The AMD A8-7650K APU Review, Also New Testing Methodology
by Ian Cutress on May 12, 2015 10:00 AM ESTOffice Performance
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.
All of our benchmark results can also be found in our benchmark engine, Bench.
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 likes high IPC and clock frequency, which indicates that the highest clocked APUs perform the best out of AMD here. Because the benchmark is single threaded, even a dual core Intel wins though.
WinRAR 5.0.1: link
Our WinRAR test from 2013 is updated to the latest version of WinRAR at the start of 2014. We compress a set of 2867 files across 320 folders totaling 1.52 GB in size – 95% of these files are small typical website files, and the rest (90% of the size) are small 30 second 720p videos.
The varied-thread workload of WinRAR seems to vary between the dual core Intels and the dual module AMD chips, showing that threads matter.
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.
In this nieve benchmark, the compiler prefers x87 style commands which AMD's Bulldozer based architectures isn't too fond of. This benchmark is meant to be a representation of crude scientific code, similar to that used in a research lab. Ultimately, Bulldozer architectures such as Kaveri prefer specific commands, especially when dealing with basic math.
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
177 Comments
View All Comments
testbug00 - Tuesday, May 12, 2015 - link
Well, Broadwell is supposed to be out in 2013 according to Tick Tock. So, Intel's at least 1.5 years to their party. However, doing smaller nodes is REALLY hard. So, hard to blame them.azazel1024 - Tuesday, May 12, 2015 - link
Err, no. Haswell was 2013 and Ivy was 2012. Broadwell would have been summer of 2014. It ended up being a mostly paper launch in late fall 2014, with parts meaningfully showing up winter of 2015 and Intel mentioned up front that Broadwell would be a mostly mobile release.So they are perhaps 6 months late on Broadwell, but unless something different happens, Intel is still claiming summer/fall for Skylake, which puts them right back on the original schedule (and probably also why Broadwell is limited, Intel has known about their 14nm node issues for awhile, so they limited it to get the node out there and more experience on it and then jumping in to Skylake with both feet).
Ian Cutress - Tuesday, May 12, 2015 - link
I have the i5-5200U here in a BRIX that I can test, though it's worth noting that the 5200U list price is $281, more than any APU.takeship - Tuesday, May 12, 2015 - link
I was just thinking it wouldn't be too far off as a total cost comparison - a tall body i5 nuc + win 8 license + ram and scrounged up HDD/ssd is just about $600, which isn't too far above what a simple box with this would run. And my suspicion is that you don't give up too much gfx perf going down to the i3 and saving a hundred. Bandwidth being the bottleneck that it is.Refuge - Tuesday, May 12, 2015 - link
This SKU is new, but the chip is just a re-badge.extide - Tuesday, May 12, 2015 - link
A rebadge from what, exactly? No... It's not a rebadge, its just a lower model sku in a lineup that we have already seen. That is not what a rebadge is. We have not seen this core in this (or a similar) config released with a different sku before.Edens_Remorse - Wednesday, May 27, 2015 - link
Quit confusing the ignorants(shut up spell check, it's a word).duploxxx - Tuesday, May 12, 2015 - link
untill by the end of the year you start to see DX12 benchmarking :) and this more power silicon gets a free bump.25W is btw the difference in just a light bulb near your desktop or the minimal powerconsumption you have of dedicated gpu for the lack of onboard Intel GPU power :)
Michael Bay - Tuesday, May 12, 2015 - link
Yes, while all the action is in mobile segment. Where, you guessed it, AMD has no foothold.Not even mentioning DX12 being largely irrelevant this AND next year outside of useless synthetics.
duploxxx - Wednesday, May 13, 2015 - link
Carrizo is all about Mobile :)useless synthetics and benchmarking is all where Intel shines, the result bares show only better result, reall life daily use you don't even notice.