AMD R7 260X Graphics Card Review

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

We’re continuing on with our coverage of the GPU14 products released this week by AMD. Today, we have a budget minded offering coming in the form of the R7 260X. Just like the R9 280X and R9 270X we’ve already had a look at, this one is a re-badged GPU based off the already existing Bonaire GPU core. So far, AMD has done a nice job revamping existing GPUs for their GPU14 lineup; let’s find out if that trend continues as we explore the R7 260X.


Specifications/Product Highlights

The AMD R7 260X specifications are listed in the slide below, which was pulled from the press deck we received.


GPU-Z verifies what we see above and adds some additional details as well. We can see confirmation of the Bonaire GCN GPU built on a 160 mm² die using the 28 nm fabrication process. GPixels/s and GTexel/s come in at 17.6 and 61.6 respectively.When compared to the Bonaire GPU core used on the HD 7790, we have a 100 MHz increase in GPU speed and a 125 MHz increase in memory speed. Sounds promising.


To give you an idea of the value AMD is trying to portray, they show a comparison of today’s pricing on the R7 260X versus what the HD 5870 was priced at when it was released. Seems like a lot of GPU for $139, stay tuned!


The R7 260X has been assigned to compete directly against the NVIDIA GTX 650 Ti. AMD’s in-house testing shows it beating the GTX 650 Ti hands down in a multitude of different games. We’ll see if this holds true in our testing later in the review.


The popular Eyefinity technology is available on the R7 260X, which can support up to six monitors using a DisplayPort Hub. That being said, six monitors might be too much to ask from a card in this class, but I would venture to say a dual monitor setup would be fine.


TrueAudio is a promising new 3D audio technology being developed in conjunction with several game development, audio middleware, and processing algorithm partners. It’s no secret audio performance on a PC has been stagnate for some time now. With only 10% of CPU utilization being dedicated to audio, game developers have to develop to the lowest common denominator. Because of this, there isn’t much time spent on enhancing audio effects, as not to impact overall game performance. Enter TrueAudio, which can do it all CPU-free. TrueAudio is a DSP and not a CODEC, meaning it replaces what you can get through audio, but not where it comes from (standard audio out ports, USB, HDMI, etc,) Below are a few slides further detailing TrueAudio. You should be able to purchase several game titles supporting TrueAudio in the coming months.

amd-r7-260x-37 amd-r7-260x-30
amd-r7-260x-32 amd-r7-260x-33
amd-r7-260x-34 amd-r7-260x-31

Meet the R7 260X

Here are several glamor shots of the R7 260X, cute little guy, ain’t it? You can see it uses a conventional downdraft style cooler, so proper case air flow will be at a premium.

This slideshow requires JavaScript.

I measured the overall length of the card and came up with 6-7/8 inches. To give you a visual sense of the card’s size, the next slideshow is a series of pictures showing it next to the R9 270X.

This slideshow requires JavaScript.

The R7 260X Up Close

Removing the cooler is a simple task that only requires the removal of four screws at the back of the card. Once off, we can see an additional heatsink covering the MOSFETs that should get a good amount of airflow from the fan. Once AMD partners roll out their own versions of the R7 260X, you’re likely to see coolers much different than this one.

Cooler Removed
Cooler Removed

Cooler & MOSFET heatsink Removed

Cooling System
Cooling System

Display connectivity options include two DVI, HDMI, and DisplayPort.


Here are some images of the R7 260X undressed. We’ll dive a bit deeper momentarily.

Bare R7 260X
Bare R7 260X

Bare R7 260X
Bare R7 260X

Bare R7 260X
Bare R7 260X

Bare R7 260X
Bare R7 260X

Moving in for a closer look reveals a 4+1 power phase design and a single 6-pin PCI-E power connector. We can also see that Hynix H5GQ4H24MFR-R2C has been used for the 2 GB onboard GDDR5 memory. The memory on the R7 260X is set to 1625 MHz (6500 MHz Quad Pumped), but the Hynix specifications say it’s capable of 1750 MHz (7000 MHz Quad Pumped) at 1.5 V. Hopefully, that equates to a fair amount of memory overclocking, but we’ll see. And finally, we have a good view of the Bonaire GPU core.

