AMD A10-7800 Review: Testing the A10 65W Kaveri
by Ian Cutress on July 31, 2014 8:00 AM ESTCPU Benchmarks
The dynamics of CPU Turbo modes, both Intel and AMD, can cause concern during environments with a variable threaded workload. There is also an added issue of the motherboard remaining consistent, depending on how the motherboard manufacturer wants to add in their own boosting technologies over the ones that Intel would prefer they used. In order to remain consistent, we implement an OS-level unique high performance mode on all the CPUs we test which should override any motherboard manufacturer performance mode.
HandBrake v0.9.9: link
For HandBrake, we take two videos (a 2h20 640x266 DVD rip and a 10min double UHD 3840x4320 animation short) and convert them to x264 format in an MP4 container. Results are given in terms of the frames per second processed, and HandBrake uses as many threads as possible.
The latest Intel processors have the lead for low quality Handbrake conversion, and despite the generational gap between the FX-4350 and the A10-7800, the extra MHz is preferred here.
For large frame manipulation, the latest architectures mixed with the most threads perform best.
Agisoft Photoscan – 2D to 3D Image Manipulation: link
Agisoft Photoscan creates 3D models from 2D images, a process which is very computationally expensive. The algorithm is split into four distinct phases, and different phases of the model reconstruction require either fast memory, fast IPC, more cores, or even OpenCL compute devices to hand. Agisoft supplied us with a special version of the software to script the process, where we take 50 images of a stately home and convert it into a medium quality model. This benchmark typically takes around 15-20 minutes on a high end PC on the CPU alone, with GPUs reducing the time.
Dolphin Benchmark: link
Many emulators are often bound by single thread CPU performance, and general reports tended to suggest that Haswell provided a significant boost to emulator performance. This benchmark runs a Wii program that raytraces a complex 3D scene inside the Dolphin Wii emulator. Performance on this benchmark is a good proxy of the speed of Dolphin CPU emulation, which is an intensive single core task using most aspects of a CPU. Results are given in minutes, where the Wii itself scores 17.53 minutes.
Dolphin seems to work best with high single core speed and Haswell.
WinRAR 5.0.1: link
PCMark8 v2 OpenCL on IGP
A new addition to our CPU testing suite is PCMark8 v2, where we test the Work 2.0 and Creative 3.0 suites in OpenCL mode. As this test is new, we have not run it on many AMD systems yet and will do so as soon as we can.
The combination of processor graphics and OpenCL support push the AMD APUs up to the top of our PCMark tests.
Hybrid x265
Hybrid is a new benchmark, where we take a 4K 1500 frame video and convert it into an x265 format without audio. Results are given in frames per second.
Cinebench R15
Cinebench is typically Intel territory for high IPC processors, but when it comes to multithreaded rendering, extra threads help.
3D Particle Movement
3DPM is a self-penned benchmark, taking basic 3D movement algorithms used in Brownian Motion simulations and testing them for speed. High floating point performance, MHz and IPC wins in the single thread version, whereas the multithread version has to handle the threads and loves more cores.
All the calculations in 3DPM deal with floating point numbers, a known sink for AMD compute.
FastStone Image Viewer 4.9
FastStone is the program I use to perform quick or bulk actions on images, such as resizing, adjusting for color and cropping. In our test we take a series of 170 images in various sizes and formats and convert them all into 640x480 .gif files, maintaining the aspect ratio. FastStone does not use multithreading for this test, and results are given in seconds.
Web Benchmarks
On the lower end processors, general usability is a big factor of experience, especially as we move into the HTML5 era of web browsing. For our web benchmarks, we take four well known tests with Chrome 35 as a consistent browser.
Sunspider 1.0.2
Mozilla Kraken 1.1
WebXPRT
Google Octane v2
Facebook
Twitter
RSS
147 Comments
View All Comments
austinsguitar - Thursday, July 31, 2014 - link
just for shiz and giggles.... how about qualcomm start making desktop grade equipment. they need to! intel now has a total monopoly right now. cmon qualcomm come at me!jamescox - Thursday, July 31, 2014 - link
Enthusiast make cpu performance to be much more important than it actually is. For most everyday use (web surfing, web video, etc) there isn't going to be too much subjective difference between any modern cpus. For gaming, the cpu is also becoming mostly irrelevant. If you play at 1080p or greater with a dedicated gpu, you have to go down to a very low-end cpu before it will make much of any difference. This may change when DX12 comes out, since they are supposed to be making it more multi-threaded. This may favor cpus with 4 or more threads eventually, but one should not buy computer hardware based on "eventually". I wouldn't even bother overclocking a cpu for gaming these days. I would consider water cooled gpus and such for a high-end gaming machine though.AMD is presumable working on stacking memory chips in with their gpus and apus, but this will be a ways off and may be limited to mid-range or higher. Any added step in processing adds the possibility of failure of the final package, which increases cost. It seems like they could achieve near dedicated gpu performance with an integrated memory chip similar to iris pro, but perhaps this is too expensive. It also may cannibalize dedicated gpu market. Intel doesn't have that concern since they have no dedicated gpu market.
