.

It's my understanding, correct me if I'm wrong, that PhysX uses the CPU instead of a video card on AMD cards since it's not built in. That's probably the reason for lag in Borderlands 2 and Hawken. GW2 is just a dense game. A lot of stuff happeneing all over the place.

 

Also, finished benchmarking Trine 2

Trine 2

Benchmarked around 15 minutes.

Settings: AA set to Extreme, V-Sync off, Graphics detail level set to Very High

Results:

Minimum frames per second = 42 fps

Maximum frames per second = 76 fps

Average frames per second = 60 fps

Notes:
Game is very playable at max setting at 1080p. No lag, no stutter, I might have noticed screen tearing once, but it was very faint.

 

UPDATE: Metro 2033 benchmarks finished as I waited for Planetside 2 to update!

Metro 2033
I benchmarked about 20 minuted at the Dead City level (I'm playing it in Russian, so it might be called something different).

Settings: V-Sync off, graphics set to Very High, Advanced physics on, AA set to AAA

Results:

Minimum frames per second = 4 fps (Fraps caught a half-second hiccup)

Maximum frames per second = 76 fps

Average frames per second = 52 fps

Notes:
Surprisingly the game was not as laggy as I thought it would be. Actually, it was pretty smooth for the most part. There were a few laggy areas, but framerate never really dropped to a crawl. It did have a few stutter moments, which is where that single-digit minimum came from. My main issue with performance of this game was screen tearing. It's very noticeable in Metro 2033.

 

Gonna benchmark Planetside 2 next

I wouldn't expect much out of Haswell, I asked the same question and everyone said that it was just gonna be a less power consuming chip with a very tiny improvement to performance. And yes, PhysX with AMD cards does use the CPU, if you have a good cpu then you shouldn't see much stutter. I saw a display of PhysX on BL2 with an i5 and a 7870 and it was very smooth.