R7 260X PCB Layout
R7 260X PCB Layout

Power Delivery
Power Delivery

Power Delivery
Power Delivery

Power Delivery
Power Delivery

Bonaire GPU Core
Bonaire GPU Core

Hynix GDDR5 vMem
Hynix GDDR5 vMem

Overclocking and Benchmarks

Test Setup

Motherboard ASUS Maximus VI Formula
CPU Intel i7 4770K Haswell
Memory G.SKill TridentX DD3-2666 MHz 2x4GB
SSD Kingston HyperX 3KSSD 240 GB
Power Supply Corsair HX1050 Professional Series
Video Card AMD R7 260X Graphics Card
Cooling Swiftech Apogee HD CPU Water Block – 3X120 mm Radiator – MCP35X Pump

Overclocking for Stability

Just as we found in our earlier R9 280X and R9 270X reviews, overclocking software hasn’t quite made jump to these new GPUs as of yet. They allow core speed and memory speed to be adjusted, but we don’t yet have any voltage control. However, that won’t stop us from seeing what the card can do without voltage manipulation.

After quite a bit of tinkering, I finally settled on a GPU core speed of 1250 MHz and a memory speed of 1825 MHz (7300 MHz Quad Pumped). That’s a +150 MHz increase in core speed (12%) and a +200 MHz increase in memory speed (11%). Not too shabby considering no voltage increases were available. Because of the lack of voltage control, and that I’ve basically reached the maximum this card can do without it, we’ll forgo the “Pushing the Limits” section of the review. We totally expect this to change once the AMD partner cards arrive in our hands for review… sooner than later!

HWBot Heaven @ 1250 / 1825

3DMark Fire Strike @ 1250 / 1825

Benchmarking Methods

Our new “GPU Testing Procedure” has been in place for several months now, but If you are not yet familiar with it, click the provided link to learn more. Below is a quick synopsis of the procedures we adhere to.


  • i7 4770K @ 4 GHz
  • Dual Channel DDR3-1866 9-9-9-24
  • GPU @ stock
  • Monitor capable of 1920×1080


  • All Synthetic benchmarks set to their default settings
  • Unigine Heaven (HWbot) is run using the “extreme” setting
  • Aliens vs. Predator – 1920×1080 with highest settings offered (4x AA, textures set to highest)
  • Battlefield 3 – 1920×1080 at Ultra settings (4xAA/HBAO by default)
  • Dirt 3 – 1920×1080 with 8x MSAA and all settings enabled and at Ultra where possible
  • Metro 2033 – 1920×1080, DX11, Very High, 4x MSAA/ 16x AF, PhysX OFF, DOF enabled, Scene: Frontline
  • Civilization V – 1920×1080, 8x MSAA, VSync OFF, High Detail Strategic View: Enabled, Other Settings: High, using full render frames value ( / 60)
  • Batman: Arkham City – 1920×1080, VSync off, 8xMSAA, MVSS and HBAO, Tessellation set to high, Extreme Detail Level, PhysX Off

This will more than likely be the last reference design card we review in the AMD R9 and R7 series, so we’re going to chart all three that we have reviewed to date. We wanted to give our readers a quick way to compare the performance between the three cards we received at launch – the R9 280X, R9 270X, and today’s R7 260X. Because the R7 260X is pointed directly at the NVIDIA GTX 650 Ti, we’ll also include results from that. However, please note the GTX 650 Ti we have for comparison is actually EVGA’s GTX 650 Ti Boost SC edition. That card has a faster core speed, faster memory bus, and faster memory speed than the reference design GTX 650 Ti that AMD undoubtedly used for its in-house testing. So, we’re likely to see disparities in our test results versus what AMD shows. But, as long as the R7 260X hangs tight to the GTX 650 Ti Boost, AMD’s claims would definitely be substantiated.

