Last week I had the chance to check out our first R9 390 with the Sapphire Nitro. Well today we get to take a look at what XFX has to offer. This gives us a great chance to compare the cards against each other along with all of the cards previous tested. This helps put more perspective on their individual performance to find out what card is best for your specific build. So today I’m going to run the XFX R9 390 Double Dissipation Core Edition through all of our in-game, compute, and synthetic performance tests as well as the more specific per card tests like noise and cooling performance. In the past the XFX cards have performed well, but it might have its hand full with the triple fan cooler on the Sapphire.
Product Name: XFX R9 390 Double Dissipation Core Edition
Review Sample Provided by: XFX
Written by: Wes
Pictures by: Wes
Amazon Affiliate Link: HERE
Part Number : R9-390P-8DF6
Product Description : AMD RADEON R9 390 1015M 8gb Dual Dissipation
UPC Number : 778656067660
|Processor & Bus||
Bus Type : PCI-E 3.0
Chipset version : Grenada Pro
GPU Bus (bit) : 512
GPU Clock : 1015MHz
Performance Category : Enthusiast
Stream Processors : 2560
Memory Bus : 512 bit
Memory Clock : 6.0 GHz
Memory Size : 8GB
Memory Type : DDR5
AMD Eyefinity Technology : Y
AMD HD3D Technology : Y
AMD PowerPlay Technology : Y
AMD Stream Technology : Y
Highlight Features - 1 : Double Dissipation
Highlight Features - 2 : Unlocked Voltage
|Environmental||RoHS : Y|
Display Port ready : 1.2
Dual link Support : Y
HDMI Ready : 1.4a
Max Supported Resolution (ANALOG) : 2048 x 1536
Max Supported Resolution (DIGITAL) : 2560 x 1600(DVI);4096 x 2160(HDMI;DP)
Output - Display Port : 1
Output - DL-DVI-D : 2
Output - HDMI : 1
Card Profile : Dual
Thermal Solution : DD Fansink
Thermal Type : Dual Slot
Card Dimension (cm) : 29.5 x 14.3 x 4.2
Card Dimension (inch) : 11.61 x 5.63 x 1.65
6-pin to 4-pin power cable : 1
8-pin to 6-pin power cable : 1
Driver Disk Installation Guide : 1
Installation DVD : 1
Promotional Bundles : PSU Cross Marketing Insert
Quick Installation Guide : 1
External Power - 6-pins : 1
External Power - 8-pins : 1
Minimum Power Supply Requirement : 750 watt
XFX Recommended Power Supply : XFX 850W PSU
The outside of the box for the XFX 390 had a completely different look than past XFX cards. The background has a black to grey fade that uses triangle shapes. The front of the box has a HUGE R9 logo along with a smaller 390 and AMD logos. Over on the right are a few key features like the completely unlocked voltage, extended heatpipes, ghost 3.0, and the 8GB ram capacity. The background design carries over onto the rear as well. Here we have three sections split up. One just lists everything inside of the box an also has a list of the outputs the card has. The next expands on the feature list and it is highlighted in a bright red. The last section just has a few AMD specific features listed out. Sadly there isn’t a photo of the card you are actually buying on the front or the back or a specification listing. If nothing else the size of the card would be nice to see to help make sure it will fit If you are buying it in a retail store. That said I do like the background design tha XFX went with.
Inside you pull out a second box. Inside the card is wrapped up in a static protective plastic bag and it sits in a cutout in cardboard. Up under it you will find the driver disc as well as a quick install guide and an install guide for the driver CD (I’ve never seen one of those before). For accessories you get two all black adapter cables. One is a two Molex into a single 6-pin adapter and the second converts two six pin cables into one eight pin.
Card Layout and Photos
So as it turns out the angular design on the packaging for the XFX R9 390 DD was a hint at wat came inside of the box. Moving o the 300 series they redesigned their fan shroud to have that same shape. I will be honest when I say I feel like they are regressing a little at a time. With the 7000 series XFX had a crazy but very unique design that helped them stand out and oozed quality. For the 200 Series launch they dropped the metal fan shroud and went with a great looking blacked out design with a glowing XFX logo. This new design, while still decent looking, doesn’t compare to their previous designs in my opinion. We do still get the dual fan design but this design loses the business formal look of the older cards and isn’t as bold as the 7000 Series. The one thing that we do have going on is an all-black design that should match any build you want to put it in. The card isn’t as long as the Sapphire 390 but it is still really tall, almost an inch over the top of the PCI slot. Oddly enough they didn’t seem to use that space to get even larger fans in even though there is room.
The fans on the 390 DD Core have nine curved blades and a very glossy finish. In the middle they put silver covers with the XFX logo on them. This silver is a little weird with the rest of the card being all black, I would love for the stickers to be black with silver XFX logos to help keep things consistent.
