GTX 970 Roundup (EVGA, GALAX, Gigabyte)

[SKYMTL]

NVIDIA’s GTX 970 has received quite a bit of face time lately and with good reason. When paired up with a good heatsink and some slightly higher clock speeds, it can give the GTX 980 a run for its money. Meanwhile, in SLI the inexpensive GTX 970 provides an amazing amount of performance relative to the setup’s cost of less than $700.

Part of the GTX 970’s allure is the wide variety of cards NVIDIA’s board partners have launched. Since this is a “virtual” design, they’ve been given a free hand to stretch the limits of board design provided minimum specifications are met. As a result we’ve seen everything from typical blower-style setups to some wild cooling setups and exceedingly high clock speeds. All of these are priced well under $400 which makes them extremely good values.

In order to get a better grip on how some of these board partner-specific versions line up against one another, we decided to get a few into the lab and test them for ourselves. The trio of cards in this roundup represents a relatively small sampling of what’s available today but they’re some of the most popular around. EVGA, GALAX (more on them later) and GIGBAYTE have all contributed but ASUS and PNY's GTX 970 will still be represented in the performance charts.

Before we get too deep into the specifications for each of these cards, let’s explain who GALAX is. Basically GALAX was created as an international amalgamation between the original Galaxy brand and their KFA2 offshoot. According to our conversation with their representatives, this will give global customers full access to both companies’ full product lineups whereas before certain designs were reserved for different regions. Supposedly this will improve their overall customer support and availability but there is one caveat: these cards will only be available through GALAX’s online storefront. That’s good news for pricing since they are effectively cutting out the distribution channel but it also raises questions about cross-border shipping costs.

Speaking of those shipping costs, they aren’t free and Canadian buyers will have to pay taxes at the border since GALAX won’t pre-clear you package. That means a shipping cost of about $20 alongside the usual taxes from your province and a $15-$25 brokerage handling charge for customs clearance. Needless to say, after all of this is taken into account, Canadian customers won’t be saving any money versus buying from a local etailer.

With that out of the way let’s start with the EVGA GTX 970 SC ACX 2.0 or the artist also known as Superclocked in EVGA’s lineup. Despite it not even being close to the highest clocked card in their GTX 970 stack (the SSC and FTW are even faster) the SC close to the top of the hill in this review, at least in terms of raw frequencies. It is however tied with the Gigabyte card for most expensive at $349. There have been some minor complaints about EVGA’s new ACX 2.0 heatsink design’s fan profiles (particularly when at idle) but a new BIOS was launched to take care of many of those concerns.

GALAX’s entry into this roundup is their GTX 970 EX OC which boasts clock speeds that lie somewhere between the ASUS DirectCU II OC and Gigabyte G1. Gaming and EVGA’s SC. Its price reflects these qualities as well but in terms of actual onscreen differentiation between these cards, there likely won’t be much considering they’re all rated to within about 100MHz of one another. However, we wouldn’t be surprised if GALAX’s unique design and the EX OC’s very aggressive $329 price point draw in buyers.

The $349 Gigabyte G1 Gaming is arguably one of the most talked-about GTX 970’s currently on the market. It is big, brash and comes in two versions: the “standard” WindForce 3X OC edition which is actually clocked substantially higher than NVIDIA’s reference specifications dictate and the G1 Gaming that costs a good $20 more. With that being said, it may be one of the more expensive GTX 970’s around, but this card also boasts some awesome clock speeds.

The specifications of EVGA, GALAX and Gigabyte may not allow you to see any onscreen framerate differences; their respective lengths are wildly varied. Gigabyte’s G.1 Gaming is an absolute behemoth at just over 11 ¾” long and it likely won’t fit into many smaller chases. Meanwhile, GALAX’s EX OC rings in at 10 ¼” which broadens its compatibility and EVGA’s is by far the most compact of the bunch at just 9 ½”.

While there are plenty of points of differentiation, some things are becoming pretty standard in the industry these days. It shouldn’t come as any surprise that every one of these cards has a 3 year warranty and strictly constrained memory overclocks. Availability is also a concern among every entry since production still hasn’t picked up to a point where it’s in line with actual demand. If that particular situation doesn’t improve by the Christmas rush, these board partners and NVIDIA will be leaving a lot of money on the table.

EVGA, GALAX and Gigabyte all have pretty compelling reasons as to why their particular card is the one to buy. Considering they’re priced within spitting distance of one another and overclocking typically varies from one sample to another, only overall performance stability, temperatures and acoustics will sway consumers one way or another.

