AMD A10-7850K (Kaveri)

[section_title title=AMD A10-7850K Kaveri] Introduction For years, AMD is pushing the idea of treating CPU and GPU equally when comes to computing power through the concept called "heterogenous computing". After all, why would you waste all of the power and space in a PC if certain workload can be better suited for the GPU as oppose to CPU. Even better would be to have both CPU and GPU working concurrently in parallel to finish a task faster. AMD started talking about such concept with the Fusion platform in 2006 which leads to the company’s acquisition of ATI. While the concept of heterogenous computing sounded good, getting there is not an easy task. To achieve such goal, it requires a fundamental redesigning of the processor. A lot of work and engineering is needed not just in the hardware but also software. AMD started moving toward true heterogenous computing slowly with the Brazo platform that put the CPU and GPU on the same chip. Then with the Llano, AMD integrated the Northbridge to the same chip and created the Radeon Memory Bus. Trinity/Richland, AMD fine-tuned the architecture and brings us Turbo Core, and allows the GPU to address the system memory. Essentially, all of the APUs that was released were more of a CPU with a GPU on the same chip than a true co-processors. Finally, with the release of Kaveri, AMD has gotten all of the interconnect between the CPU and the GPU worked out so that it can finally treat the CPU and GPU as equal partner and delivers the true "heterogenous computing" that it has set to achieve. With Kaveri, AMD has further refined the Bulldozer microarchitecture that is now in its third generation. Codenamed Steamroller, the CPU on the Kaveri is able to deliver a better performance at lower power consumption. The GPU on Kaveri does get a major upgrade from the Cayman derived VLIW4 to GCN microarchitecture, the same that is found on the current generation of the AMD GPU. However, what is the most exciting about the Kaveri in our opinion is not the processors upgrade but rather the underlying interconnects between the CPU and the GPU that AMD has implemented with the heterogenous system architecture (HSA). HSA, hQ, hUMA, With HSA, AMD has finally found the last puzzle that is needed in order to achieve a true heterogenous computing where both the CPU and GPU are treated co-processors with coherent access to virtual memory and system level atomics for synchronizing workloads across the cores. HSA eliminates the need for data copying back and forth between the CPU and the GPU instead, both cores are able to access the same data in the same place on the memory. Kaveri brings us three new technologies hUMA, hQ, and HSAIL. We have talked about hUMA in the past. The quick summary about hUMA is that it finally allows the system memory to be seen by both CPU and the GPU. Both cores are able to access the same memory so that no longer system needs to copy data back and forth between the two cores. hUMA implements pointer system where an entire data structure with all of the embedded links can be passed to GPU. Having both CPU and GPU share the same memory space not only removed the unnecessary overhead from copying data between the two cores, it also allows both cores to on the same data at the same time to even concurrently to finish work faster. Besides hUMA, AMD also brings hQ to the HSA to further improves on the efficiency between the different processor cores. hQ brings user-mode queuing to reduce overhead of job dispatching between CPU and GPU. In addition, GPU can enqueue work for itself and even able to queue CPU call back function. Now with hUMA and hQ, AMD essentially allows both CPU and GPU to behave like co-processors in the APU as oppose to have the GPU depended on the CPU. As the HSA is not only limited to just CPU and GPU, it is possible that other processors can be added to the mix where it can also share the same memory space so that work can be distributed to whoever is best suited for a given task. AMD has added ARM Cortex A5 to some of the APU for hardware security and it also has plan to build their own ARM chip. So it is possible in the future we will see ARM chip that is also HSA compatible where it will able to work in conjunction with CPU and the GPU as well as other special processor cores. AMD has laid down the foundation and have the hardware ready with HSA and now it just need to get the developers onboard. Traditionally software development has always lagged behind the hardware development. And to get the software developers to create applications or optimized their current