AMD A10-7890K Glorious FPS for Less

A10-7890K - Introduction Very recently we took a long hard look at the AMD A10-7800 APU and walked away extremely impressed with what that rather inexpensive APU could do; however, we did have a few people question why we were reviewing that older model when AMD had already upgraded the Kaveri processor architecture and released a newer improved generation – the Godavari APU series. The answer to that is twofold. First the A10-7800 represents a great value in its own right and deserved its own moment in the spotlight…. but mainly we did it because we had a 7890K in our hands and wanted to be able to show precisely what has been changed from one generation to the next. Today we will be putting the $150 (USD) version of the Godavari (AKA Kaveri 2.0) A10-7890K APU under the microscope to see what has been upgraded and precisely what the additional $40 nets you the consumer. In order to do this, we will not only be comparing it to the previous A10-7800 and a modern i3, but also putting it up against an Intel i5 – specifically Intel i5-6400 which costs $170 or twenty dollars more than the A10-7890K. We will also be using the same RAM and motherboards as the previous review so as to keep things as fair as possible. We will also be using that amazingly potent AMD Wrath CPU cooler as it is now the stock cooler for this AMD APU. On the surface this is certainly a very interesting four-way competition as the 7890K costs only $150 but offers consumers four Kaveri 2.0 Compute Cores running at a maximum speed of 4.1Ghz (or 400MHz more than the A10-7800) and eight video Compute Cores running at 866Mhz (or 146Mhz more than the eight which is in the A10-7800). That certainly does seem like one heck of deal; however, at this price point consumers can almost get an Intel i5-6400 which offers four Intel 'Skylake' cores running at a maximum speed of 3.1Ghz, and 24 Execution Unit based onboard video processor (HD Graphics 530) running at 950Mhz. So when the dust settles and all the numbers are crunched will the AMD A10-7890K APU be able to do what its direct predecessor was able to do? Or will the increase in processor speed, and refreshed cores simply not be able to justify its noticeable increase in price? Let's find out! A10-7890K - Specifications CLOSER LOOK: A10-7890K As the A10-7890K is an APU design the best place to start is by describing exactly what an APU is. An APU is an Accelerated Processing Unit and basically is a CPU (Central Processing Unit) with an GPU (Graphics Processing Unit) built right on to the same piece of silicon (aka 'on-die GPU'). However, this does necessarily mean a larger number of pins are required. To be precise AMD's FM+ APU's use a 906 pin design (with the pins being on the APU and not the motherboard's socket) instead of 941 pins AMD's AM3 processors require. How is this possible? Simple. Good engineering… and a lot of time, effort and experience with doing a true dual purpose processor. To be precise Godavari is a fifth generation design and is the second generation of AMD APU's that uses a Heterogeneous System Architecture (HSA). As we saw in the A10-7800 review, HSA does make a difference and does allow core components that date back to early 2014 to compete with Intel 2016 designs… and win on the video front andoverall value front – while doing a darn decent job on the CPU front as well. When Godavari was announced many assumed this second generation HSA APU would be a radical change from Kavari. That is not the case. Instead Godavari is heavily based on Kaveri. In fact, beyond some minor low level tweaking to the architecture that increases efficiency, Godavari is a Kaveri… on steroids. Basically AMD focused on making the best Kaveri design possible on a 28nm fab process and then moved on to the upcoming 'Zen' architecture that will be new and potentially radically different. On the surface this does sound disappointing, to say the least, but the fact of the matter is Intel does the exact same thing with their 'tick-tock…tock' cycle. Put another way, besides the missing node shrink AMD are doing exactly what Intel has done for many, many generations. For those who really, really want a brand new architecture October should be an interesting month (as 'Zen' is alleged to be landing). In the meantime, Godavari has a lot of horsepower under the hood and is still one potent all in one processor. So what is a Kaveri 2.0… err Godavari APU made up of? Basically it is the same as a A10-7800 (or if you want to pedantic a A10-7850K) but at higher clock rates. This means two CPU blocks for a total of four Steamroller based Compute Units running at a maximum speed of 4.3Ghz (up from 3.9Ghz in the fourth generation A10-7800). For the 'GPU' portion of the core the A10-7890K has eight video Compute Units… just as with the previous Kaveri. However, while it is indeed true these eight graphic Compute Units are based on an R9 2-series design, they are now running at 866Mhz (up from 720Mhz) and still boast 512 stream processors, 128 Texture Units and all the other sundries that make up a modern AMD 2-series R-7 'GPU'. Also as expected to make all twelve of these Compute Units work together as one cohesive whole AMD's HSA layer resides 'on top' of them. This in turn means the four CPU Compute Units can create and send work directly to the eight video CU's and vice versa (via what AMD calls heterogeneous Queuing, and the industry simply calls Dynamic Parallelism). Better still this cross flow of GPU calling on CPU functions (and vice versa) is all done at the kernel level and does not require OS, API, or even drivers to do