Ok, with the disclaimer out of the way, let’s dive into our synthetic benchmark results. Here we see the R7 260X indeed holds tight to the GTX 650 Ti Boost and actually beat it in the 3DMark Vantage test. Once the R7 260X was overclocked, it was able to overtake the GTX 650 Ti Boost in our 3DMark Fire Strike test. At stock settings, we actually achieved a better 3DMark Fire Strike score than AMD noted in their press deck. Nothing to complain about in the synthetic tests, and the R7 260X seems to live up to its advanced billing.

3DMark Vantage Results
3DMark Vantage Results

3DMark 11 Results
3DMark 11 Results

3DMark Fire Strike Results
3DMark Fire Strike Results

HWBot Heaven Results
HWBot Heaven Results

Moving on to our game benchmarks, we see a similar pattern with the R7 260X holding close to the GTX 650 Ti Boost. When overclocked, it even managed to top the GTX 650 Ti Boost a couple of times. The R7 260X showed nice scaling as well, which is always a good thing! Again, nothing to complain about here, and performance levels were just as expected.

Aliens Vs. Predator Results
Aliens Vs. Predator Results

Batman: Arkham City Results
Batman: Arkham City Results

Battlefield 3 Results
Battlefield 3 Results

Civilization V Results
Civilization V Results

Dirt 3 Results
Dirt 3 Results

Metro 2033 Results
Metro 2033 Results

As a side note, I spot checked the AMD in-house testing results that I could match up to our testing. Of the two direct comparisons I found (3DMark Fire Strike and Battlefield 3), our test results actually came out a little better than what AMD showed. It’s nice to see accurate marketing claims from AMD, kudos for that!

Temperatures and Power Consumption

Our temperature testing procedure entails running HWBot Heaven at both stock and overclocked settings. The results are normalized to 25 °C ambient. Testing includes the fan control set to auto, and then again with the fan speed set to 100%. Because the cooler’s heatsink is made entirely of aluminum, things got pretty warm when the fan control was set to auto. The fan never ramped up much past 50%, which contributed to the high temperatures. Those high temps don’t concern me much as AMD partner cards will undoubtedly have much better coolers on them.

With the fan set to 100%, we see a marked improvement in temperatures across the board. They are still a little higher than we like to see with the fan set to 100%, but adequate. As far as fan noise goes, lower speeds result in very little noise. When ramped up to 100%, the fan makes itself known; but not what I would call obnoxiously so.

Temperature Testing
Temperature Testing

Our power consumption testing is done with a Kill-a-Watt, and wattage usage recorded at idle and load. We run both HWBot Heaven and 3DMark 11 (Combined Physics Test) to hopefully get the maximum power draw the video card can produce. I take this testing one step further and also provide results while the video card is overclocked.

Power consumption for the R7 260X is insanely low. With this card installed in our test bed, the whole system at idle drew a mere 98 watts. Heck, I’ve got light bulbs in my house that draw more than that… Crazy! The most power draw I could get was when the R7 260X was overclocked and benchmarking HWBot Heaven. Even at that, the highest power draw recorded was 231 watts. Combine power consumption like this with a little better cooler from AMD partner cards, and you’ve got something that would be great in a HTPC application too.

Power Consumption
Power Consumption


As I’m sure most of you are aware by now, NVIDIA lowered some of their pricing to better compete with these new AMD R9 and R7 offerings. It’s hard to tell how this “price war” is going to shake out over the next few weeks, but what I see right now won’t sway a Fan Boy of either side from changing allegiance. AMD says the R7 260X is priced at $139; but when partners slap their own coolers on, expect a $5 to $10 increase in that price.

If TrueAudio takes off like AMD hopes it will and the upcoming Mantle API is widely adopted by game developers, the value of AMD graphics cards will be even more appealing to potential buyers. But, just like the pricing tug-of-war, we’ll have to wait and see how this all plays out over the coming months.