I have also wondered why they don't put a gpu-like memory interface on these apus and just solder everything onto the board. This would deliver dGPU like performance and would be great for mobile and SFF; it would be very similar to what is in the PS4. It would have low cpu performance due to power limitations, but not that many applications actually require more cpu performance these days.
AMD has more gpu power available on-die, so they need to add something like Intel's quick sync or other software/hardware features. The APUs should not lose in video encoding and such when they have more processing resources available. Is the OpenCL accelerated version of handbrake available yet? Several of the applications under cpu benchmarks can be gpu accelerated, and several of them that cannot be gpu accelerated are probably not really relevant. Any of these chips are fast enough for web browsing. I am still using an old core2duo laptop for web browsing, and it mostly works fine. The memory size (<4 GB) seems to be more of a limitation than the cpu speed.
MrVeedo - Thursday, July 31, 2014 - link
these chips serve their purpose in my opinion. amd does seem to get a lot of bashing, some deserved I will admit, but again it all depends on what you will be doing with your pc. if you shop around, you can pick up a nice board and cpu for relatively cheap. my most recent build is focused around the asrock fm2+ itx motherboard and a 7850k. using a fractal node 304 case, I have 6 hard drives and a 250gb msata ssd, along with a xonar stx sound card. my killawatt meter is in front of me and its reading a 44 watt draw while typing this. obviously the drives are asleep at the moment. the cpu is clocked at 4.1, gpu 960, mem speed at 2133. just to give an idea on power consumption.Cerb - Sunday, August 3, 2014 - link
With HP and Lenovo both using AMD APUs in mainstream notebooks, including HP in their business lines, clearly they are doing something right with the latest APUs, even outside of our bubble (like having good enough real-world performance per dollar and Watt).Cerb - Sunday, August 3, 2014 - link
I get CPU-limited in games a few years old, don't even have a >=1080p monitor, and AMD doesn't even produce a CPU as fast as mine. "Gaming," is quite varied in terms of performance. AAA shooters, and most games also on consoles, will not need much CPU, because even the latest consoles are running Atom-class CPUs (the prior gen would get laughed at by current Atoms). Games tuned for each platform, or just on PCs, tend to be much more CPU heavy, or at least capable of being so (usually resulting in what amounts to a different game).Cerb - Sunday, August 3, 2014 - link
Re more processing: they don't have more CPU processing resources than Intel's CPUs. If Handbrake intends to maintain quality over speed, FI, don't expect much gain from any GPU acceleration. GPU accel works well for some video processing, and depending on program used, can make a big difference in speed of editing videos, but not so much encoding. The image viewer is all I saw that definitely should be able to be accelerated by the GPU, though the 3D particle bench might be able to be. The rest are very much cases for CPUs, not at all GPUs. Not that there isn't a lot that can be accelerated, but of the chosen things to test, there's not much there that would be a good case for it.nofumble62 - Thursday, July 31, 2014 - link
AMD finally could produce a credible alternative for Intel Celeron.Cerb - Sunday, August 3, 2014 - link
That's a bit low. I'd take a Vishera or Kaveri over a Celeron any day, as a current daily user of a Haswell Celeron. If you were used to a 4MB/die or 6MB/die Core 2, or a Phenom II w/ L3, or better, a new brand Celeron would be a palpable step backwards, while most of AMD's APUs are quite capable, and only feel slow when faced with long-running CPU-limited tasks.wintermute000 - Friday, August 1, 2014 - link
a.) Lack of power test results is worrying.b.) Why are the gaming benchmarks done @ 1280x1024? Why not do 1368x768 and 1680x1050 and 1920x1080? i.e. the resolutions of most monitors
Aside from that, looks grim for AMD, absolutely no point to these APUs - even if the power usage is indeed efficient, we're looking @ desktop not laptop chips. The only scenario I can see them being useful is in tiny HTPC/mITX builds where you're not using a dGPU. Otherwise as many have commented, a pentium + dGPU will smoke kaveri at not much extra cost. AMD better pray that OpenCL takes off in a big big way.
Personally, I would love a ~50-80 dollar version with just a minimal IGP and ECC support for low power file server / virt platform.... intel cores are expensive :)
Anonymous Blowhard - Friday, August 1, 2014 - link
Um, you do know that the G3220 supports ECC RAM, right?