Is an RX 460 really an upgrade over an APU?

This time the answer is yes…


The with launch of AMD Radeon RX 460 last month we’ve come to a point where we can reconsider the age-old question of the poor gamer; is buying a cheap discrete graphics card worth it? Ages ago before the dawn of decent on-die graphics the answer to the question was always yes. Then in about 2011 or so AMD an Intel started looking at graphics performance as a valid area for competition between their processor offerings.

APU’s Changed the PC Gaming Market

The advent of APUs and on-die graphics left us in a weird place where sub $100 low-end discrete graphics solution really didn’t perform better than what came on your chip. AMD and Nvidia saw this trend and the responded by not updating their entry-level GPUs. You can still buy graphics cards like AMD’s old Radeon R5 250 or Nvidia’s GeForce GT 720 but they are truly woeful solutions that fail to offer a performance increase significant enough to justify their expense for most use-cases.

With its Polaris generation of chips AMD has opted to fill out the bottom of its product stack with the relatively powerful RX 460. AMD’s old bottom end offering the R5 250 had 6 GCN compute units, AMD’s APUs have 8 GCN compute units, and the RX 460 has 14 GCN compute units. Given that AMD’s RX 460 shares a similar price point to what the old R5 250 had when it launched in 2013 it’s good to see that it offers more than double the raw execution units.

Quantifying the Gap between APU and Discrete

On paper the RX 460 is a winner, but how does it perform in reality? To find out we broke out our A10-7870K test system and found the highest playable settings for all of the games we’re testing in this article. To make a long story short those playable-on-an-APU-setting were 1080P, high details, and no AA. At those setting we saw between 30 and 60 FPS in the popular esports titles that newer PC gamers tend to favor.

Then we dropped our RX 460 in and ran the same tests again at the A10-7870K’s best playable settings to quantify the performance gap between the two solutions.

AMD 7890K

Too Long; Didn’t Read

If you’re a subscriber you can now find our testing methodology, exact hardware configurations, a detailed justification of these benchmarks, and our raw testing data at the very end of the article. We’ve opted to place that information behind our paywall with the acknowledgement that few people actually read those parts of our reviews and with the goal of making these articles easier to read for people who aren’t interested in those aspects of our content.

Transparency Statement

For the sake of transparency, we want you to know that AMD provided both of the APUs, the GPU, and the DDR3 memory we’ll be testing in this review. All the other components were purchased at retail and without the knowledge or consent of those companies.


The RX 460 ranges from merely 1.5 times faster than an A10-7870K to over 3.7 times faster in our testing. This is a convincing win for discrete graphics especially when you consider that our RX 460 sample retails for $140.

A Graphics Card is Clear Step Up

What we have today with the RX 460 is an entry level graphics card that is clearly and without question better than any APU. If you’re looking to upgrade to discrete graphics you will absolutely notice an improved gaming experience by dropping and Radeon RX graphics card into your system. In many ways the future can be a terrifying prospect, but if you’d told me five years ago that in 2016 low-end discrete graphics card wouldn’t be a term synonymous with garbage I probably would have suggested that this reality is the best possible outcome.

Note: The following is for professional and student level subscribers.

Disclosures: Charlie Demerjian, Thomas Ryan and Stone Arch Networking Services, Inc. have no consulting relationships, investment relationships, or hold any investment positions with any of the companies mentioned in this report.

The following two tabs change content below.
Thomas Ryan is a freelance technology writer and photographer from Seattle, living in Austin. You can also find his work on SemiAccurate and PCWorld. He has a BA in Geography from the University of Washington with a minor in Urban Design and Planning and specializes in geospatial data science. If you have a hardware performance question or an interesting data set Thomas has you covered.