applications for HSA, AMD and other members of the HSA foundation has standardized HSAIL, an intermediate language that bridges the gap between the high-level language and the machine code. The idea here is to allow developers to continue to use their current high-level language like OpenCL or C++ and the LLVM compiler will generate HSAIL binary object format which the HSA Finalizer ingest then translate to the native machine code. HSAIL is not an API instead it is an intermediate language that creates runtime and map to the hardware of the vendor’s ISA. As HSA is an open standard not AMD’s own proprietary, this intermediate translation language allows the same software to work across all different hardware GPU, CPU, ARM, DSP, etc. HSA foundation includes not just AMD but alsoother industry giants like Samsung, Qualcomm, ARM and others. While currently, AMD is the only active member at creating hardware and pushing the software development, it is very possible if AMD is able to succeed at getting developers onboard to create software to take advantage of HSA, we will see others to jumps onboard with other hardware designs. Furthermore, as mentioned, AMD’s future roadmap that includes both x86 and ARM SoCs, HSAIL would allow software to be used on either processor without the need to be rewritten. Kaveri marked the world first OpenCL 2.0 ready chip. Though, the driver is not currently available to take its hardware advantage. AMD promised a driver to be released toward the end of the year. It will first be on Linux system. [section_title title=KAVERI: Steamroller CPU and GCN GPU] Meet KAVERI: STEAMROLLER CPU AND GCN GRAPHICS Compute Cores As a company, AMD is always trying to get the industry to think differently. Way back in the Athlon era, the company introduced the concept of performance numbering system as oppose to the clock speed. That worked out well as it gets people to understand that clockspeed alone does not dictates the overall performance. When the Fusion platform was launched in 2011, AMD started to call their processors with the integrated graphics APU. Now, AMD is once again trying to redefine what we think as a processor as it is now introducing the concept of “compute cores”. The compute cores combines the number of CPU and the GPU cores together. In the case of the A10-7850K, it comes with 12 compute cores that breaks down to two CPU and eight GPU cores. This maybe a marketing gimmick to get people to think that its APU has more cores and thus offers better performance. Obviously, educated buyers will know that it is not just the number of cores but other factors such as architecture and clock speed together will dictate the performance. We understand AMD’s reason for doing this as it is trying to push the concept of the total compute power of APUs with HSA and heterogenous computing. Unfortunately, this can be little bit misleading as you can have two APUs with same numbers of compute cores that is made up different numbers of CPU and GPU cores. Furthermore, the compute cores gives us no information regarding to the clock speed of the CPU and the GPU, so even if two APUs shared the same compute cores, it may not give us similar performance. So, counting just the compute cores is not as clear indicator to a chip’s raw performance. At least, AMD is being very open about the number of CPU and GPU core on the box labels where you will see the number of the CPU and GPU cores in parenthesis next to the total cores. Additionally, at its current state, AMD has a much simpler product line up in their APU where we only have two models in the A10-7000 series where the difference is the CPU clockspeed and the number of the GPU cores as the GPU is clocked at the same 720MHz on both models. So it is not too difficult to distinguish between the two. However, going forwards, if the number of the models increased, then it can be a very challenging to distinguish one unit from another. Currently, it is still hard to quantify a CPU/APU without talking about pure CPU and pure GPU performance as the HSA compatible software selection is very limited. HSA is still in its infancy; once and if the technology has gained a wider market, we may have to finally rethink how to quantify a processor’s total compute power that takes into account of all of the processor cores inside the chip and the clock speed. Steamroller CPU Core: 28nm While the Richland is build on the 32 nm fabrication process, Kaveri is build on Global Foundaries 28 nm SHP technology. The die reduction allows AMD to pack even more transistors on Kaveri. Kaveri's die size is comparable to