Introduction

AMD today announced its Radeon RX 590 graphics card. This is an unexpected product launch given the current competitive environment, and nobody expected something new from AMD until 2019. The Radeon RX 590 is designed for the vast majority of PC gamers who still play at Full HD (1080p) resolution and is priced under $300. With a number of AAA game launches lined up for the holiday, AMD is going after the crowd that's either upgrading or gifting a graphics card for gameplay at 1080p with all details maxed out in every game. Rival NVIDIA hasn't managed to address this segment with its RTX "Turing" architecture yet, and there is a big price-performance gap between its "Pascal" GeForce GTX 1060 and $360 GTX 1070, which AMD is targeting with the RX 590.

The Radeon RX 590 packs none of the exotic HBM tech from its RX Vega siblings and uses existing GDDR5 memory, which has AMD and its partners enjoy more headroom in which to adjust prices. It is based on the "Polaris 30" silicon, which is essentially a "Polaris 10" die built on the latest 12 nm FinFET node at GlobalFoundries, yielding significant energy-efficiency dividends AMD is cashing in on to increase clock speeds by 15 percent. The engine clock has been dialed up to 1545 MHz, compared to the 1340 MHz of the RX 580.



Unlike the RX 580, the new RX 590 only comes with 8 GB of video memory (no 4 GB variant), and the card's memory setup is unchanged: 8 Gbps GDDR5 over a 256-bit wide memory interface, which yields 256 GB/s of bandwidth. The "Polaris 30" silicon features the same core-configuration as its predecessors, with 2,304 stream processors spread across 36 compute units, 144 TMUs, and 32 ROPs. There's still no ray-tracing machinery to rival RTX, or other new features.

The Radeon RX 590 launch is completely board-partner driven, and there is no reference-design card. In this review, we're taking a look at the XFX Radeon RX 590 Fatboy. This card features a PCB that closely resembles AMD's reference design "Polaris" PCB, mated with an aluminium fin-stack cooler that's ventilated by a pair of fans. The card draws power from an 8-pin and a 6-pin PCIe power connector. There's also a factory overclock on tap, which has the XFX RX 590 Fatboy running at 1580 MHz out of the box. Memory clock is unchanged at 2000 MHz.

Architecture

At the heart of the Radeon RX 590 is the new "Polaris 30" silicon, which, as we mentioned earlier, is essentially the same chip that powers the RX 580 and RX 570, but is fabricated on the 12 nm FinFET process. The shrink to 12 nm improves power and thermal characteristics of the chip, allowing its designers to increase clock speeds. AMD missed the opportunity to make small but meaningful hardware changes, such as updating the display controllers with support for technologies such as VirtualLink, or at least updating the multimedia engine with hardware acceleration of newer video formats. The die size of the GPU is unchanged, which suggests that only a smaller transistor design is used, without any changes to the rest of the circuitry.


The architecture block diagram for the "Polaris 30" is hence identical to "Polaris 10" from 2016. A command processor distributes workloads between four Shader Engines, which distributes a geometry processor, rasterizer, and 8 ROPs among 9 compute-units, each. The compute-unit (CU) is the indivisible sub unit of the "Polaris" architecture and performs the bulk of the number crunching on the GPU. Each CU packs 64 stream processors, sharing them with tiny data caches, a scheduler, a scalar unit, and four TMUs.


The "Polaris 30" GPU has 36 compute units and a total of 2,304 stream processors, 144 TMUs, and 32 ROPs, all of which are numbers that are unchanged from the RX 580 and RX 480. The L2 cache is town-square for the GPU, which is addressed by not just the front-end, but the four Shader Engines, 256-bit wide GDDR5 memory interface, PCI-Express 3.0 x16 host interface, multimedia engine, and display controllers. One area where AMD's architecture excels NVIDIA's is with asynchronous compute, in which the GPU can divide graphics and compute workloads more granularly between its compute units. The "Polaris 30" silicon features two hardware schedulers dedicated to asynchronous compute.

Packaging and Contents

You will receive:

• Graphics card
• Documentation
• Molex to PCIe power adapter
• 6-pin to 8-pin PCIe power adapter
• Driver DVD

The Card

The XFX Radeon RX 590 Fatboy probably earned its name from being shorter in length than most cards in its class, while being taller and thicker. It uses a dual-fan cooling solution. A pair of 80 mm fans guide air through an aluminium fin-stack heatsink. The cooler is longer than the PCB underneath it. The backplate extends into the tail-end. Dimensions of the card are 27.0 x 13.0 cm.


Installation requires three slots in your system.


Display connectivity options include three DisplayPort 1.4, a HDMI 2.0b, and a dual-link DVI-D. The DVI connector has no analog wiring, so D-Sub dongles won't work and an active adapter has to be used.


The board uses a combination of 6-pin and 8-pin PCIe power connectors. This input configuration is specified for up to 300 watts of power draw.


XFX includes a dual-BIOS feature with their card, which will come in handy when it comes to recovering from a failed BIOS flash. Both BIOSes are identical.


The Radeon RX 590, like every other current AMD GPU, supports up to 4-way CrossFire X via PCIe.

Disassembly

Taking the XFX Radeon RX 590 Fatboy apart is straightforward because there are no exotic screws in your way. The cooler features an aluminium fin-stack heatsink with fins arranged such that air is guided along the front and back, rather than the sides. A copper base conducts heat from the GPU to four 6 mm thick copper heat pipes, which distribute heat across the fin stack. A secondary aluminium base pulls heat from the memory chips. Additional aluminium bases attached to the fin stack draw heat from the VRM MOSFETs.


As we mentioned earlier, the metal backplate extends into the rear-end of the card. There's an insulating sheath, although with no thermal pad to pull some heat from behind the GPU, which is a waste of so much metal.

On the next page, we dive deep into the PCB layout and VRM configuration.