 

[SKYMTL]

EVGA's GTX 970 SC ACX 2.0

EVGA GTX 970 SC ACX 2.0

EVGA’s GTX 970 SC comes in two different forms: one with the slightly older ACX cooler and the ACX 2.0-equipped one we see above. While there have been some minor updates to the internal heatsink design through fin array and heatpipe placement, the main improvements have been rolled into the fans. They now feature an advanced 11-fin design with swept blades to better direct the airflow, a reinforced magnetic o-ring, an updated motor and double ball bearings. Supposedly this leads to a drastically enhanced lifespan and better cooling performance than the previous design.

<object width="640" height="360"><param name="movie" value="//www.youtube.com/v/JERkKJvcDNM?hl=en_US&version=3&rel=0&controls=0"></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="//www.youtube.com/v/JERkKJvcDNM?hl=en_US&version=3&rel=0&controls=0" type="application/x-shockwave-flash" width="640" height="360" allowscriptaccess="always" allowfullscreen="true"></embed></object>​

Initially some complained about the performance of EVGA’s AXC 2.0 heatsink and there were erroneous reports that pointed towards the HDT-style base not making proper contact with the small GM204 core. This couldn’t be further from the truth since the ACX 2.0 contact plate is optimally placed for efficient heat transfer. EVGA did however trace the initial teething problems back to their fan speed profiles, which have since been fixed through a BIOS update. This essentially shuts off the fans in most idle situations (though it only happened a few times in our testing in Windows 8) and optimizes both temperatures and fan speeds when gaming.

As with all of the other board partners, EVGA has used their own in-house PCB for their GTX 970. There isn’t anything particularly interesting back here other than the presence of 2GB worth of Samsung K4G41325SC GDDR5 modules. These are present on nearly every GTX 970 we have come across thus far and are relatively good overclockers.

EVGA is using a standard 6+6 pin layout for their power input but this shouldn’t negatively affect overclocking since there’s plenty of current overhead given the core’s efficiency. Meanwhile, the backplate is pretty straightforward, though it lacks native support for more than one 4K display unless you use a secondary DisplayPort MST hub. Its DisplayPort, HDMI and dual DVI outputs do provide more than enough options for the vast majority of gamers though.

 

[SKYMTL]

GALAX’s GTX 970 EX OC

GALAX’s GTX 970 EX OC

The original Galaxy brand was known as one of the few NVIDIA board partners that was willing to go out on a limb and try some innovative designs. That mantra was eventually carried over and expanded upon within the KFA2 lineup so the expectations were pretty high expectations for the GALAX EX OC despite its lower $329 USD price (or about $415CAD when everything is added up). Luckily it met many of those, at least from its exterior design perspective.

The heatsink is covered by an anodized aluminum shroud which is a significant departure from the somewhat cheap-feeling plastic that some other board partners use. There is also a pair of directed 80mm fans which have been engineered to run at minimal RPMs without sacrificing cooling performance.

<object width="640" height="360"><param name="movie" value="//www.youtube.com/v/Xr-ByCuML1Y?version=3&hl=en_US&rel=0&controls=0"></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="//www.youtube.com/v/Xr-ByCuML1Y?version=3&hl=en_US&rel=0&controls=0" type="application/x-shockwave-flash" width="640" height="360" allowscriptaccess="always" allowfullscreen="true"></embed></object>​

Those fans also feature GALAX’s Silent Extreme Technology which disables one of the card’s two fans when the BIOS detects an idle or low power usage scenario. While some other cards have exhibited similar behavior, the EX OC actually pulses the second fan to maintain temperatures in higher load situations rather than simply enabling it outright. This insures excellent acoustics but it will also be something that could impact the results in our Performance Over Time testing.

​

The EX OC’s heatsink is a pretty impressive affair that is no wider or longer than the heatsink itself. It features a quartet of large heatpipes and a polished aluminum base plate for better heat transfer. The only negative that we can see is that small form factor users won’t appreciate its lack of a direct exhaust design but the GTX 970’s core doesn’t output all that much heat to begin with anyways.

GALAX has also installed an anodized aluminum backplate which perfectly hugs the PCB’s contours. It not only looks great but it also disperses the heat from the rear-mounted GDDR5 modules. For those wondering, the EX OC uses a 5+2 phase PWM.

​

Input power is handled by an 8+6 pin layout while the backplate is a simple affair with a pair of DVI connectors, a DisplayPort and a HDMI output. There’s also an enlarged grille for exhaust airflow but the amount of air that actually escapes is minimal at best.

 

[SKYMTL]

Gigabyte’s GTX 970 G1 Gaming

Gigabyte’s GTX 970 G1 Gaming

With its insanely large heatsink, the Gigabyte GTX 970 G1 Gaming cuts an imposing figure given that many GTX 970’s have been moving towards more compact form factors. Granted, this design’s nearly 12” length will limit its compatibility with some cases but if you want the ultimate in cooling performance, look no further.