Richland where it has areal density of 245 mm² compare to Richland’s 246 mm². However, Kaveri is packed with 2.41 billion transistors; almost double the number compare to 1.3 billion on the Richland. The increase in the transistor count is largely due to the GPU as 47% of the die on the Kaveri is devoted to GPU. GPU design focuses on the transistor density, this is the reason why AMD decided to go with the Global Foundry 28nm SHP. The trade off here is that CPU frequency at higher TDP is lowered but AMD make it up with higher instruction per clock which the company estimates up to 20% improvement over the previous generation. For AMD to compete against Intel and it needs to address its weakness. Thus, the focus on Kaveri is efficiency and power consumption. As a result, the maximum clockckspeed on the Kaveri has been reduced. Our A10-7850K has a base clock of 3.7GHz and turbo speed of 4.0 GHz compare to the A10-6800K’s 4.1GHz and 4.4GHz respectively. However, refinement on the CPU architecture allows the chip to perform just as fast as its predecessor at the same clock speed or even faster yet at the same time consumes much less power. Kaveri CPU core is refined Piledriver that is still based on the Bullozer’s module design with two integer units paired with a float point unit. The third generation chip is codenamed "Steamroller". The chip has up to 2MB of L2 cache per module or up to 4MB total as we will only find Kaveri comes in either single module (dual-core) or dual modules (quad core) variant. There is no new ISA on Kaveri as it still supports FMA4/3, AVX, AES, and XOP. With Steamroller, the fetch on the APU has been improved where AMD reduce the i-cache misses by 30% and reduced the mispredicted branches by 20% and increase in the scheduling efficiency by up to 10%. The max-width dispatches per thread has been increased by 25% and improved in store handling. Together AMD estimates these improvements helped to gain the 20% IPC over the Richland. At the same time, AMD also lowered the power consumption usage on the CPU. Desktop Kaveri will come in 45W, 65W, and 95W, a 5 watts reduction from the 100W TDP on the fastest Richland/Trinity. KAVERI also brings configurable TDP to the desktop. So, it is possible now to purchase a faster APU say A10-7850K and run it at 45W or 65W when it is constrained by the power or cooling. Then, once the constraints are lifted, you can run the APU at its full potential. Kaveri's memory controller also gets a speed boost where it now supports DDR3-2133. Only a selected model (namely A10-6800K) support DDR3-2133 while most off the 6000 series APUs only support DDR3-1866. GPU: GCN While the CPU seems a minor improvement, the GPU on the Kaveri gets a major overhaul. Kaveri now uses the same design as “Hawaii”, the same GPU found on the AMD current generation R9 290X dedicated graphics. Hence it supports 8xAA and 16xAF, DirectX 11.2, AMD Eyefinity, 4K Ultra HD, and DisplayPort 1.2. Kaveri’s GPU features up to 8 graphic compute unit where each unit features 64 unified shader, 4 texture mapping units, and one render output units. The GPU features IEEE2008 compliance, texture fetch units, registers and precision improvements. The GPU is clocked at 720MHz on all three models (A10 and A8) that AMD has launched. We are not sure if it will be the same for future models. As with previous APUs, Kaveri also supports AMD dual graphics. The Kaveri will be able to pair with the R7 based Radeon GPUs with DDR3 memory for additional performance. Having the same microarchitecture as the desktop GPU, Kaveri inherited all of the features that is found on the Radeon graphics. Kaveri features TrueAudio, Mantle, dedicated video coding engine (VCE) 2.0, unified video decoder (UVD) 4, The VCE 2 adds support for the H.264 YUV420 B frames and H.264 YUV444 I frames for 60Hz wireless display. The UVD 4 adds improved error resiliency for H.264/AVCHD. Mantle is a low level program API for game developers to improve gaming performance by reducing the overheads. Single thread performance on AMD CPU has been lagging behind Intel for quite awhile now so this could be one area where AMD get a lot of benefit if more workloads are pushing toward its GPU. As you may recall, TrueAudio was on all of the AMD GCN 1.1 or higher GPUs (R9 and R7 GPUs). The idea here is to offload the audio processing to the special DSP on the chip as oppose to the CPU. Thus, CPU resources are not tied to the audio processing in game and can be used for other workloads or power can be saved with lower CPU usage. With TrueAudio to take off the workload for the audio processing, it