From the end of the card we can get a better look at the cooling design under the fans and fan shroud. We can see five copper heatpipes running through a heatsink that is about ¾ inch thick. This pulls the heat out over the whole heatsink to help both fans cool things down. The heatsink fits very tight up against the caps at the end of the card as well. With the open end and bottom design and with the fans blowing down against the PCB we know that nearly all of the heat from the card will vent into the case, not out of the back of the card like most reference designs work.
Up on the top edge we have that same open design, in fact we can see most of the heatsink and the heatpipes where they turn to run over the GPU. The top edge of the fan shroud does have the XFX logo embossed into the textured shroud. We also can see that there is no need for crossfire bridges as well. Down at the end we have two power plugs. One is an 8-pin and the other is a 6-pin, XFX didn’t go the direction everyone else has been with flipped power plugs. So the clips on the power plugs are on the heatsink side and they didn’t really leave any room to get at them. When testing the card I had a lot of trouble getting my fingers down in there to unplug the cables each time.
I love that XFX didn’t follow everyone else as far as connections go. They kept the dual DVI ports with a full HDMI and full DisplayPort. Most cards have moved to just one DVI port and a lot more DisplayPorts. While my personal setup has a few DisplayPort monitors, most people haven’t moved that direction from my experience. The rear slot also has the XFX logo cut into the vent as well like XFX cards from the past few generations.
The rear of the card has a full length backplate. The backplate has round holes for ventilation over most of the back but they did manage to fit a bright white XFX logo on the back in the proper orientation for it to look correct when installed in your PC. The backplate is also careful to have a cutout that still lets us see the serial number sticker on the PCB as well. My only complaint here are the bring white circles around the four mounting points, I’m not sure why they needed to highlight them so much, especially with it cutting into the mostly blacked out design to do it.
Our Test Rig and Procedures
|Our Test Rig|
|CPU||Intel i7-3960X||Live Pricing|
|Memory||Corsair Vengeance 1600 MHz DDR3 RAM Quad Channel (4x4GB)||Live Pricing|
|Motherboard||Asus Rampage IV X79 Motherboard||Live Pricing|
|Cooling||Intel Active Thermal Solution RTS2011LC||Live Pricing|
|Power Supply||Cooler Master Gold Series 1200 Watt PSU||Live Pricing|
|Case||High Speed PC Test Bench||Live Pricing|
|Our Testing Procedures|
|3DMark||The same goes for the most current version of 3DMark using the Fire Strike benchmark in normal, extreme, and ultra settings|
|Unigine Heaven Benchmark 4.0||Using the “Extreme” preset|
|Unigine Valley Benchmark 1.0||Using the Extreme HD preset to get an average FPS|
|Bioshock Infinite||Using the Adrenaline Action Benchmark Tool we run Bioshock Infinite on the “Xtreme” quality setting. This has a resolution of 1920x1080, FXAA turned on, Ultra Texture detail, 16x Aniso Texture Filtering, Ultra Dynamic Shadows, Normal Postprocessing, Light Shafts on, Ambient Occlusion set to ultra, and the Level of Detail set to Ultra as well. We also run this same test at 2560x1440 using the same settings as mentioned above.|
|Tomb Raider||Using the Adrenaline Action Benchmark Tool we run Tomb Raider on the “Xtreme” quality setting. This has a resolution of 1920x1080, Exclusive Fullscreen turned on, Anti-Aliasing set to 2xSSAA, Texture Quality set to Ultra, Texture Aniso set to 16x Aniso, Hair Quality set to TressFX, Shadow set to Normal, Shadow Resolution on High, Ultra SSAO, Ultra Depth of Field, High Reflection quality, Ultra LOD scale, Post Processing On, High Precision RT turned on, and Tessellation is also turned on. We also run this same test at 2560x1440 using the same settings as mentioned above.|
|Hitman: Absolution||Using the Adrenaline Action Benchmark Tool we run Hitman: Absolution on the “Xtreme” quality setting other than the MSAA setting is turned down from 8x to 2x. That setting puts the resolution at 1920x1080, MSAA is set to 2x, Texture Quality is set to High, Texture Aniso is set to 16x, Shadows are on Ultra, SSA is set to high, Global Illumination is turned on, Reflections are set to High, FXAA is on, Level of Detail is set to Ultra, Depth of Field is high, Tessellation is turned on, and Bloom is set to normal. We also run this same test at 2560x1440 using the same settings as mentioned above, except on the “high” setting.|
|Sleeping Dogs||Using the Adrenaline Action Benchmark Tool we run Sleeping Dogs on the “Xtreme” quality setting. That means our resolution is set to 1920x1080, Anti-Aliasing is set to Extreme, Texture Quality is set to High-Res, Shadow Quality is High, Shadow Filter is set to high, SSAO is set to High, Motion Blur Level is set to High, and World Density is set to Extreme. We also run this same test at 2560x1440 using the same settings as mentioned above.|
|F1 2014||We use the built in benchmark for F1 2014. We use the Ultra setting and then test at 2560x1440 and 1920x1080|