As our testing indicates, the R7 260X performed quite admirably for a card in this price range. Once game titles that take advantage of TrueAudio are released, this card will have something the more expensive R9 270X/280X cards do not… support for it. That will add even more value to the asking price. An inexpensive gaming video card with awesome audio potential sounds like a pretty sweet deal to me.

Overclockers_clear_approved Click the stamp for an explanation of what this means.

Dino DeCesari (Lvcoyote)


Leave a Reply

Your email address will not be published.


  1. ima amd fanboy but seems to me 650ti is a better deal, specially if you find one used i paid 108 for mine(used evga lifetime warranty), and its the 2gb ssc midget version.
    would be nice if it wasnt a double slot card either. for itx applications most of the smaller cases only have one pci cutout since its itx spec. mine is like that but since i dont have the stacked dvi ports they are side by side it still fit fine and with the cooler setup just like that one it doesnt have to exhaust out the back.
    Hi Guys,
    The only information I have right now lists three upcoming game titles.
    Murdered -Soul Suspect
    That doesn't mean that's all of them, but for now it's all I have on the list.
    Looks like its competing with a 650 TI or 650 TI boost with very comparable specs. Nothing special at all and i would say, its a bit newer and improved but ultimately its not able to exceed the Nvidia offers, its simply a good match. Its probably a bit stronger than a 650 TI but slightly weaker than a 650 TI boost. But totaly depends on situation of course, guess to some extend it can match the 650 TI boost, but its not truly helping at gaming because ultimately no matter if we use 650 TI, 650 TI boost or 260X, they all need to be "cranked" down to medium settings, in term they want to be used for gaming (with good FPS).
    Surely a great HTPC GPU for those who want a truly powerful and beasty "master of all classes" HTPC GPU. But at 1080P gaming it wont go past "medium settings" else the card will struggle to maintain good FPS in many cases.
    In the same price range a even better gamer GPU could be gotten, for example a old 6950 or 7850, they both may perform better and are priced the same or even lower (6950 could be gotten for less than 100 $ and 7850 may drop to below 130 $ very soon). So for me the 260X is simply a great, quiet and very effective HTPC GPU for those who arnt happy with any of the weaker solutions. Especially at high AA settings, the higher bus bandwidth of the 6950 and 7850 may be able to improve overall stability of FPS (max FPS is never a matter, min FPS is the critical stuff).
    The entire 7000 series and all the R series are basically GNC based architectures* (the 2 new "true next gen consoles" too) and in theory they all should be benefiting from Mantle (because they basically use the same drivers). But im not sure about the whole Mantle thing, so its gonna be a interesting research. After all, AMD did release very few infos about that new API so far but i think it can be useful for a alternate OS such as Linux. The main problem Linux got is the API support, most games are only supporting Direct X and i truly wonder if Mantle is able to make some changes.
    *There is just very small differences, resulting into GCN 1.0, 1.1 and 2.0 "specs" but it doesnt mean to much. The GCN 1.0 is the "old" 7000 series, the 1.1 is simply with a new audio chip as far as i know, and the only "truly improved GCN" is the GCN 2.0, but its main performance is coming from a increased ressource count and just very few from a minor GCN improvement.
    I heard, Mantle can even be used on Nvidia or Intel Hardware, so its not bound to AMD. Mantle is able to provide full features, no matter the OS (so users of Win 7 dont get disdavantage vs. Win 8), so its pretty interesting. BF4 will be supporting Mantle, but not much more known so far.
    Still curious why Overclockers isnt able to review "the real beast" but the baby puppys instead. Its pretty unusual and a hint that AMD isnt truly trying to compete hard enough, just trying to match Nvidia in some way... its some sort of flower war i feel. Prehaps there is a hidden agreement between those companys not to use the blade of a Katana but its sheath instead, so both are able to decrease cost and it may be helping in order to fight the real monster... Intel.
    Lol ivy the nda isnt up on them yet thats why they have posted the reviews. Its always been that way the flagship gpus are always released later they still have two more that are faster than the 7980