is possible if games are able to deliver a more true to life audio with greater affect if games are written to use its dedicated DSP. And this is ultimately going to be the major task for AMD not just on the APUs but also on their Radeon graphic cards with TrueAudio DSP-- get the software developers on board to use it. Currently, there are a less than a handful of games—Thief and Battlefield 4– that can take advantage of the TrueAudio and Mantle. Given to the fact these technologies are not found just on the APUs but also Radeon GPUs, we think that it would have a much wider support by game developers. But only time will tell. [section_title title=Chipset and Line up] Chipset Now days, CPU and chipset are coming from the same manufacturer, we often get a new chipset along with a CPU release. AMD has traditionally offered a much longer upgrade path with their motherboard where the boards often lasts two generations. The FM2 socket introduced with the Trinity is compatible with the Richland but sadly with Kaveri, you would need a new motherboard. Kaveri requires a new FM2+ socket where it has slightly different pin configuration with two additional pins. The FM2+ boards will actually be backward compatible with Richland/Trinity FM2 APUs as well as Kaveri but unfortunately the reverse is not true. If you have gotten a FM2+ board that is running an FM2 APU (Richland/Trinity), a BIOS update is all you need in order to use Kaveri. The official supported chipset or Fusion-Controller Hub (FCH) for the Kaveri are A55, A78, and A88X. The A55 was the chipset that AMD has launched since the FM1 and was continued to the FM2 chipset and now AMD continues to use the same chipset for the FM2+ socket. The A78 FCH, from what we can tell, is the same as the A75 except that it gained FM2+ socket support. Finally, the A88X is a new chipset that is slightly upgrade from the A85 chipset where it now adds support for not just FM2+ socket but also the PCI Express 3.0. This marks the first chipset from AMD that supports PCI Express 3.0. It also supports CrossFire where you are able to split the 16 PCIE lanes into dual 8 lanes. Other than these, the chipset still has the same number of USB and SATA ports: 4 USB 3.0, 10 USB 2.0, and 8 SATA 6 Gbps ports. Chipset A55 A78 A88X PCIE 2.0 2.0 3.0 PCIE Configuration 1x16 1x16 1x16 or 2x8 SATA Ports 6 SATA II 6 SATA III 8 SATA III USB Ports (3/2/1.1) 14/2 4/10/2 4/10/2 RAID 0,1,10 0,1,5,10 0,1,5,10 Line UP At launch, AMD released four models of APUs. The two 95W models are A10-7850K and the A10-7700K. Both APUs come with four CPU cores except that the A10-7850K is clocked at 3.7 GHz base and 4.0GHz Turbo while the A10-7700K is clocked at 3.4GHz base and 3.8GHz Turbo. In addition, the A10-7700K has two fewer GPU cores than the A10-7850K but both APU have their GPU clocked at 720 MHz. The 65W and 45W model carried the A8-7600 branding with six GPU cores clocked at 720 MHz and comes with four CPU cores clocked 3.3 GHz base and 3.8 Turbo for the 65W model and 3.1 GHz base and 3.3 Turbo for the 45W model. AMD since added a new A10-7800 that has the same number of cores as the A10-7850K but has a lowered clockspeed and lower TDP. In addition, the A6-7400K is also added where it has two CPU cores with four GPU cores. This particular model is different than others as it only comes with two CPU cores and 1MB L2 cache. It also supports DDR3-1866 as oppose to 2133MHz but has a slightly higher (36 MHz) GPU frequency than other APUs. [section_title title= Test Setup and Overclocking] Test Setup Test Rig Case Open Air CPUs AMD A10-7850 KA10-6800K A10-5800K Intel Core i3 2100 Core i5 2500K Core i7 3770K Motherboards AMD: MSI A88X-G45 GamingIntel: Gigabyte GA-Z77X-UD3H Ram Kingston G.Skill RipJawX 2133MHz CPU Cooler Thermalright SI-128 Hard Drives Seagate Barracuda XT 3TB SSD 1x OCZ Vertex 3 Max IOPS 240GB SATA III 6Gb/s SSD OVERCLOCKING Overclocking the A10-7850K is very simple as it features unlocked multiplier. We were able to quickly raise the CPU clockspeed from the default 3.7 GHz to 4.3 GHz in a matter of just a few minutes by raising the core voltage to 1.4V. While we were able to get the APU to boot into Windows at higher clockspeed it will not pass Prime95 stress test. Even when we raise the voltage by 0.025V, we were still not able to get the Prime to run stable for at least 5 minutes. So we settle for the 4.3GHz. [section_title title=Benchmarks: AIDA64] AIDA64 CPU Comparing the A10-7850K against the previous two generation APUs from AMD, we can see that the CPU performance for the Kaveri is a bit of a mixed bag. On tests such as