so. Instead it is bare-metal crosstalk that is a lot more efficient than a discrete CPU and GPU… or even an older APU's ways of doing things. More importantly is the twelve cores now share the entire RAM amount as one large memory unit instead of dividing it upon POST into two parts… that could not be shared or even 'seen' by the other Compute Unit type. This is called hUMA (heterogeneous Uniform Memory Access) and works darn good and requires a lot less TLC than the old way of dividing and hoping one type does not run out of memory. The works loads are also now equally spread across all the CUs and each CU knows what the others are working on so as to not cause issues with writing to memory sector that another is reading ('Platform Atomics'). Basically the A10-7890K is a super-clocked A10-7800 (well really a A10-7850K) but at a higher TDP (now 95 instead of 65). Also noteworthy is the additional $40 'surcharge' that the '7890K demands over the older '7800 is it is a 'black' processor. In layman's terms this means it is an unlocked processor that allows for rather easy overclocking of both the 'CPU' portion and the 'GPU' portion. In testing we easily pushed the four CPU cores to 4.5Ghz and pushed the GPU cores to 1040Mhz. That is damn impressive considering we did all that and still did not hit the 72-degree thermal wall. Needless to say AMD's Wrath stock cooler is one beastly cooler that allows consumers the luxury of spending money on other things… without really limiting overclocking. Color us impressed. Closer Look: Wraith CPU Cooler Even before we removed this cooler from its cardboard box we knew that AMD had created something special… as this heavy box was downright massive. This impression was then further reinforced once we actually looked at it. So much so we had to double check that yes it does indeed have an AMD sticker on it as we thought that it might have been a new aftermarket CPU cooling solution that somehow got mixed in with the rest of the shipment! To be honest the sheer weight of this cooler is reassuring but it was the size that really grabbed our attention. For a 'stock' cooler it is massive and it really does look more like an aftermarket cooler. Better still is while the fan and fan shroud are now upsized – more on this in a moment – the aluminum fin array has also gotten the jumbo-sized treatment. AMD states that this new cooling tower, and yes, it is now large enough to be a downdraft tower cooling array, offers 24% more surface area but that is only half the equation… or maybe even only a quarter. Yes more surface area means higher efficiency and transferring heat from the array to the air, but that is putting the cart before the horse. The other part of the equation is the additional weight. To be specific this CPU cooling solution tips the scales in at 455grams, whereas the older/traditional AMD stock cooler was 330grams. This is a 37.9% increase and as every enthusiast knows the more metal there is to transfer heat from the IHS to the cooling array the cooler the CPU cores will be… which in turn means a lower fan speed is need to keep the CPU cool. Yes it is a positive feedback loop and AMD knocked it out of the park. By flipping the wraith over we can see that what AMD did was take the original… and simply made the fin array larger. Though this is overly simplistic as the fins are solid aluminum and AMD's engineers simply made them taller and wider…and added more of them. Also on the positive side, the base of the wraith is cooper and not aluminum – something that Intel should take note of – and uses four heat pipes. This is the same number as the original but they are longer and thus can handle even more heat before being overloaded. The other half of the equation is the fan. The best that could be said about the original AMD fan was that at least it was better than Intel's… but that is not saying much. This new fan however is bloody impressive all on its own. To be precise this is a Delta 92mm fan that is rater for 60CFM at 3200rpm; however equally impressive is it is rated for an impressive 5.34mm of static pressure! What this means is even at lower rpms it still will have plenty of power to push the air through the fin array – something that the older AMD fan also had troubled with. To further improve the fans abilities to push air through the fin array AMD has included a nice large fan shroud that removes the dead zone in front of the fan hub and evens out the fan pressure. As an added bonus, thanks to rubber mounts used, there basically no fan vibration noise created. Now the one downside to using a larger CPU cooler is that it can cause installation issues. Thankfully this is not an issue with the AMD wraith… or at least not as long as you are not sticking it on a mITX motherboard and trying to shoehorn it into a shoebox sized case. Other than this this cooler is every bit as easy to use as the original (it even uses the same stock mounting setup), and does not block RAM or GPU slots. Color us impressed. Testing Methodology For this review all CPUs were tested at stock 'factory' settings with no overclocking, boosting, or tweaking done to their profile. To ensure consistent results, a fresh installation of Windows 8.1 was installed with latest chipset drivers and accessory hardware drivers (audio, network, GPU) from the manufactures website. The full list of the applications that we utilized in our benchmarking suite: Cinebench R15 3DMark 8 WinRAR x264 HD Benchmark AIDA64 PCMark 8 Sleeping Dogs BioShock Infinite Battliefield 4 Assassins Creed: Black Flag Call of Duty: Ghosts System Components Used: Processors: A10-7800, Intel