    Not true, the 7970, the old flagship was the first 7000 series they released and it took them way to much time in order to release the "lower series". Same for Nvidia, the first Kepler they was releasing is the GTX 680... Funnily enough, the time AMD was releasing the 7970 was just a "boasting press release" and not a official one. They wasnt ready for a quick delivery of the customers. With the R series, they apparently are gonna handle it a bit different and arnt trying to make "press releases but without a real release": So i expect to have a good bunch of cards available soon, right after "press release".
    But with the R series, AMD decided to make a flower war, and isnt trying to "impress" to hard. Will be interesting how Nvidia will be handling it when Maxwell is released. Prehaps they gonna throw in the flagship in order to "impress".
    Btw: 7980 sounds like a myth, and it wouldnt even be easy to cool that chip down (without watercooling) when there is even more SPs. In term there is no improvements on the architecture i wouldnt want that "jet engine". The main problem nowadays is the cooling, not any other factors such as PSU, even the smallest PSUs (smaller than a card) could handle a Titan nowadays. People should stop looking at power supply elements (inside card or outside), it says nothing at all... the cooling is the only factor able to enable or disable performance (apart from chip quality, cooling is the culprit to OCing too).
    The 290X +290 got some architecture improvements (not a new architecture but they did improve the GCN a little). I expect them to run at least 10% more effective (GFLOP/W) compared to the "old GCN" but it could be even more (of course thats affecting TDP too). Else the massive amount of ressources used is difficult to implement without the card running hot (it would result into a jet engine, comparable to a 7990). Cooling is the main issue when building stuff inside SFF (small form factor PCs) too and any hardware able to perform well at such a environment got a increased value. Compact sized cooling elements (such as micro WC) will be the most advanced and most important stuff soon, in a hardware market with increased demand for smaller systems because the amount of people seeking for compact solutions is increasing. The most valuable PC is the one with few size but good cooling, not unable to exist but hardest tech able to create (thats the challenge a expert should try to execute).
    lol how can you type so much with saying literally nothing. you could have said all that with two sentences. and 7980 was obviously a typo.
    and either way they never have released an entire series at one time. never. they always stagger them weather it is the flagship gpu's first or the mid - low end ones first they have done it both ways. and i can bet were gonna have the 290 r the 290x for reviews the "big boy cards" that we "never get" because this must be such a bad forum?.
    p.s. the entire last two paragraphs you posted were completely and utterly pointless. there's an entire thread dedicated to rumors for gpu's, actually the entire speculation about neither company trying hard is covered in there too.
    Look, this time the lower R series is nothing more than "old chips" and those was already available for very long. They was probably only changing the model number and putting in a new sound chip, thats it. So its no wonder the lower series was that easy to release.
    The situation would be different when the architecture is completly new (comparable to the old 7000 series). In that term they are gonna create a fully winged chip, and the lower series will usualy be released after because they are usualy fully winged chips with "broken legs", so they have to be run at lower ressources. Some of the even lower series will have a own much smaller die in order to save up cost but its contraprductive to release the weak chips first because they arnt having the required amount in order to supply the huge mainstream at release. Its easyer to produce a few "fully winged" chips at release at a high price, so the supply is easyer to maintain. There is a good reason for the staggering, just have to understand the matter.
    The lower end R series has been released faster than the flagship because there is already tons of chips available... its not a new chip at all. ;)