Unlike its lower-priced sibling, the G1 Gaming receives a few additional features as well. Being part of the Super Overclocked family, its cores are thoroughly screened for the most overclocking headroom possible. While that may not be all that important given the G1 still adheres to NVIDIA’s Power and Voltage limitations, it could be handy if custom BIOSes ever become available.

Another slightly less advertised inclusion is a dual BIOS layout. Unlike some cards where these can be automatically switched between, Gigabyte’s GTX 970 will detected if a BIOS flash has failed and will then automatically load from the stored backup to insure functionality. This may not be a useful feature for most gamers but overclockers will use it as a great failsafe if a flash goes wrong.

The WindForce 3X heatsink design has gone through many different iteration but this one is perhaps the most impressive. Despite the GM204 core being quite efficient, this cooler runs the entire card’s length and then some, providing space for a trio of large 90mm fans. There are also several other secondary heatsinks that are installed onto critical components like the PWM and memory modules.

Flipping the card around we see a full-length backplate that includes slits for optimal airflow. In order to retain a relatively clean looked design, that backplate continues past the PCB’s edge and out into thin air in an effort to align with the WindForce 3X heatsink. That does cause a bit of a problem….

With such a thin strip of metal that is supported by only two small cross-members (for some reason Gigabyte decided to give this area an “airy” look by stopping the main backplate in line with the PCB), you will invariably get flexing in this area. Our card had its overhang bent at a pretty alarming angle but it was easily returned to normal. From a design standpoint this may have seemed like a good idea but in practice the overhand looks pretty cheap and unfinished.

​

Power input is accomplished with a 6+8 pin design. These connectors are actually inverted so the heatsink can remain as low as possible to the PCB while maximizing vertical fin area.

The I/O panel is another unique area on this card. Gigabyte claims their Flex Display technology can automatically detect and set up display groups of up to four monitors (three primaries plus one accessory display). At first glance this doesn’t sound any different from NVIDIA’s usual Surround technology but the G1 Gaming can do its thing while utilizing a combination of DisplayPort, HDMI and DVI.

 

[SKYMTL]

Test System & Setup

Main Test System

Processor: Intel i7 4930K @ 4.7GHz
Memory: G.Skill Trident 16GB @ 2133MHz 10-10-12-29-1T
Motherboard: ASUS P9X79-E WS
Cooling: NH-U14S
SSD: 2x Kingston HyperX 3K 480GB
Power Supply: Corsair AX1200
Monitor: Dell U2713HM (1440P) / ASUS PQ321Q (4K)
OS: Windows 8.1 Professional


Drivers:
AMD 14.7 Beta
NVIDIA 344.07 Beta


*Notes:

- All games tested have been patched to their latest version

- The OS has had all the latest hotfixes and updates installed

- All scores you see are the averages after 2 benchmark runs

All IQ settings were adjusted in-game and all GPU control panels were set to use application settings

The Methodology of Frame Testing, Distilled

How do you benchmark an onscreen experience? That question has plagued graphics card evaluations for years. While framerates give an accurate measurement of raw performance , there’s a lot more going on behind the scenes which a basic frames per second measurement by FRAPS or a similar application just can’t show. A good example of this is how “stuttering” can occur but may not be picked up by typical min/max/average benchmarking.

Before we go on, a basic explanation of FRAPS’ frames per second benchmarking method is important. FRAPS determines FPS rates by simply logging and averaging out how many frames are rendered within a single second. The average framerate measurement is taken by dividing the total number of rendered frames by the length of the benchmark being run. For example, if a 60 second sequence is used and the GPU renders 4,000 frames over the course of that time, the average result will be 66.67FPS. The minimum and maximum values meanwhile are simply two data points representing single second intervals which took the longest and shortest amount of time to render. Combining these values together gives an accurate, albeit very narrow snapshot of graphics subsystem performance and it isn’t quite representative of what you’ll actually see on the screen.

FCAT on the other hand has the capability to log onscreen average framerates for each second of a benchmark sequence, resulting in the “FPS over time” graphs. It does this by simply logging the reported framerate result once per second. However, in real world applications, a single second is actually a long period of time, meaning the human eye can pick up on onscreen deviations much quicker than this method can actually report them. So what can actually happens within each second of time? A whole lot since each second of gameplay time can consist of dozens or even hundreds (if your graphics card is fast enough) of frames. This brings us to frame time testing and where the Frame Time Analysis Tool gets factored into this equation.