i3-6100, A10-7890K, Intel i5-6400 CPU Cooler: AMD Wrath CPU cooler, stock Intel cooler, stock AMD cooler Memory: 32GB GSkill TridentX DDR3-2400, 32GB Crucial Ballistix Sport DDR4-2400 Motherboard: Gigabyte GA-F2A88XM-D3H, ASUS Z170-A Power Supply: EVGA Storage: Crucial MX300 750GB L.E. A10-7890K - Software Benchmarks CINEBENCH R15 CINEBENCH is a real-world cross platform test suite that evaluates your computer's performance capabilities. CINEBENCH is based on MAXON's award-winning animation software CINEMA 4D, which is used extensively by studios and production houses worldwide for 3D content creation. winRAR Benchmark WinRAR is a popular file compression and decompression tool that also happens to be multi-threaded application. While it does have a built in benchmark we have opted to use a custom 3GB test consisting of small, medium, and moderately large files. We then time how long it takes to compress the files into one .rar file with all WinRAR settings left at their defaults. x264 HD Benchmark x264 HD Benchmark is a benchmark that allows you to measure how fast your PC can encode a 1080p video clip into a high quality x264 video file. It allows for an easy comparison because everyone running it will use the same video clip and software. The x264 video encoder has a fairly accurate internal benchmark (in frames per second) for each pass of the video encode and it also uses multi-core processors very efficiently. All these factors make the x264 HD Benchmark an ideal tool in comparing the video encoding performance of different processors and systems. Adobe Photoshop is a notoriously slow loading program under the best of circumstances, and while the latest version is actually pretty decent, when you add in a bunch of extra brushes and the such you get a really great torture test which can bring even the best of the best to their knees. Let’s see how our review unit faired in the Adobe crucible! PCMark 8 PCMark 8 is the latest iteration of Futuremark’s system benchmark franchise. It generates an overall score based upon system performance with all components being stressed in one way or another. The result is posted as a generalized score. Overall we are rather impressed with what this APU can do. Yes, it is based on an older architecture but the end result is all that matters: good performance. At this the AMD 10-7890K does offer performance that is more than reasonable for its asking price. Basically this APU costs more than an i3 but less than an i5 and performs in-between these two 'extremes'. 3DMark Ice Storm Benchmark The latest version of 3DMark from FutureMark includes numerous benchmarks including one they have "Ice Storm" and we have used the 'Extreme' settings. A10-7890K - Gaming Benchmarks Sleeping Dogs Sleeping Dogs is an open world action-adventure video game developed by United Front Games in conjunction with Square Enix London Studios and published by Square Enix, released on August 2012. Sleeping Dogs has a benchmark component to it that mimics game play and an average of four runs was taken. The settings used in the testing below are the stock Medium display settings and a resolution of 1920x1080. BioShock Infinite BioShock Infinite is a first-person shooter video game developed by Irrational Games, and published by 2K Games released in March 2013. The game has a benchmark component to it that mimics game play and an average of four runs was taken. The settings used in the testing below are Low for quality and a resolution of 1920x1080. Assassins Creed: Black Flag Assassins Creed: BF is an historical action-adventure video game, published by published by Ubisoft and released in October 2013. Like Battlefield 4 this games does not include an in-game benchmark. This makes it perfect for more real world gaming testing. To obtain repeatable results we have used FRAPs and recorded a custom run through Havana’s marketplace. An average of four runs was taken. The settings used in the testing below are Medium settings for quality, VSync disabled and a resolution of 1920x1080. Call of Duty: Ghosts Call of Duty: Ghosts is a first person shooter video game, published by Activision and released in October 2013. Since it does not include an in-game benchmark, it perfect for more real world gaming testing. To obtain repeatable results we have used FRAPs and recorded 90 seconds of the single player ‘Ghost Stories' level, starting as soon as we resume control of Logan back on Earth. An average of four runs was taken. The settings used in the testing below are medium stock settings for quality, VSync disabled and a resolution of 1920x1080. As you can see while the 'CPU' portion of the AMD A10-7890K is basically in-between an Intel i3 and i5 in its overall abilities, the A10-7890K's 'GPU' is a relative monster. A monster so large that it is more than up to the task of moderate game play. This is something that Intel is still not really capable of. A10-7890K - Wattage, Heat and Sound System Power Consumption To obtain accurate results we have connected the system to a Power Angle power meter that has in turn been attached to a 1500watt UPS. This ensures only 120volt power is supplied to the PSU and removes an variances that could potential crop up because of brownouts and power spikes. In order to stress the video card we have once again used Unigine’s Valley benchmark and ran it for 20 minutes to determine peak system power consumption. For idle results we have let the system idle at the Windows 8.1 desktop for 25 minutes and recorded the peak idle power consumption. This is one area that Intel is clearly ahead of AMD in; however, in return for more power consumption