Some of you may have noticed we didn’t have coverage of the AMD Radeon team's launch of the RX 6800 and RX 6800 XT last week. Sadly we haven’t gotten in the reference cards, at least not yet. But I have had the chance to spend some time with one of the aftermarket RX 6800 XT’s, the XFX Speedster MERC 319 AMD Radeon RX 6800 XT BLACK to be exactly. The full name is a mouthful, but I was excited to see how XFX would follow up the 5000 series. They have changed things up completely with a new series of cards (Speedster) as well as Merc which stands for mercury, not mercenary which designates this is their high-end model. Today we get to get our first official look at the 6800 XT as well as put XFXs new cooler to the test. But only after I take a closer look at what all it has going on. So let’s dive in!

Product Name: XFX MERC 319 RX 6800 XT

Review Sample Provided by: XFX

Written by: Wes Compton

Amazon Affiliate Link: HERE


Model Number


Product Name

AMD Radeon™ RX 6800 XT

Bus Type

PCI-E 4.0

Stream Processors


Memory Bus

256 bit

Memory Clock

16 Gbps

Memory Size

16 GB

Memory Type


Card Profile

2.9 Slot

Thermal Solution

3 Fan

Clock Speeds


Base Clk: up to 1925MHz

Game Clk: up to 2110MHz

boost Clk: up to 2360MHz


Factory Default Setting

Base Clk: up to 1900MHz

Game Clk: up to 2090MHz

Boost Clk: up to 2340MHz

Power Phases

14+2 phase design (AMD was 10+2)

Display Connections

2x DisplayPort

1x USB-C


Card Dimension (cm) 34 x 13.9 x 5.7

Card Dimension (inch) 13.39 x 5.47 x 2.24


3 Years

Before diving into everything I do always take a look with GPUz to double-check that the listed specifications match up with what I am getting in my testing. Our Merc 319 has a boost clock speed of 2340 MHz which lines up with the balanced BIOS mode. This is also a nice overclock over a stock 6800 XT as well which has a clock speed of 2250 MHz. Sadly the BIOS revision wasn’t getting picked up by GPUz so we can’t document that. But I was testing on the latest public driver which is 20.11.2.

image 1


Along with changing up their naming completely once again, XFX also changed their packaging up as well. It does still have the same black background that they have used before. But like EVGA they flipped things to a vertical orientation. Down at the bottom, the red Radeon wrap around with the model name is still there. Next to that are matching red boxes that highlight the VRAM (which is 16GB) and that this is a PCIe 4.0 card. But above that, the black background takes over with a slight overlapping grid design on it. They have “Speedster Series” or well Speedster Seri in the background with red outlined letters. Then over that, they have MERC 319. Up top, they have all of that information in a small section with the XFX logo. So here is how the branding works out. The Speedster Series is their main series that covers all of the cards. The MERC stands for mercury which is XFX’s top in designation, the idea is to be able to have other categories later that could cover other areas like budget-focused for example. Then the 319 is the cooler designation. This cooler has three fans with the 1 and 9 being for the 100mm and 90mm fans used and the 1 is first because there are two 100mm fans vs the one 90mm. I would have loved to see a picture of the card itself on the front, but they did at least include pictures of the card on the back. Two that cover most of the angles of the card. They list off what you will find in the box but beyond that, the rest of the information is your generic Radeon info. I would love to see the card dimensions and the display connection listed here if nothing else so anyone shopping in retail will be able to know if it will work for them. Power connections/wattage requirements would be nice as well.

image 5

image 6

Inside the box, the MERC 319 comes wrapped up in a thick plastic bag. On top of that XFX also ships it with coatings, you have to remove that cover the entire fan shroud and the backplate so it is well protected from scratches. As for bigger damage, it comes sitting in foam cut to fit it with another foam panel that sits on top. There aren’t any accessories or adapters included. So with the card, you get a paper warning to download the driver online and a book with warranty and installation instructions.