|Total War: ROME II||Ultra setting tested at 1920x1080 and 2560x1440, built in forest benchmark|
|Middle-earth: Shadow of Mordor||Using the built in benchmark we test with ultra settings at 1440p|
|Sniper Elite 3||Ultra setting tested at 1920x1080 and 2560x1440, built in benchmark|
|GRID Autosport||Ultra setting tested at 1920x1080 and 2560x1440, built in benchmark|
|Theif||Tested using the “Very High” setting at 1920x1080 and 2560x1440|
|Folding @ Home||Using the Folding @ Home benchmark we test both single and double precision using the explicit result|
|Unreal Heaven Benchmark 4.0 heat testing||We run through Unreal Heaven using the “Extreme” preset for 30 minutes to test in game cooling performance.|
|Power Usage||Using Unreal Heaven Benchmark 4.0, we get our “load” power usage number from the peak power usage during our test. We get our numbers from a Kill-A-Watt connected to the test benches power cord.|
|Noise Testing||Our Noise testing is done using a decibel meter 3 inches away from the video card on the bottom/fan side of the card. We test an idle noise level and then to get an idea of how loud the card will get if it warms all the way up we also turn the fan speed up to 50% and 100% and test both speeds as well. The 100% test isn’t a representation of typical in game noise levels, but it will show you how loud a card can be if you run it at its highest setting or if it gets very hot.|
To start off my testing I went with the tried and true synthetic benchmarks. We use these to look at the difference in performance between the cards as they have results they don’t change as much between driver revisions and they are consistent. The results don’t show you how the XFX 390 will perform in game, I will get into that testing next. The first benchmark I went with was 3DMark Fire Strike in all three settings to see how the cards handle various resolutions and graphical settings. The XFX came in with a very similar setup to the Sapphire 390 that I tested last week. The XFX has an extra 5MHz on the GPU clock speed and they have the same memory clock. That small difference did seem to show in 3DMark with the XFX edging out the slightest lead over the Sapphire in all three tests. This put the XFX 390 in right below the GTX 980 and above the GTX 970 just like the Sapphire as well.
Next I ran through both Unigine benchmarks as we slowly move over to the Valley Benchmark in the future. In Heaven the XFX pulled a .1 FPS edge over the Sapphire and .4 in the Valley Benchmark. In both tests the GTX 980 was out ahead along with the GTX 970 being behind, no big surprises here, other than 5MHz getting you almost a half an FPS in Valley Benchmark lol.
Like I mentioned in the last section, the synthetic benchmarks are about comparing card to card and the in game benchmarks are all about finding out what YOU will experience with the XFX R9 390. To do that I ran the card through nine different modern games with a mix to try to hit what everyone plays. Each game is tested at its highest settings and with the exception of one game all of the games are tested at both 1080p and 1400p. So how did the XFX perform? Well I like to break down the numbers into three categories, one is anything under 30 FPS because I consider that to be unplayable. Next is over 30 but under 60 FPS, this is typically playable but a picky gamer would notice. Then we have 60 FPS or above, aka the sweet spot. At 1080p the XFX 390 blasted through all eight games tested with each having at-least 60 FPS and most had a lot more than that. When I turned up the resolution to 1440p things did get a little harder. Here we had 5/9 games in the 30-60 FPS range where picky gamers will need to turn down the settings slightly but still playable. The other 4 were up over 60 FPS as well, even with the higher resolution.
So how does all of this compare to the Sapphire 390? Well it had the same results going off the breakdown. In about half of the benchmarks the Sapphire actually pulled slightly ahead at 1080p but even in those cases the extra 5MHz on the XFX helped pull ahead ever so slightly at 1440p. When compared to the Nvidia cards the XFX 390 outperformed its direct competitor the GTX 970 consistently and came in below the more expensive GTX 980 consistently as well.
Next I ran through our compute benchmarks including a newly added Cinebench R15 benchmark. To start things off I tested the XFX 390 using the Folding at Home benchmark to get an idea of how well it would handle both single and double precision work. With the single precision setting the XFX 390 came in just above the Sapphire 390 but in the middle of the charts compared to cards like the Titan X and the GTX 980/980 Ti. With double precision turned on though we really see it step up, well actually we see Nvidia step down. Over the past few years Nvidia has been gimping the double precision of their consumer cards to define the difference between the consumer cards and their workstation product line. This let the R9 390 pull up above much more expensive cards. My last benchmark doesn’t have a lot to compare to just yet but we are nor using Cinebench R15’s OpenGL benchmark to test how well cards can render. The extra 5MHz on the XFX made for a little over three additional FPS over the Sapphire.