Frame times simply represent the length of time (in milliseconds) it takes the graphics card to render and display each individual frame. Measuring the interval between frames allows for a detailed millisecond by millisecond evaluation of frame times rather than averaging things out over a full second. The larger the amount of time, the longer each frame takes to render. This detailed reporting just isn’t possible with standard benchmark methods.

We are now using FCAT for ALL benchmark results, other than 4K.

 

[SKYMTL]

Analyzing Performance Over Time

Analyzing Performance Over Time

As we already know, NVIDIA cards react in a particular way to temperatures. The Boost algorithms attached to every core penalize high temperatures (typically anything north of 80°C) with lower clock speeds while keeping temperatures under that point allows for increased frequencies provided voltage and power constraints oblige. While there may be a ton of factors being taken into account for a Maxwell-based card to spit out satisfactory performance, the equation for gamers is relatively simple: reduce temperatures to potentially boost performance.

With every one of the cards in this roundup boasting some form of custom cooling solution, they should be able to hit their maximum stated Boost speeds and even exceed them in normal gameplay conditions. Another interesting aspect of these cards is their default voltage, each of which is slightly different in order to achieve advertised frequencies.

As expected, not one of these custom cards gets anywhere close to NVIDIA’s self-imposed 80°C temperature limiter and performance across the board is actually quite impressive given the clock speeds these GTX 970’s are running at.

EVGA’s ACX 2.0 heatsink is arguably the most impressive of the bunch not because of its temperatures but rather what it can achieve despite having such a compact footprint. It remained steadfastly at 64°C throughout testing, even after hours of gaming.

GALAX was a bit of a dark horse here since we didn’t quite know what to expect. With temperatures hovering around the 62°C, it was more than able to keep up with the competition.

Believe it or not, we’re mostly ambivalent about Gigabyte’s performance here. On one hand it’s just awesome to see a pre-overclocked GTX 970 hitting temperatures below the 55°C mark. However, the G1 Gaming has a ridiculously-sized cooler and let’s be honest, there’s a point of diminishing returns since frequencies will only boost so far, even when (limited) additional voltage is applied. If you want what is likely the best air-cooled thermal solution around on your GTX 970, look no further than this one but if a more compact setup is on the horizon you may want to look elsewhere.

With such low temperatures from every card, clock speeds are of course completely stable. Gigabyte leads the pack with a continual 1356MHz which has been achieved with 1.225V coursing through the core. That’s a full 100MHz higher than the reference cards we’ve tested and actually bests Gigabyte’s own stated maximum Boost frequency of 1329MHz.

Meanwhile, both EVGA and GALAX require 1.2V to achieve their clocks with the Superclocked lagging behind Gigabyte’s monster by just 30MHz. GALAX isn’t all that much faster than the reference cards we’ve tested but it remains blissfully consistent, unlike those selfsame reference designs.

As they say: performance is what it is. Every one of these cards delivers framerates that align perfectly with their relative clock speeds, beating the reference design by sometimes wide margins.

 

[SKYMTL]

Thermal Imaging / Acoustics / Power Consumption

Thermal Imaging

Starting things off with the EVGA GTX 970 SC ACX 2.0 we can see that its relatively diminutive heatsink is fully capable of cooling off the overclocked GM204 core without adversely affecting component temperatures. There are some slight hot spot on the PCB’s underside but that’s to be expected with GDDR5 modules running for all they’re worth. Even then, there’s nothing to be concerned about here.

GALAX’s card does have a few more hot spots than EVGA’s but that doesn’t mean temperatures are running amok. On the contrary; everything remains cool and the backplate allows for additional memory module heat dissipation.

Anything less than absolute perfection here would have been a major blow to Gigabyte since their entire goal for the G1 Gaming was to design the highest performance cooling solution possible. It met those expectations and then some. Temperatures across the card were literally perfect despite it being the highest clocked in this roundup.

Acoustical Testing

What you see below are the baseline idle dB(A) results attained for a relatively quiet open-case system (specs are in the Methodology section) sans GPU along with the attained results for each individual card in idle and load scenarios. The meter we use has been calibrated and is placed at seated ear-level exactly 12” away from the GPU’s fan. For the load scenarios, Hitman Absolution is used in order to generate a constant load on the GPU(s) over the course of 15 minutes.

Let’s start things off by talking about EVGA since their ACX 2.0 cooler initially met with a good amount of critique for higher than expected fan noise. With the latest BIOS (an easy upgrade for anyone who previously bought this card) things are brought to heel with idle noise being nonexistent since, like the ASUS STRIX, the fans completely shut off in many idle scenarios. Granted, our Windows 8 install with a fully animated Metro UI did cause spin-up in some cases but for the most part, the SC’s silence was golden. Even when operating at full load, there was nothing to complain about.

