Should AMD’s Radeon RX 590 have come with a larger memory bus?

That’s the question on everyone’s minds as the Radeon RX 590 begins selling overseas. When the leaks for the RX 590 came out, everyone assumed that this was going to be a GPU based on the Xbox One X GPU featuring 2560 shader cores and a 384-bit memory bus. After all, that’s what AMD’s own documents suggested back in 2016, indicating that the second digit in the model number indicated the size of the memory bus. Now that convention is broken with the RX 590’s release, AMD is finally able to talk about their change in tactics.

After the launch of the RX 590, which we knew about for roughly half a day before the official launch, I emailed my contacts at AMD asking if AMD’s branding plans had changed for the RX 500 series. As shown below, AMD’s plans for the RX 400 family, which was the debut of the “RX” brand, laid out a neat, easy to understand naming structure that aimed to make it clearer to consumers what kind of product they were buying, and why it was better than a product that may be based on a similar name from the previous generation.

“The nomenclature for AMD Radeon RX graphics products that was laid out for the 400 series does not apply to the 500 series graphics cards,” said AMD Director, Product PR & Regional Communications for the UK Christine Brown. “Although the memory bus width of the RX 590 is identical to the RX 580 (and RX 480 before it), the previous nomenclature that was laid out in Munich was intended specifically for those parts.”

Brown added that the RX 590’s brand “was chosen to keep the models within the series uniform in differentiation and we feel it will be easier for our target customers to understand the performance differences using this branding.”

At an editor’s launch day in Munich, Germany in 2016, the company showed the slide above that would make things clearer for the consumer trying to understand the naming scheme. Theoretically, a Radeon RX 485 would be a mid-step revision of the RX 480 and it would be either faster or cheaper for AMD to produce. Perhaps there’s a new memory technology that it’s using, maybe the die design is different, or perhaps there’s a process improvement. A second revision to the brand would be enough to indicate that the card was materially better. Because it was on the eighth tier, it would be suited for 1440p gaming, and not 4K.

In that same scenario, an RX 490 would have sported a larger memory bus bigger than 256 bits, and would have been suitable for 4K gaming. AMD never made an RX 490, but if it did exist, it would probably be the GPU from the Xbox One X.

With the launch of the RX 500 series, AMD didn’t mention that the naming scheme was abandoned. It was assumed based on existing documents and launch slides that the company would continue to use the naming scheme, and launched the RX 500 series and not a revision of the RX 400 series because of material changes to the board design and performance. The RX 580 is always clearly faster than the RX 480 in benchmarks, for example. With the RX 590, we’ve gone back a step. It is faster, but it also consumes more power than is reasonable for its price point, and there isn’t any differentiating thing that makes it stand out in the lineup, apart from its name.

Because the RX 590 is AMD’s first discrete GPU available to the public using Global Foundries’ 12nm production process, this practically serves as pipe cleaner for the company. They get to test the process for the arrival of a new family of cards that they’ll surely be preparing for 2019, and iron out any bugs before launch. I look forward to seeing what they can do with it, but I also hope that they take the time to fix their branding on the GPU side of things.

GG: South Africa’s new sex ed curriculum is absolutely necessary so shut up, Karen