    Still, AMD could have delayed the release (of the lower series) in order boast around using theyr flagships first, but it doesnt truly change anything because either they gonna sell the old chips in the shape of a 7000 or the new R series... its both almost the same. The true reason why i call it a flower war is because AMD could have released (press release) the flagships way sooner and not delay the reviews so hard... because the card is basically fully alive for several weeks already but AMD isnt trying to make a "premature press only release" anymore. At least many game devs already got the stuff running for weeks, and are certainly happy with, at the same time we are getting releases of rebadged hardware for the public... (whoot). At least they could put the new sound chip on any of them but im not sure thats the case.
    which we all already know they are old chips with the "updates" thats why they can crossifire a 280x and a 7970. duh. its not like AMD and NVIDIA havent been doing the same crap since the 8800/1900 series, both sides have been rebadging cards almost every single series. i dont see why you are making such a big deal over it this time. 770 is just a rebadged 680... 9800gtx rebadged 8800gtx. ect ect.
    Besides, the 260X got the new audio chip is that correct? So i truly would say its a great HTPC card, but as a (max setting) gamer card it will lack muscles (especially for future games). So its actually the GCN 1.1 spec, when i got that right.
    Of course, the entire matter about AMD deciding not to implement the sound chip inside any mid range card is leaving a very bitter smell on my tongue, that cant be serious yet they still are serious. So AMD is kinda offending the whole mid range market doing so and only giving them cheap rebrands, but at least at a good price.
    The rebadging was rarely used on non OEM cards and usualy just a few cards but not entire series with the exception of flagship(s). It wasnt common for the dedicated high end market, but i believe that the customers simply want to see "something new" to waste bucks on, so rebrands/rebadge are welcome in order to make them feel to get something new.
    the audio thing will not effect movie stuff so it wont belp htpc's or anything, its more of a sound physics processor per se. like different sounds for environments and sounds made by physical interaction distances, surroundings all things like that will effect the sound in real time so its not just playing a general mp3 file that is the same no matter what comes in contact with something.
    Well, in term the new chip is only affecting sound physics (that means real time processing using a game engine, prehaps even non game engines), and they are implementing it at a weak gamer card (performance is about 650 TI boost level), its rather hilarious. The aged 650 TI AMP (OCd) is about the same level of performance with same card size and about same GFLOP/W, almost 1 year older than this new flashy release. OC potential of the 260 X is low too (same as AMP) because its already clocked very high.
    Technically the 260X is a 7790 rebrand but with a new "sound physics chip", yet those sound physics calculations could be done on the CPU (prehaps even GPU) itself but its using valuable processing ressources. There is another weakness on "CPU" bound calculations: The CPU simply isnt fast enough in order to make calculations without delay, but its bit difficult to understand. A CPU, such as any part of the PCs hardware, is running at a certain frequency and is exchanging informations with many other parts. Thats a complicated architecture constantly busy with a lot of tasks, which is comparable to a "ground bounce", and its inferior for the very best audio quality having such a delay*. So indeed, i feel that a good dedicated audio chip can be supreme: A dedicated audio solution (a own chip) is able to reduce that delay a huge margin because it wont do any other actions apart from sound processing and isnt "busy" with many other tasks. *Funnily enough, the delay is increasing at higher CPU load or when to much CPU cycles are used (CPU time), so its not even stable and difficult to create a great sound that way. Its one of the reasons why i do believe, a good CPU giving better (more accurate) audio because it got more cycles left in order to deliver the bits more accurate in time, that will even affect static sources (without real time physics). Some people dont take the sound chip matters serious because they dont know the background or are underestimating that matter but i dont, i find its important having such a solution for further advancement of the audio processing capabilitys. Although, i think, when there isnt very good audio equipment the hearable difference is very limited, so the equipment is critical for further improvement
    its not just playing a general mp3 file that is the same no matter...