Cooling, Noise, and Power
While we did see a slight difference between the different R9 390’s due to the slightly different clock speed. The best way to compare cards with the same GPU is in this section. Here we can see how well the cooler design cools, how quiet it runs, and just how much power the cards pull. To start things off I went into our power testing by running Heaven Benchmark 4.0 and noting the highest peak wattage as well as an idle wattage. While the two 390’s were less than 10 watts apart, it was enough to lower the XFX down below a few other cards leaving the Sapphire 390 as the highest wattage single card tested still.
Next I did noise testing on the XFX at 100% fan speed, 50% speed, and at idle. Here the Sapphire ran quieter even with all three of its fans turned all the way up. The XFX 390 also didn’t turn the fans off during low power usage like nearly every other new card so while it was on the lower end of our idle testing it is still noisier than the Sapphire R9 390.
Lastly I warmed the XFX up by running Heaven Benchmark and noted where the temperatures leveled off. Just like the Sapphire 390 it leveled off at 71 degrees. I consider this to be a little on the warm side for a card with an aftermarket cooler but still better than what you would normally see on a reference card.
Once I finished out regular testing I did want to see how well the XFX would overclock, especially compared to the Sapphire. To do that I broke the overclock down into three different sections, the GPU clock, the memory overclock, and lastly the two put together. The whole time I documented everything and included it below along with the FPS results from our 3DMark 11 benchmark that I use to verify that the overclock works. With the results we can see what works for actual performance and what doesn’t.
To start things off I did a baseline benchmark at the stock clock speeds. From there I started in on the GPU clock speed jumping right to 1200MHz and failing, from there 1100MHz pass so I had to work in. After trying a few I found that 1130MHz was my peak GPU overclock without running into artifacts. For the memory I went through the same thing only it took a lot more tests to get up to a point where it failed, that was 7400MHz where I had a hard black screen crash that required a manual reboot. In the end I was able to get stable results at 7300MHz on the memory.
Next I attempted to run the two overclocks together but I ended up having artifacts. Normally that would make me think the problem was the memory but this time around I had artifacts when I pushed the GPU clocks too hard. So I turned the GPU clock down to 1120MHz and ran again without any issues. This was 10MHz less than the Sapphire managed and more importantly for some reason the Sapphire managed to put out a full FPS higher in almost all tests when overclocking. That said the XFX still managed to jump from 81.54 up to 89.90FPS with its overclock, a nice improvement.
|GPU Clock Speed Overclocking|
|GPU Clock Speed||Pass/Fail||FPS Result||Notes|
|Memory Clock Offset Overclocking|
|Memory Clock Speed||Pass/Fail||FPS Result||Notes|
|7400MHz||Fail||N/A||Black screen, hard crash|
|GPU and Memory Overclocks Together|
|GPU Clock Speed||Memory Clock Speed||FPS Result||Notes|
Overall and Final Verdict
So going into today we already know that the R9 390 out performs the GTX 970 that is is priced to compete with and that the leftover R9 290X’s are also a good value at that same price point. What I really wanted to know was just how good the XFX R9 390 Double Dissipation is in comparison to the Sapphire 390. While XFX cards have performed great for us in the past, when it came down to their coolers they always have a hard time hanging on compared to their competition. This time around though the cooling performance was on par. For noise I would have preferred XFX to introduce a mode that turns the fans off at low usage like other companies but in my testing I didn’t see them turn off. The noise numbers were slightly higher than the Sapphire card that was running an additional fan even. The XFX made up for that in most of the benchmarks were its very small 5MHz GPU clock speed advantage helped it pull slightly ahead. For power draw I still marked it down for pulling a lot of power because when compared to Nvidia’s offerings it does, but the XFX did pull less than the Sapphire 390.
I was a little surprised by the new cooler design though. The past few XFX card designs have really grown on me so it was disappointing to see them move away from that to something that I’m not that big of a fan of. The angular and glossy finish is a little weird. On the plus side the blacked out design does look good. You also get a backplate as well were a lot of the other manufactures have been going away from that. I also think they need to do a little more testing ahead of launch because this cooler design was really tight around the power plugs and that shouldn’t happen.
So is the XFX the way to go? Well the slight performance advantage is nice but the fan shroud design isn’t. Where the XFX really has an advantage is as of this review it is selling for less. Because of that I think that it is the better buy and with its smaller size it will be less likely to have fitment issues as well.
Live Pricing: HERE