PhotoWorxx, Hash, and VP8, it comes ahead with 2% to 20% improvement in performance. However, tests such as Queen, FPU Mandel, and FPU Julia, it comes in even slower than the A10-5800K. Clearly when a software demands higher clockspeed, the Kaveri will suffer slightly due to it's reduction in the frequency; but as you see with the Hash test, if a software is optimized with the latest instruction sets, then it could offset the frequency reduction and delivers a good performance gain. Memory Kaveri brings DDR3-2133 memory support. We ran the memory test with DDR3-2133 on all of our AMD systems and with the Intel system, we ran the memory at 1600 MHz as that is the officially supported frequency for the Core i5 2500K. The memory read on the Kaveri has improved by 5% over the A10-6800K. The biggest improvement is in the memory write where we get 9.4% higher bandwidth. Memory copy also gets a big improvement where we observed a 25% improvement of over the A10-5800K. Though, the memory latency on Kaveri took a big hit where it is 30% slower than the A10-6800K. [section_title title=Benchmarks: CineBench and TrueCrypt] CINEBENCH R11.5 Cinebench in particular loved higher frequency and as a result, the A10-7850K took a hit here where it is about 10% slower than the A10-6800K (which is expected as the turbo frequency is reduced by 10%) Multi-thread performance also took a bit of hit but only about 2% despite the 400MHz reduction in the clockspeed compare to the A10-6800K Cinebench 11.5 OpenGL has the A10-7850K coming in only about 2% faster than the A10-6800K. Cinebench R15 Let's switch to the latest edition of Cinebench to see if the performance difference remains the same. Like the Cinebench 11.5, the latest edition also has the A10-7850K coming behind the A10-6800K. Similar to what we have already seen, the single thread shows a much bigger difference in the performance but the multi-thread is much closer between the two where there is just a very small difference. However, OpenGL tests shows a much better performance gain for the A10-7850K where it is 8.5% higher than the A10-6800K. TrueCrypt A10-7850K performs quite well in TrueCrypt where it get 10% improvement over the A10-6800K despite 400MHz reduction in the clockspeed. Truecrypt often scale quite linear with the CPU clockspeed of the same architecture so this test gives us a good idea just how much architectural improvement on the Steamroller over the Richland/Trinity. [section_title title=Benchmarks: Media Expresso, x264, HandBrake] Media Transcoding With MediaExpresso, we did not notice any difference in the performance between the AMD APUs. The latest X264 HD benchmark gets a little bit of performance boost where we can see the A10-7850K comes ahead of the A10-6800K despite running at slower clockspeed. Same with the HandBrake that supports OpenCL and the A10-7850K gets a sizable improvement over the A10-6800K. The improvement on the UVD and VCE and the GNC architecture on the GPU most likely contribute to the performance improvement. Still with the improvement, Intel CPU still dominates over AMD thanks to the QuickSync transcoding. [section_title title=Benchmarks: PCmark 7 and PCMark 8] PCMark 7 PCMark 7 does not support OpenCL so as a result, the A10-7850K is slower than the A10-6800K. With this test, you can get a sense that under the worst case scenario, the A10-7850K would give about 90% of the performance of the A10-6800K. PCMARK 8 The latest PCMark 8 adds support for OpenCL. Thus, A10-7850K performs much better here. There is 6% improvement in the Work, 20% in Creative, and 30% in the Home benchmark. [section_title title=Benchmarks: 3DMark11, 3DMark, and games] 3DMark11 When comes to gaming, the A10-7850K is about 20% faster than the A10-6800K with 3DMark 11 benchmark. Looking at the 3DMark Fire Strike, we can see the A10-7850K GPU can deliver 20 to 30% improvement over the A10-6800K. However, we can see that the CPU performances are mixed. We toss in a couple of the real-world gaming to see how well the integrated GPU performs. As you can see, the A10-7850K is able to driver 10~20% improvement. Notice that with Metro 2033, while the difference in the average frame rates is small, the minimum frame rate on the A10-7850K is 10% higher than the A10--6800K. The R7 GPU on the A10-7850K delivers the performance about $50 discrete GPU. It is still not fast enough to give us 1080p game play at medium to high setting but if we tone down the setting just a tad, it is possible to have at least a smooth game play. At 1680x1050 resolution, the A10-7850K is probably playable at medium settings. Obviously, with