image 7

image 8

Card Layout and Photos

I don’t think any other graphics card manufacture has reinvented their card's looks and styling as much as XFX has over the years that I have worked with them. There has often been some carry over, but they have had designs with an all carbon fiber looking fan shroud, full metal shrouds with silver and orangish-red trim, that same design only with black, cards that feature triangles, a completely flat design, and many many more. Their RX 5000 series had a new look as well along with changes mid-cycle to improve on performance. This time around for the 6800 XT the new Merc 319 has again received a new look and I can only describe the shroud design as a flat golf ball. The new styling is a large flat black finish with dimples all over the shroud, anywhere they couldn’t put a fan. It is also significantly larger than past cards coming in at 13.39 inches long or 340mm long. It is a three-slot thickness at 57mm thick or 2.24 inches and it is 139mm or 5.47 inches tall which puts it around an inch and a half taller than the top of the PCI bracket. Like our first 3080 aftermarket card, this one is just huge to handle and if the two are any indication it looks like cases are going to need to keep making more and more room for video cards. In fact, just any one of the three dimensions is enough to prevent these cards from fitting in some SFF cases.

image 9

image 10

So like with the Thicc III last generation, the Merc 319 has a triple fan configuration. Only this time around we get two larger fans and one smaller fan, last time it was one large fan and two smaller fans. The two large fans are 100mm and the center fan is 90mm and these are axial, pushing air down into the fan, not to the sides like a blower card. The fans also look different fan the last generation of XFX fans with more blades but similar to some of the older designs like the RX580. The new center badge is a little weird because it isn’t round, making it look like it's missing something but it does put the XFX X on each fan. They missed the chance to make the center fan have an F though.

image 11

image 17

Even without the Thicc branding, the Merc 319 does still have a similarly thick heatsink which is split up into three sections and connected together with heatpipes. XFX has a 2mm thick 99mm x 85mm in size copper cold plate with 7 heatpipes in total to help pull that heat out across the cooler. The heatsink is aluminum and isn’t a universal design, you can see when looking around the edges of the card that it is shaped around each tall cap and thicker at the end of the card beyond where the PCB ends. The heatsink is laid out horizontally and at the end of the card XFX doesn’t have the end completely blocked off, but it is partially blocked with a weird gap at the top that makes the shroud look like it is floating.

image 18

image 19

image 20

The top edge of the card has a few things going on. For power, it has two 8-pin power connections which are a few inches away from the end of the card, right at the end of the PCB. They are flipped around backwards with the PCB notches to allow room for the clips. The PCB is also notched down here to keep the power connections low to allow room for the cables. This means that XFX is using the house money to gain extra space on some of the top of the card, using the space that the cables already stick up which is nice. Next to the power plugs, there is a small switch that switches between the normal Balanced BIOS and the Rage mode BIOS. This only changes the boost clock slightly, going up from 2340MHz to 2360MHz, but also changing the power limits and fan profiles as well. The top edge of the Merc 319 also features a raised section at the top which has the Radeon RX 6800 XT model name which is also backlit. Then on the shroud itself, they have the XFX logo as well. I have to say, not enough cards put the actual model on the cooler at the top edge. Frankly, I could care less about seeing Radeon or GeForce like most cards get or just the card brand. With higher-end models, I think it is much nicer to show off the GPU model itself and cards used to do this but have gone away from it in the name of reusing parts. XFX has managed to do this by making the model name section its own part, so the rest of the shroud can be reused if they want.

image 15

image 16

image 22

XFX went all out on the branding on the backplate of the card. It has the XFX logo cut into the thick metal backplate along with a whole array of holes of various sizes for ventilation. This is in addition to making sure to use thermal pads between the backplate and card to help pull heat out. In fact, the entire card is loaded with thermal pads, 20 in total making sure all of the VRM, memory, and other components are all integrated into the cooling. The back has pads right on the back of the GPU and on the opposite side of the memory and VRM as well. The solid backplate also helps keep things ridged, which will be important with a card this big. They also have MERC painted in bright white across the back. But I think it’s the blow through section at the end that I like the most. Like what Nvidia did with their Founders Edition cards, the section at the end past the PCB has the thicker heatsink and a blow through design the keeps airflow unrestricted.

