Overclocked Kaveri Benchmarks
|A10-7850K||Core i3 4340
Core i5 4670
AMD A88X Chipset
ASRock FM2A88X Extreme6+
AMD A88X Chipset
Intel Z87 Chipset
|Memory||16GB (2x8GB) Crucial Ballistix Sport XT 1866 10-10-10-30
8GB (2x4GB) Radeon Memory GAMER Series 2400 11-12-12-31
(on Overclocked system)
|Hard Drive||OCZ Vector 256GB
|Video Card||Diamond Multimedia ATI Radeon HD7870 2GB
Sapphire Radeon R9 280X (for gaming tests)
|Motherboard Drivers||Catalyst 13.30 RC3 Chipset|
Catalyst 13.12 SB
|Video Drivers||GPU: Catalyst 13.12|
IGP: Catalyst 13.30 RC3
|GPU: Catalyst 13.10
IGP: Intel 10.18.10.3345
|Operating System||Windows 8 Professional x64 RTM
All updates as of Dec 1, 2013
We’ll be looking at most of our CPU test suite, with some left out such as OpenCL benchmarks and IGP benchmarks since we’re only looking at CPU performance right now. We’ll be including the A10 7850K at stock speed, along with Intel’s Haswell based Core i3 4340 and Core i5 4670. In our original Kaveri review, the 7850K would sometimes struggle to keep up with the Core i3, so it should be interesting if its unlocked nature gives it such an advantage that it meets or exceeds it in performance.
Again, for a more in-depth CPU review, be sure to check out our original Kaveri A10 7850K review.
SiSoft Sandra 2014
We always start off our CPU reviews with SiSoft Sandra’s compute tests, as it gives us an idea of how each CPU performs in a ‘pure arithmetic’ and an ‘accelerated arithmetic’ scenario, showing us where each may have strengths and weaknesses. None of these are truly real-world performance results, but are still interesting to look at.
We already know where Kaveri’s weaknesses lie in having only one FPU scheduler per Steamroller module, but the overclock allows it to overcome this issue. At least enough to catch up to the Core i3 which also only has two cores, although runs four threads.
With the more optimized and more “real world like” Multimedia benchmark, even a huge overclock is not enough to catch up to Core i3’s performance.
Gaming With Overclocked Kaveri
We usually break our gaming results into separate pages, since there is quite a bit to go over. This time we’ll just put them all on one page, since we don’t expect huge performance improvements just by overclocking the CPU. For details on how we test each game, check out those particular pages in our Kaveri review.
We are using a Radeon R9 280X for these tests, and the settings shown on each chart. For a video of each benchmark run, you can check out our Youtube channel. Each frame time chart can be clicked on for higher res version with more detail.
Assassin’s Creed IV saw a moderate performance boost of around 2 FPS, but as you can tell by looking at the frame time chart, this doesn’t actually amount to much. Both Intel lines are much more steady, almost a perfect 60 FPS score. The Core i5 line is even tighter than the Core i3 line, despite the nearly identical FPS result. This tells us that the Core i5 actually delivers a smoother experience compared to Core i3, which is something you’d never know just by looking at FPS.
The FPS improvement in Battlefield 4 is much more substantial, but again if you look at the frame time chart, Core i5 is clearly a better performer.
Crysis 3 puts a beating on the R9 280X, but it still manages to be fastest on the Core i5. If you look at the frame time chart, you will see a much ‘tighter’ line for the Core i5, which means there is a much smaller variance of frame times. This leads to a much smoother feel that is usually not shown when the results are lumped up into a single FPS score.
Metro Last Light scores within the margin of error, so overclocking the CPU does nothing here. Although all four CPUs allow the R9 280X to produce pretty consistent frame times this time around, the Haswell CPUs achieve a much higher framerate overall.
Tomb Raider appears to allow the R9 280X run at pretty much full potential with these settings, no matter which CPU is being used. You can see that the Core i3’s frame times result in a ‘fatter’ line although it probably isn’t ‘fat’ enough to be noticeable compared to the others.