AMD A10-7890K APU Review

Not open for further replies.


Mar 12, 2012
i see no point in this when you can get a 7860k godrvari apu, at a lower watt rating, thats cheaper and overclock it.

The one value proposition of the 7890k that I see is the bundled Wraith Cooler. I wish toms had a performance benchmark on it because if it's good enough to cool your overclocked 125W APU (and even the 7860k should be rated at 125W TDP) then you won't have to spend the money on aftermarket cooling like I used to recommend. But considering you can get capable aftermarket cooling for as little as 15-20$, the 7860-7890 price difference is going to have to be quite small.
$180 for 7890K?
This is too much. AMD needs to drop prices a bit cause APUs have potential in the market but with these prices, it will target only people that don't have enough space...
You can buy an FX6300 with R7 360 2GB that they will blow 7890K out of the water for the same money.
PCPartPicker part list / Price breakdown by merchant

CPU: AMD FX-6300 3.5GHz 6-Core OEM/Tray Processor ($88.99 @ SuperBiiz)
Video Card: XFX Radeon R7 360 2GB Core Edition Video Card ($100.99 @ SuperBiiz)
Total: $189.98
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2016-03-22 08:50 EDT-0400


Mar 16, 2016
I have no clue why no one has made a 1080p 60HZ IPS free-sync display, there are TN ones for around $150, but who the hell wants a non 144hz TN display in this current year?


The main takeaway I got from the article is that FreeSync doesn't make much sense on budget APU builds right now, because of the cost displayport-capable mainboards add. For that value segment, HDMI FreeSync is more interesting.

Still, I bet that anyone looking at a budget APU rig isn't aiming for ANY adaptive sync solution, at least not yet. I mean if you eliminate that from the equation you can build the entire system for even less money and either use a pre-existing display or buy one for way less.
As an AMD fanboy, this just came out as a disappointment. I hope Excavator does well. I might just wait for a Zen APU in 2017.

The Zen is due this October. Should have 8 cords and multithreading and about the same instructions per clock as the Broadwell. Will be a great chip.


Aug 11, 2011
How about a DX12 test like running Ashes of the Singularity on the APU against the x4 760K+R7 250 against the APU+R7 250? In my opinion a very good use of an APU is to complement a dedicated GPU, even an entry level one.


Jan 10, 2011
I hope AMD becomes more competitive, this is the wrong road to take. APUs should be cheap and reliable for general computing and cost less than a cpu+discrete gpu.
That less cost should be giving same performance/price ratio as a cpu+discrete gpu.

It is the wrong way to try and push APUs being more expensive for some fps more. I would also like to see APUs crossfireX (not sure its called that) being something cool and actually give some oomph.


Dec 29, 2015
You compared this chip to the $370 Intel i7-6700k, WHY? Then you compared it to the out of date Haswell i3, WHY?

Give a nice i5-6500 a try with the same HD 530 GPU in it that the i7 has. Heck, the i3-6100 has the same GPU and may well perform nearly as well as the i7 did.



Not every reviewer has every CPU available all the time.

The i7 shows Intel's best-case IGP performance while the i3 shows where the A10 stands in terms of CPU performance. I'd call those fair enough reference points with the take-away being that if you use an external GPU, the i3 performs about on par on the CPU side but if you want to use the IGP, you are better off with the A10 which performs ~20% better on the IGP side than even Intel's best.

The point of using the i3 instead of i5 was to show that the A10 competes at i3 level. If you want to say...yeah but A10 graphics are better than I'd say take a cheaper 760k and pair it with low end gpu, 250. That'd outperform the A10 as in results above. Also, by using the i7 you can see how the APU compares editing and CAD work...APU is left in i7 dust.

So, I think using the i3 and i7 was sufficient. Keep in mind the more CPUs used the more time it takes to publish, and there are more pressing matters than that.


Apr 27, 2015
People you should have tested this at 720p resolution! apus are not made to be used on 1080p, unless paired with a discrete vga.. and only 1 of those cases was that.. next time use the correct resolution please.


Apr 4, 2011

This is called "unbiased (representative) sample" and is a method to get the best results with only a few samples to increase the efficiency of a working process. This is the difference between an unplanned benchmarking of tons of similar samples or a well-designed selection process before. Less (but good selected) samples means also a better (to understand) overview without unnecessary filler data (bars). This looks more important but is at the end really senseless. :D

AMD said, this APU is for 1080p online gaming. So I tested it as requested. ;)

Especially for the 720p guys a brandnew benchmark:

(from me and the German site)
Not open for further replies.