consumers get access to an integrated GPU that is actually worthy of the name 'Graphics Processing Unit'. Temperature Results For all temperature testing the cards were used in an open test bed environment. Ambient temperature was kept at a constant 20°C (+/- 0.5°C) and if the ambient room temperatures rose above 21°C or dropped below 19°C at any time, all benchmarking was stopped until proper temperatures could normalized. For Idle tests, we let the system idle at the Windows 8.1 desktop for 25 minutes and recorded the peak temperature. For Load tests we ran AIDA64 for 20 minutes. AMD really knocked it out of the park with this new Wraith design and it does pay dividends. So much so that it is easily up to the task of pushing the A10-7890K to rather lofty overclocking heights. Sound Level Test Results While everyone "hears" noise differently there is one easy way to remove all subjectivness and easily compare different fans: use a sound level meter. This way you can easily compare the various fans noise envelopes without us coloring the results and see what fans fit within your personal comfort level. Of course, we will endeavor to try and explain the various results - which are taken at a 15 inch distance from the GPU’s fan(s) - to help you gain an even better understanding of how loud a cooler's stock fan is, but even if you discount our personal opinions, the fact remains numbers don't lie. For Idle tests, we let the system idle at the Windows 8.1 desktop for 25 minutes and recorded the peak temperature. For Load tests we ran AIDA64 for 20 minutes. Even with a higher TDP processor the AMD Wrath cooler is more than up to the task. Once again this cooler simply destroys what Intel includes with their processors as it not only keeps the APU cooler but also quieter. A10-7890K - Summary and Closing Statement Summary Once again AMD does lag behind Intel's equivalent in the CPU department, but the GPU built into the A10-7890L simply dominates Intel's iGPU. Basically for most consumers the AMD will provide a decent if not stellar real world processing power, but is an excellent choice for those who actually want to use the integrated 'GPU'… and let's face it if you are not intersected in the onboard graphics why even bother with an APU?! Closing Statement Let's start with the obvious. The new Godavari design's improvements over Kaveri are best described as refinements that focus mainly on performance per watt…. rather than overall computing performance. It has the same 'CPU' cores, the same 'GPU' cores, and even the same number of each (4 and 8 respectively). It is 'only' a refinement on an already good design. These refinements however allow the processor to run cooler at a given speed and allowed AMD to boost the speed (and thus the performance) of both the CPU and GPU Compute Cores. What these refinements don’t do however are allow this new A10-7890K to compete with Intel's top of the line models. This to us is perfectly fine… and expected. While yes the AMD A10-7890K will not have Intel and their latest i7's, let alone the 2011-v3 6950X, quaking in its boots very few average consumers will care about that. For the same price as an i7 CPU customers can buy the A10-7890K, a motherboard… and maybe even have some money left over to put towards the RAM. Better still is this processor is not a 'cheap' processor instead it just inexpensive and can offer people a heck of a lot of bang for the buck. That is the key to this processors success: its overall value. Simply put, when it comes to value AMD has once again knocked it out of the ballpark… and past the parking lot behind the bloody ballpark. As we saw throughout testing this $150 processor easily slots in between a modern i3 and i5 options from Intel on the general processing front… but eats both of those processors graphics abilities for lunch. In some cases, you basically need to take both the i3 and i5's results to beat the A10-7890K! Better still is even when compared to the less expensive A10-7800 this newer processor easily justifies its 36-percent increase in price. On the surface that may seem counter intuitive as the results from the modest clock boost and moderate improvements in the Compute Cores are not 36% higher. But there is a big ass terisks to that statement… as that 36 percent may sound downright massive but its only forty dollars. For a mere forty dollars extra consumers get a nice boost increase in performance, and get easy overclocking. Yes, that 'K' at the end means it's an unlocked 'black' processor and with very little effort consumers can easily get a further boost in performance. This boost is actually enough for it to take on the i5 and make it a real horserace to say the least. Then when you add in the Wrath cooler and its abilities vs an i5 and things are even better for AMD and their extremely frugally priced powerhouse APU. No matter how you look at it, this ease of overclocking combined with improved cores makes the A10-7890K an excellent value. After all, near i5 performance for darn near i3 asking price is all on its own amazing. However when you add in the fact that DDR3 RAM is a lot cheaper than DDR4… well that value really is knocked so far out of the park it nearly reaches orbit. For that reason, we have zero qualms with awarding this Godavari APU our Real Deal award. Needless to say when looking at a reasonably priced system we strongly recommend not overlooking what AMD and their 'APU' options have to offer… no matter what some 'Internet pundits' have to say. Otherwise you will once again be overlooking one hell of a value and a darn decent processor in its own right.

No Result