AMD APU, you could always toss in a compatible graphic card for crossfire and with the A10-7850K, you could add a cheap R7 card that would yield even better performance. Still, it is probably better for most gamers to spend a little more and go with a midrange GPU for a bit higher visual effects. [section_title title=Benchmarks: OpenCL Performance] OpenCL Performance We turn to LuxMark to measure the APU's compute power. Starting with the CPU,we can see that the A10-7850K actually is slower than the older APUs due to the fact it is running at lower clockspeed. However, GPU compute on the A10-7850K has doubled from the A10-6800K. Combining the CPU and GPU, we can see that the A10-7850K is about 30% faster than the A10-6800K. While the additional improvement still not able to get AMD's APU to compete against the Intel's offering, it at least putting the chip in a much more competitive category. Obviously, AMD still needs software developers to optimize their codes in order to truly unleash its fullest potential. Basemark 1.1 Looking at the basemark benchmark and we can see that Kaveri is about 60% faster than the A10-6800K. When we break down into individual tests, we can see that the performance ranges from 30% to 100% improvement over the A10-6800K. The image test gets the best improvement probably due to the elimination of the needless copying of data between the GPU and the CPU since this test focus on the bandwidth limitation stemming from moving data between the CPU and the GPU. [section_title title=Power Consumption and Configurable TDP] Power Consumption The idle power for the Kaveri is 46W, a little bit higher than the the A10-6800K but lower than the A10-5800K but they are essentially stayed the same. Under load, however, we see a big drop in the power consumption where the A10-7850K consumes 12 watts fewer than the A10-6800K. In fact, despite its 95W TDP, we see the A10-7850K consumes actually 83 watts more power under load compare to the idle. One thing we did notice with the Kaveri is how the power is being distributed in the chip. In the past, when we stress both CPU and CPU, we saw both cores will run at the maximum frequency. However, it appears that Kaveri gives the GPU the priority when both CPU and GPU are stressed as we notice that the clockspeed on the CPU is reduced to 3.0 GHz while the GPU clockspeed remained 720MHz. Kaveri reserved about 30 watts of power to the GPU and the rest of the power is left for the CPU to use. Configurable TDP: 45W, 65W, and 95W Kaveri features configurable TDP where you would be able to lower the chip's power consumption by reducing the clockspeed. As we mentioned earlier in the power consumption section, Kaveri seems to give preference to the GPU when is under load and this is the same when the TDP of the chip is altered. When we set the processor’s TDP to 45W, we noticed that the CPU often runs somewhere between 2.5GHz to 3.0GHz where 2.5GHz seems to be the base clock when all 4 cores are active and 3.0GHz is the Turbo clockspeed. At 65W, the CPU seems to run at 3.5GHz for most of the time regardless if it is lightly threaded or heavily threaded. We noticed that the GPU clockspeed does not seem to vary much where it runs at 720 MHz at any TDP. We decided to run a few benchmarks to show the performance difference with the APU running at 45W, 65W, and 95W. The chart below gives us a sense just how much performance differences are there. We included the percentage of performance improvement compare against the 45W. Specifications 45W 65W 95W CPU Max Frequency (MHz) 3.0 3.5 3.7 Cinebench R15 CPU (single/multi) 71/228 82/270 (15.19%/18.42%) 93/318 (30.98%/39.47%) Cinebench R15 (OpenGL) 37.19 40.56 (9.06%) 46.07 (23.87%) 3DMark Fire Strike Overall 1200 1303 (8.58%) 1354 (12.83%) 3DMark Fire Strike Graphics 1305 1422 (8.97%) 1473 (12.87%) 3DMark Fire Strike Physics 3042 3744 (23.08%) 4469 (46.91%) PCMark-Home 3045 3232 (6.14%) 3299 (8.34%) x264-2 pass 18.03 21.21 (17.64%) 24.555 (35.91%) Looking at the performance scaling as we move from 45W to 65W and we can see that it delivers 6 to 20% improvement. The single thread and GPU performance seems to be about 10% improvement whereas the multi-thread performance shows a larger gain at around 20%. Moving up to 95W shows a similar trend though the improvement over the 65W TDP shows a smaller range from 2% to 17%. [section_title title=Conclusion] Conclusion On the CPU front, Kaveri did not make a giant leap that we have hoped. It is a bit of a mixed bag where the reduction in the clock speed hurts its performance in apps such as Cinebench that demands on