image 12

image 13

image 14

I’ve said it with most cards recently, but it applies here as well. This card desperately could use a flat black PCI bracket. Not only would it help match the rest of the styling of the Merc 319 cooler. But also matches most cases as well or in the case of white, it contrasts nicely. There is some ventilation here even though the card design shouldn’t be pushing much this way and it has the XFX logo cut in because I don’t think there is an angle other than the end of the card that doesn’t say XFX somewhere. As for display connections, XFX went with something different than the normal layout. You end up with two DisplayPort connections and one HDMI. Then the small one at the end is a Type-C, much like the Type-C that Nvidia was including last generation for VirtualLink though I don’t know if that is supported here. It’s a shame that everyone else dropped that because I still think it would be nice. If nothing else you do get an extra type-C connection.

image 21

Before diving into testing, I did want to check out the lighting and the Merc 319 is limited in its lighting to the backlit stuff up on the top edge. The Radeon RX 6800XT in the main section looks amazing. Especially mixing the red and blue. The XFX down at the end doesn’t look bad as well even though I don’t normally like backlit branding aka billboards much. I would have loved to see the Radeon dropped and just have the RX 6800XT with MERC up on top backlit. There aren’t any other lighting accents or any controllable RGB at all in fact, nor any way to turn it off from what I have seen as well for those who don’t want any lighting at all.

image 23

image 24

image 25

Test Rig and Procedures

Test Rig

CPU: AMD Ryzen 3900X

Motherboard: Asus Crosshair VIII HERO WiFi

Memory: G.Skill Trident Z Royal 3600MHz 16-16-16-36

Storage: Corsair MP600 2TB

Cooling - Noctua NH-U12S

Power Supply - Corsair AX1200w

Case - Primochill Wetbench

OS - Windows 10 Pro 64-bit

Our Testing Procedures


The same goes for the most current version of 3DMark using the Fire Strike benchmark in normal, extreme, and ultra settings. Tests are also run in the DX12 focused Time Spy benchmark as well as the Time Spy Extreme test. Port Royal is also used on video cards that support DirectX Raytracing

Unigine Superposition

1080p Medium, 1080p Extreme benchmarks along with the VR Maximum and VR Future tests, both done at the Vive resolution


Cyan and Blue rooms tested, use Average FPS for the result

Borderlands 3

Built-in benchmark testing with the ultra detail setting and medium detail setting, done at full screen with default settings at 1080p, 1440p, and 4k.

Metro Exodus

Using built-in benchmark, testing at ultra and normal details at 1080p, 1440p, and 4k.

The Division 2

Built-in benchmark at Ultra detail with V-Sync turned off at 1080p, 1440p, and 4k resolutions.

Total War: Three Kingdoms

Built-in benchmark using the Battle Benchmark setting. Tested at 1080p, 1440p, and 4k at both Medium and ultra detail settings

World War Z

Tested at 1080p, 1440p, and 4k in both Medium and Ultra Detail using the built-in benchmark.

Shadow of the Tomb Raider

Built-in benchmark, tested using the Medium texture setting and again at the highest texture detail setting. Both tested at 1080p, 1440p, and 4k

Far Cry 5

Built-in benchmark, tested at 1080p, 1440p, and 4k with the Ultra and Medium detail settings

Warhammer 40,000 Dawn of War III

Built-in benchmark, Image and Texture settings set to the maximum setting and V-Sync turned off. Tested at 1080p, 1440p, and 4k

Watch Dogs: Legion

Built-in benchmark testing at ultra and high details. Tested at 1080p, 1440p, and 4k

Shadow of the Tomb Raider

Built-in benchmark, tested using the Medium texture setting and again at the highest texture detail setting. Both tested at 1080p, 1440p, and 4k

Far Cry 5

Built-in benchmark, tested at 1080p, 1440p, and 4k with the Ultra and Medium detail settings

Passmark Performance Test 9

Test using the GPU Compute Score inside of Passmark’s Performance Test 9


Using the new Blender Benchmark with the Quick Benchmark setting set to use the GPU, not the CPU. The result is in total seconds the test took, lower is better. All cards tests were done using the 2.90 build for compatibility with the latest cards

Basemark GPU

GPU tests were done using the OpenGL and DirectX12 APIs

Geekbench 5

GPU Compute test is run using the OpenGL and the Vulkan tests

Power Usage

Results come from a Kill-A-Watt hooked up in line to the power cord for the test rig. Two tests are done, one using the AIDA64 Stress Test and the second uses the 3DMark Fire Strike benchmark on the Performance setting using the combined test.