    MP3 is such a aged, lossy and inaccurate format that it truly is hilarious why it became such a huge success on the mainstream... any true audio equipment would be able to reveal a lot of loss. Although, its perfectly hitting the real issue. Because of those facts i do believe: Most people are having inferior equipment and thats why they cant even hear the difference and are taking the audio matter not serious at all. They may just have in mind: "I want gaming power... i want eye candy and FPS and a good monitor and what else. But good audio, whats that? Lets just take those tiny, almost for free, cheap loudspeakers from Wallmart and then lets have fun... :D"

    Although, i believe its correct that the mid range may get the first 20 nm refresh. Its a very lucrative market and it may create lot of hassle leaving those buyers in the cold.
    Although, i think, when there isnt very good audio equipment the hearable difference is very limited, so the equipment is critical for further improvement

    so your saying without a good set of speakers or headphones you wont be able to hear the difference the sound chip is going to make? its way beyond audio quality. you could tell the difference with a pair of $1 ear buds.
    as for the load on the cpu im sure it will be very minimal, idk what the main cpu of the system would have to do with it besides assigning it minimal resources the gpu will be doing all the calculating.
    and i dont understand why you keep saying their lowest end card that is just as fast as their last series mid cards is a piece of crap just because its a rebrand. so the 770 is a piece of crap too then? its a gtx 680.
    considering this "low end turd" is nearly as fast as a 6970... i dont see the problem. heck my main card is an oc'd 6950 and it doesnt break a sweat playing any current games. and the 650x is a low end card no one should be purchasing it expecting to play with max settings on every new game.
    in short i dont understand why you keep saying this series is all crappy.
    260X is a 7790 and about same performance such as a 650 TI OC (there is OCd out of the box cards) or a 650 TI boost, so those cards arnt "low end" they are challenging the 260X, thats fact. The aged 6950 is at least comparable in performance and the 7850 is a bit ahead, not by that much but they got the edge as far as i can tell. On top of that the 6950 can now be gotten for very few bucks, so price/performance is higher. Regarding AA, the slow memory bus and low ROP count of the 260X is able to cut its wings, and in term of high AA, even the 6950 may outperform that card. So its a rather mediocre gamer card and Nvidia got comparable offers for a full year already.
    I think, regarding audio quality, we have to separate the "physics" from the "general" audio quality. The physics itself isnt general audio quality, but even the physics quality is increased a lot when there is good audio equipment resulting into general quality. Its possible to use headphones and hear a noticeable "increase" regarding sound physics but its still no match compared to real audio equipment.
    So, there is probably not much general audio quality needed in order to hear "good sound physics", as long as that kind of physics isnt much more than the increase or decrease of noise in order to create a audible space. But even more complicated when the exact direction of a sound have to be executed. Gonna be hard with just 2 speakers, but according to AMD they can do magics and create the audible space out of 2 speakers or a pair of headphones, and that makes me officially impressed. Prehaps they was boasting to much, i cant say without excessively testing it. I think a majority is simply underestimating the sound matter in general and got in mind that they can get good sound without serious investment. It may not work that way because audio equipment is just as important than a good monitor, as far as my knowledge goes. As far as my experience goes, especially when it comes to the implementation of the exact location of sonic waves (this is part of the sound physic), even real high end 5.1 and 7.1 systems are challenged doing so and even the room itself is very important, it will need a good acoustic.
    But in term AMD is able to do magics and bring some outstanding sound physics to everyones ears... i will just gladly shake theyr hands and bow down.

    as for the load on the cpu im sure it will be very minimal, idk what the main cpu of the system would have to do with it besides assigning it minimal resources the gpu will be doing all the calculating.

    Most of the sound physics is currently done CPU side not GPU as far as i know. The GPU is usualy busy at rendering and sometimes visual physics too (Havok, PhysX, TressFX). Although the sound physics development has been stuck for some time, a CPU simply got its limitations handling it. So the devs didnt pay a lot of attention to sound physics anymore, thats at least the stuff told by AMD. I cant tell, i never made research regarding sound physics but i find it a interesting and important matter with a lot of possible improvements.
    Its a mid end card its supposed to be mediocre its not a bad thing. Seriously drop that point it is moot it makes no sense the card is cheap you cant expect it to play bf4 on ultra or even bf3. No mid card plays any mainstream titles with max settings
    There are no games with true audio even out yet its still in developement no nothing is computing it. It does not exist yet that is why true audio is being integrated into the cards now so when its finally implemented it will work.
    The soundcard does all the processing for the rest and most people that would use better quality audio have a receiver that can decode the audiostream digitally itself. Movies wont have to generate any "sound physics" they just stream a pre generated audio stream. Then the receiver or sound chip decodes it and decides what channel thw sound goes to.