GALAX started off with a blank slate since there were no expectations attached to their GTX 970 EX OC. However, they proved a careful hand in heatsink design can achieve some great results and the EX OC became one of the quietest cards we’ve tested.

The GTX 970 G1 Gaming is nothing short of a revelation. It boasts extremely low temperatures alongside what are nearly the lowest acoustical results we’ve ever seen. Really, there isn’t anything more to say about this achievement.

System Power Consumption

For this test we hooked up our power supply to a UPM power meter that will log the power consumption of the whole system twice every second. In order to stress the GPU as much as possible we used 15 minutes of Unigine Valley running on a loop while letting the card sit at a stable Windows desktop for 15 minutes to determine the peak idle power consumption.

Please note that after extensive testing, we have found that simply plugging in a power meter to a wall outlet or UPS will NOT give you accurate power consumption numbers due to slight changes in the input voltage. Thus we use a Tripp-Lite 1800W line conditioner between the 120V outlet and the power meter.

Overclocking leads to higher power consumption but regardless of clock speeds, the GTX 970 remains the best possible option in the performance per watt category.

 

[SKYMTL]

Assassin’s Creed IV: Black Flag / Battlefield 4

Assassin’s Creed IV: Black Flag

<iframe width="640" height="360" src="//www.youtube.com/embed/YFgGnFoRAXU?rel=0" frameborder="0" allowfullscreen></iframe>​

The fourth iteration of the Assassin’s Creed franchise is the first to make extensive use of DX11 graphics technology. In this benchmark sequence, we proceed through a run-through of the Havana area which features plenty of NPCs, distant views and high levels of detail.


2560 x 1440

Battlefield 4

<iframe width="640" height="360" src="//www.youtube.com/embed/y9nwvLwltqk?rel=0" frameborder="0" allowfullscreen></iframe>​

Amidst its teething problems since its release, BF4 has been a bone of contention among gamers. In this sequence, we use the Singapore level which combines three of the game’s major elements: a decayed urban environment, a water-inundated city and finally a forested area. We chose not to include multiplayer results simply due to their randomness injecting results that make apples to apples comparisons impossible.

2560 x 1440

 

[SKYMTL]

Call of Duty: Ghosts / Far Cry 3

Call of Duty: Ghosts

<iframe width="640" height="360" src="//www.youtube.com/embed/gzIdSAktyf4?rel=0" frameborder="0" allowfullscreen></iframe>​

The latest Call of Duty game may have been ridiculed for its lackluster gameplay but it remains one of the best-looking games out there. Unfortunately due to mid-level loads, getting a “clean” runthrough without random slowdowns is nearly impossible, even with a dual SSD system like ours. Hence why you should ignore any massive framerate dips as they are anomalies of poor loading optimizations. For this benchmark we used the first sequence of the 5th Chapter entitled Homecoming as every event is scripted so runthroughs will be nearly identical.

2560 x 1440

Far Cry 3

<iframe width="560" height="315" src="http://www.youtube.com/embed/mGvwWHzn6qY?rel=0" frameborder="0" allowfullscreen></iframe>​

One of the best looking games in recent memory, Far Cry 3 has the capability to bring even the fastest systems to their knees. Its use of nearly the entire repertoire of DX11’s tricks may come at a high cost but with the proper GPU, the visuals will be absolutely stunning.

To benchmark Far Cry 3, we used a typical run-through which includes several in-game environments such as a jungle, in-vehicle and in-town areas.


2560 x 1440

 

[SKYMTL]

Hitman Absolution / Metro: Last Light

Hitman Absolution

<iframe width="560" height="315" src="http://www.youtube.com/embed/8UXx0gbkUl0?rel=0" frameborder="0" allowfullscreen></iframe>​

Hitman is arguably one of the most popular FPS (first person “sneaking”) franchises around and this time around Agent 47 goes rogue so mayhem soon follows. Our benchmark sequence is taken from the beginning of the Terminus level which is one of the most graphically-intensive areas of the entire game. It features an environment virtually bathed in rain and puddles making for numerous reflections and complicated lighting effects.


2560 x 1440

Metro: Last Light

<iframe width="640" height="360" src="http://www.youtube.com/embed/40Rip9szroU" frameborder="0" allowfullscreen></iframe>​

The latest iteration of the Metro franchise once again sets high water marks for graphics fidelity and making use of advanced DX11 features. In this benchmark, we use the Torchling level which represents a scene you’ll be intimately familiar with after playing this game: a murky sewer underground.


2560 x 1440