Noise Testing

Our Noise testing is done using a decibel meter 18 inches away from the video card on the bottom/fan side of the card. We test at both 50% and 100% fan speeds. 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. Under load testing is also done, measuring the noise levels of the card when under load in AIDA64 over a half hour. This is done using a Protmex PT02 Sound Meter that is rated IEC651 type 2 and ANSI S1.4 type 2. Tests are done set weighted to A and set to a slow response using the max function. The ambient noise level in the testing area is 33.3 decibels using the test settings.

Temperature Testing

Using AIDA64, the GPU stress test is run for 30 minutes or until the result has leveled off. The test is run twice, once with the stock fan profile and a second time with 100% fan speed.

Synthetic Benchmarks

As always I like to start my testing with a few synthetic benchmarks. 3DMark especially is one of my favorites because it is very optimized in both Nvidia and AMD drivers. It's nice to not have to worry about it being favored too much either way and the repeatability of the results makes it a nice chance to compare from card to card, especially with this being our first RX 6800 XT and the Merc 319 having a higher clock speed of 2340 MHz compared to the stock 2250 MHz.

The first round of tests were done in the older Fire Strike benchmark which is a DX11 test. There are three detail levels, performance, extreme, and ultra and the 6800 XT just jumps completely off the charts. Nvidia’s last few generations haven’t liked the DX11 titles as much, but it is amazing how well Big Navi did here.




The next two were both based on the Time Spy benchmark. One is the standard test and then there is the extreme detail level. In these, the gap between the 6800XT wasn’t nearly as large. It did still come in above the overclocked 3080 in the basic Time Spy test, but with the extreme, the numbers were flipped with both 3080’s coming in ahead. The Merc 318 was still in this lead pack (without the 3090) which is out ahead of the 3070 as well as the older 2080 Ti. If you would have asked me last year, my guess would be AMD matching or just barely beating the 2080 Ti, so this alone is huge. Not counting the 6900 XT in the future as well.



Because my whole section of RTX and DLSS testing doesn’t apply to AMD I have the ray tracing test from 3DMark included here as well. Port Royal uses Microsoft DirectX Raytracing which AMD has support for as well. The 6800 XT didn’t catch up to the RTX 3080 here but did outperform the 2080 Ti with its score of 9400.


The last test was using the Unigine based Superposition benchmark and I tested at 1080p with medium detail and again at 1080p with the extreme detail setting. In the extreme detail setting which this graph is sorted the two 3080’s did come in slightly ahead. But the medium detail setting also impressed here being significantly higher than all of the 3000 series cards.


VR Benchmarks

As for Virtual Reality, I love it but it is more demanding than traditional gaming. This is partially because of the resolutions needed to render for two eyes and because they render more than what is immediately visible. But also because of post effects to get the proper “fisheye” effect for it to look proper in your eyes with the HMD. You also have to have much higher expectations for frame rates in VR, skipping frames or lower FPS can cause motion sickness in VR. Because of that, I ran a few tests.

My first test was again in Superposition. This time I tested the VR Maximum and VR Future tests using the Vive resolution. This was similar to my other Superposition testing with the Merc 319 coming in just behind the two RTX 3080’s in the more demanding VR future testing and doing well in the VR maximum test.


In VRMark the Merc 319 topped the chart with 122.53 FPS in the always demanding blue room test. This is well over the 1090 target FPS as well which only a few cards reach. Then it stands out in the cyan room test with 35 FPS over the Suprim X 3080.


In-Game Benchmarks

Now we finally get into the in game performance and that is the main reason people pick up a new video card. To test things out I ran through our new benchmark suite that tests 9 games at three different resolutions (1080p, 1440p, and 4k). Most of the games tested have been run at the highest detail setting and a mid-range detail setting to get a look at how turning things up hurts performance and to give an idea of if turning detail down from max will be beneficial for frame rates. In total, each video card is tested 48 times and that makes for a huge mess of results when you put them all together. To help with that I like to start with these overall playability graphs that take all of the results and give an easier to read result. I have one for each of the three resolutions and each is broken up into four FPS ranges. Under 30 FPS is considered unplayable, over 30 is playable but not ideal, over 60 is the sweet spot, and then over 120 FPS is for high refresh rate monitors.