higher clockspeed. On the other hand, for the software that is optimized for, we see about 10% improvements in the performance. The single-thread performance seems to have suffered the most on Kaveri and this is an area where AMD definitely needs improvement. Multi-thread performance in general is on par with Richland/Trinity and often exceeds its predecessor. While we do not expect Kavreri to actually out-perform Intel, we at least hope it would be able to improve a little more especially the single thread performance as that has always been a sore spot on the AMD APU. The graphic portion of the Kaveri is an evolutionary upgrade and the move to the GNC is a much welcomed. There is about 10~20% improvements on the graphic performance. The switch to the GCN along with the array of features such as TrueAudio, Mantle, and discreet JPEG accelerator that AMD has bundled into the new APU puts the GPU on the APU at the same level as the current generation discreet Radeon graphics. While some of these features still need software support, it is nonetheless good to see the APU to have the same hardware as AMD’s own discreet GPU where it should help with gaining greater support from the developers. Obviously, AMD’s goal with Kaveri is deliver a higher performance per watt ratio. And if we take the power consumption into account, then Kaveri is clear the winner here. The A10-7850K that we have here only is rated to be 5 watts lower in the A10-6800K. In our test, we noticed the APU to consume much lower power than the Richland and Trinity. So clearly, Kaveri is a much more efficient chip. And we are glad to see AMD is putting a greater emphasis on reducing its power consumption. Consider the fact that the A10-7850K is retailed at $170, it maybe a tad too expensive to compete against Intel's offering at similar price range. It is sandwiched between the fastest Core i3 and the entry Core i5. Intel's offering at this price has lower TDP and performs better on the CPU front. Obviously, Kaveri’s GPU is faster but if you are gamer, you would be better off just pair an Intel CPU with a discreet graphic card. This makes Kaveri (A10-7850K) hard to recommend for anyone who plans to use a discreet GPU. However, there is a market that Kaveri would be perfect for, which is systems that relies solely on the integrated GPU where a discreet GPU is out of the equation. The reduction in the power consumption on the A10-7850K and the support for the configurable TDP actually makes it a very competitive product for systems where discrete graphic is not necessary or non desired. We think with its balanced CPU and GPU performance and low power consumption, Kaveri would be perfect for all-in-one PC and small form factor systems such as media servers and HTPCs. Mobile Kaveri would also be very interesting product as well if it can deliver similar performance at low power consumption. If we look at what AMD has done with the whole Fusion platform/APU. We can see that the CPU front essentially stayed the same; while each generation, there is some improvement, it also takes a bit of regression in some area. AMD continues to improve on the GPU in order to make the CPU and GPU more of an equal partner and at the same time reduce the power consumption of the system. However, this is done at the expense of the CPU performance as it has essentially reached a plateau. We cannot help but think that Kaveri should have been the first APU from AMD back in 2011 as the successor to the K10 CPUs. Obviously, it was not possible for AMD to deliver Kaveri back then, or we would have seen it. However, if AMD were able to deliver then, we may have a much different landscape today where there would be more software that can take advantage of the heterogeneous computing and utilizing HSA. Still, Kaveri is still exciting as it is a decade long work from AMD where it is finally ready to demonstrate its vision and promise. There are still lots of work needed to be done and most importantly, AMD needs to get the software developers on-board to create software to finally take advantage of the technologies and features that is on the chip. Mantle, TrueAudio, hUMA, and hQ all require software written to take advantage of their powers. At the moment, the FM2+ platform/Kaveri seems more like a developer’s platform where it is packed with many features waiting for the developers to tap into its power. If AMD is able to attract more developers to finally take advantage of its platform and the new technologies, then we maybe are on the cusp of something greater in the next couple of years as HSA finally takes mainstream.