So how did the XFX MERC 319 RX 6800 XT perform? Well in 1080p it did well with 11 of the results coming in over 120 FPS and the other 5 were all over 60 FPS and most were on the high end. With the resolution up to 1440p half were over 120 FPS and the other half were in that same 60-119 range. Then finally at 4k, we have the first two results that were below 60 with 13 over 60 and still 1 over 120 FPS. For comparison, the aftermarket RTX 3080 I just tested had the same result at 4k but did better at 1440p with two more and one more at 1080p.




Looking at the actual results we can start to see why the Merc 319 fell behind a little in the previous graphs. There were a few results like Watch Dogs Legion on the high detail setting which are clearly CPU limited. In that one with 116 FPS at 1440p, I have no doubt the 1080p result would be well over 120 FPS with an even faster CPU than the 3900X on our testbench. The 1440p result on Dawn of War 3 was tragic as well with the 6800 XT well ahead at 4K but struggling in both 1440p and 1080p, but the 1440p result came in .09 below 120 FPS. I ran all of these multiple times with the same results. Overall though the 6800 XT does come in slightly behind the RTX 3080, especially at ultra detail. But it does much better at the bumped down detail settings like in World War Z at medium detail, Metro Exodus at normal detail, and Shadow of the Tomb Raider at medium detail where it went from being behind both 3080’s at the ultra detail to being ahead with the detail stepped down slightly. That said at 4k it still came out on top 6 times with two more results where it beat out the Founders Edition but not the overclocked RTX 3080 Suprim X.

















Compute Benchmarks

Now some people don’t need a video card for gaming, they need the processing power for rendering or 2D/3D production, or in some cases people who game also do work on the side. So it is also important to check out the compute performance on all of the video cards that come in. That includes doing a few different tests. My first test was a simple GPU Compute benchmark using Passmark’s Performance Test 9 and the XFX MERC 319 RX 6800 XT topped the chart with 12580 points over the 12402 of the RTX 3080 Founders Edition.


In Basemark it didn’t have the same result. The OpenGL performance was, like the 5000 series cards, very bad. But the 6800 XT did well on the DirectX 12 test, but not well enough to pass the RTX 3080’s.


Blender is always my favorite compute benchmark because the open-source 3D rendering software is very popular and it isn’t a synthetic benchmark. Here I render two scenes and combine the total time it takes. The 5700 XT struggled when compared to Nvidia on this but here the 6800 XT does extremely well with it taking 111 seconds for both together. This is still a little slower than both RTX 3080’s but an impressive jump in performance from the RX 5700 XT even. You nearly halve your time even upgrading from the last gen card.


Cooling Noise and Power

For my last few tests, rather than focusing on in-game performance, I like to check out other aspects of performance. These are also the most important ways to differentiate the performance between cards that have the same GPU. To start things off I took a look at power usage. For this, I use our Kill-A-Watt hooked up to the test bench to record the total wattage of the system. I ran two tests with the first using 3DMark Fire Strike to put the system under a load similar to normal in game performance. Here the MERC 319 RX 6800 XT did surprisingly well with the whole system pulling 373 watts, which was WAY under what I saw with the 5700 XT and the comparable RTX 3080 which were up in the 470/480 watts range. In the second test using AIDA64 as a workload, however, the power draw jumped all the way up to 452 watts which was still better than the RTX 3080.



My next round of tests were looking at noise levels. These are especially important to me because I can’t stand to listen to my PC whirling. Especially when I’m not in game and other applications are using the GPU. XFX does turn the fans off on low power usage, which helps keep things quiet when you aren’t gaming. For my testing, though I first tested with the fan cranked up to 100% to get an idea of how loud it can get, then again at 50% to get an idea of its range. The 100% fan speed tests with the Merc 319 and its three fans weren’t fun, being nearly as noisy as some of the blower cards with the fans cranked up that high. It was still a little below the THICC III at least. Thankfully the card at 50% fan speed wasn’t bad at 40.4 decibels and I should note that 50% is using AMDs software which I would argue (by going off RPMs) is a little higher than 50%. The under load testing which is the only real important result here was a little more reasonable at 38.4, which was also quieter than the 3080 Founders Edition. I have also included the fan RPMs which reach 3409 at 100%, this one is interesting because normally the placement of the RPMs in the chart matches the 100% noise chart well but the Merc comes in much noisier than its RPM range would indicate.




To finish up my testing I of course had to check out the cooling performance. To do this I ran two different tests. I used AIDA64’s Stress Test run for a half-hour each to warm things up. Then I documented what temperature the GPU leveled out at with the stock fan profile and then again with the fans cranked up to 100%. With the stock profile, the MERC 319 RX 6800 XT did extremely well with 65c. This was also right where XFX had their Thicc III targeted as well and it isn’t bad when we also consider the stock fan profile wasn’t loud as well. I honestly thought cranking the speeds up would continue that trend but it only dropped 10 degrees off which is on the low side but also close to what the 3080 Suprim X did. The cooler performs well, but there isn’t a lot of overhead still left as it sits.




While running the stock fan profile testing I also took the time to get a few thermal images so we could see what is going on. On the fan side, there is a noticeable difference in cooling performance down at the end where the open back area is. Even the hottest area behind the center fan isn’t bad. The area under the card doesn’t look as bad as some cards get as well, which is weird given that is the edge with the most ventilation. The top edge shows that the raised branding bar is working with the air venting just below/behind that. The backplate is also pulling heat out from the back of the PCB with all of those thermal pads. You can see the hot spot right behind the GPU, more airflow over this area in a case would be ideal.

image 2

image 3

image 4

Overall and Final Verdict

The Radeon RX 6800/6800 XT launch has been exciting and a great follow up to Nvidia’s launches as well. That is of course not including the supply issues affecting nearly all of the launches packed into the last quarter of 2020. It’s clear that XFX hasn’t been sitting on their hands between the 5000 series and the 6000 series. The old cooler design is gone as well as the entire naming scheme they used to have. Thicc is no longer and we now have the speedster series with their top-end mercury card. Specifically, the RX 6800 XT Merc 319 which I took a look at today. It’s all new look which has what I like to think of as speed dimples aren’t as clean and simple as some of XFX’s past designs. But it isn’t far off, sticking with black and not going overboard with RGB lighting at all. It also dropped the chrome from the Thicc lineup for silver accents which are a little more muted.

XFX learned from mistakes from their early 5700’s and have improved the cooling design here with extra ventilation, a larger cooler, larger cool plate, and thermal pads on everything. The result is solid cooling performance, though I will say cranking the fans up doesn’t result in much more. They upgraded the VRMs over the stock AMD offering as well. My biggest complaints about the card are that it is huge and that when you do crank the fans up it is a lot noisier than it should be. The noise shouldn’t be an issue as long as you stay close to the stock fan profile which isn’t bad at all. As for the size, well you just need to make sure this beast is going to fit into your case. I know a lot of SFF cases aren’t going to be able to fit this and a lot of the other aftermarket cards from this generation on both sides. I also noticed that the MERC 319 doesn’t have a way to turn off the top lighting for anyone who might want that. Unless I missed it.

Now for pricing XFX has this their flagship card with an MSRP of $799 which is a lot more than the MSRP of the stock RX 6800 XT which was $649. The $799 is in line with some of the other RX 6800 XT aftermarket cards and cheaper than a few as well. More importantly though, until RX 6800 XT card availability improves the pricing isn’t going to matter. Especially when stock is selling on eBay for twice its MSRP. SO for now, the XFX MERC 319 RX 6800 XT looks to be a solid card and a good option if you come across availability. As far as its pricing, we will have to reevaluate that later when there is enough availability for it to matter. It is $100 less than the RTX 3080 Suprim X that I recently reviewed which puts it about right with its performance but more than a lot of the other RTX 3080’s on the market. Which would make the decision hard given that the 3080 does come out ahead in more of the games tested. 


Live Pricing: HERE

Author Bio
Author: garfi3ldWebsite:
You might call him obsessed or just a hardcore geek. Wes's obsession with gaming hardware and gadgets isn't anything new, he could be found taking things apart even as a child. When not poking around in PC's he can be found playing League of Legends, Awesomenauts, or Civilization 5 or watching a wide variety of TV shows and Movies. A car guy at heart, the same things that draw him into tweaking cars apply when building good looking fast computers. If you are interested in writing for Wes here at LanOC you can reach out to him directly using our contact form.

Log in to comment

We have 1262 guests and one member online