Introduction

AMD finally has their next generation graphics cards in the performance segment! The Radeon RX 7800 XT and RX 7700 XT being launched today fill a vast performance gap between the enthusiast-class RX 7900 series meant for maxed out 4K gaming, and the mainstream RX 7600 designed for 1080p. Both new cards that we review today are meant for the 1440p resolution, which is registering strong growth on the Steam Hardware Survey graphs. The RX 7800 XT should give you significantly higher performance than the RX 7700 XT for a price that's surprisingly not much higher. Across the competitive landscape, AMD considers the RX 7800 XT to be a competitor to the GeForce RTX 4070. The Red team is pricing its new card aggressively, at just $500, which undercuts the RTX 4070 by $100. To sweeten the deal even more, AMD is including the space opera RPG Starfield with this card.



The Radeon RX 7800 XT is based on the same RDNA 3 graphics architecture as the rest of the RX 7000 series, and takes advantage of the 5 nm EUV foundry node where it really matters—the compute units and other graphics rendering machinery. The card debuts the new Navi 32 silicon, which is a chiplet-based GPU, just like Navi 31 that powers the RX 7900 series. AMD turned to chiplets to maximize its foundry utilization of the more advanced 5 nm foundry node. It did so, by identifying the specific components of the GPU that benefit the most from the newer node—essentially all the number-crunching machinery—and those components that don't benefit as much, such as the Infinity Cache memory and the GDDR6 memory controllers. These are spun out as memory cache dies (MCDs) built on the older 6 nm node. There are four of these on Navi 32, compared to six on Navi 31, each with a 16 MB segment of the chip's 64 MB Infinity Cache, and a 64-bit wide portion of the GPU's 256-bit GDDR6 memory interface.

The Radeon RX 7800 XT is designed to max out Navi 32, enabling all 60 RDNA 3 compute units physically present on its GCD (graphics compute die). This works out to 120 AI Accelerators, 60 Ray Accelerators, 240 TMUs, 96 ROPs, and a 256-bit wide GDDR6 memory, which AMD is using to drive a generous 16 GB of memory. AMD is running their GPU at 2124 MHz Game clock, and 2430 MHz boost, while the memory ticks at 19.5 Gbps for an impressive 624 GB/s of memory bandwidth. AMD considers the RX 7800 XT a spiritual successor to the RX 6800 XT that the company released as an enthusiast-class product the last time around, pricing it accordingly.

The RDNA 3 graphics architecture powering the RX 7800 XT introduces new dual-issue rate compute units, with vastly improved idle SIMD resource utilization, and support for new math formats. The company claims generational IPC uplifts over the RDNA2 compute units to the tune of 17%. Taking advantage of the newer 5 nm node, AMD is also running the shader engines at higher clock speeds than the previous generation. The company also introduced Ray Accelerator, a component inside the CU that accelerates matrix math functions by utilizing the SIMD units, which should accelerate AI DNN building and training. The 2nd generation Ray Accelerator is designed to provide a 50% generational improvement in ray intersection performance. Another interesting change introduced with RDNA 3 is that the GPU's front-end operates at a 10-15% higher clock speed than the shader engines. At its given specs, the RX 7800 XT has a total board power of 263 W, and can make do with two conventional 8-pin PCIe power connectors.

AMD has a reference design graphics card based on the RX 7800 XT which we are reviewing here, there isn't a reference design board for the RX 7700 XT. The reference made-by-AMD (MBA) board uses the same design language as the reference RX 7900 XT. The triple-slot card for the most part looks like a smaller reference RX 7900 XT, but with two fans instead of three. A dense aluminium fin-stack heatsink is ventilated by two axial-flow fans. AMD is pricing the reference Radeon RX 7800 XT at $500. The card will be available on AMD's first-party online store, as well as through the company's board partners, with minimal re-branding.

Besides the AMD Radeon RX 7800 XT reference, we have several other custom-design card reviews, as well as reviews for the custom RX 7700 XT.

Short 10-Minute Video Comparing 9x RX 7800 XT and RX 7700 XT

Our goal with the videos is to create short summaries, not go into all the details and test results, which can be found in our written reviews.

Architecture

The Radeon RX 7800 XT and RX 7700 XT debut the new Navi 32 silicon, as part of AMD's desktop-first approach to the RX 7000 series. This is a chiplet-based GPU, just like the Navi 31 which powers the RX 7900 series, but scaled down. The centrally located graphics compute die (GCD) is based on the newer 5 nm EUV foundry process, and contains the main number crunching machinery of the GPU, all the stuff that tangibly benefits from the switch to the newer node, while the Infinity Cache and GDDR6 memory controllers—things that don't benefit as much from 5 nm, are spun off as chiplets called memory cache dies. Navi 32 gets four such MCDs compared to six on Navi 31. Each of these has a 16 MB segment of the GPU's 64 MB Infinity Cache, and a 64-bit portion of that adds up to its 256-bit GDDR6 memory interface.

The 5 nm GCD features 60 RDNA 3 compute units, each worth 64 stream processors, which works out to 3,840. This is a 50% generational increase in shaders compared to Navi 22 powering the Radeon RX 6700 series, but is a reduction when compared to the RX 6800 XT that's equipped with 72 RDNA2 compute units. AMD is counting on the increased IPC, higher engine clocks, and greater memory bandwidth of Navi 32 to come through and make the RX 7800 XT faster than its predecessor; although given its pricing, AMD considers the RX 7800 XT to succeed the RX 5700 XT, and prove to be a viable upgrade for all those who held out on the RX 6000 series over the last three years.

The Radeon RX 7800 XT maxes out the Navi 32 silicon, enabling all 60 CUs, which work out to 3,840 stream processors, 120 AI accelerators, 60 Ray accelerators, 240 TMUs, and 96 ROPs. With all four MCDs enabled, the RX 7800 XT gets 64 MB of Infinity Cache, and a 256-bit GDDR6 memory interface, which it puts to good use with 16 GB of memory—same amount as the RX 6800 XT. The GPU runs at 2124 MHz Game clock, with 2420 MHz boost, however within this band internally, the front-end of the GPU runs at 10-15% higher clock speeds than the shader engines. The memory runs at 19.5 Gbps, which gives the RX 7800 XT a handy 628 GB/s of memory bandwidth.

The Radeon RX 7700 XT is cut down from the Navi 32 silicon, enabling 54 out of 60 CUs, giving it 3,456 stream processors, 108 AI accelerators, 54 Ray accelerators, 216 TMUs, and 96 ROPs. This SKU has three out of four MCDs enabled, which gives it 48 MB of Infinity Cache, and a 192-bit GDDR6 memory interface that is uses for its 12 GB of memory size. AMD runs the GPU at higher clock speeds than the RX 7800 XT, with 2171 MHz game clocks, and 2544 MHz boost. The memory is slightly slower, at 18 Gbps, giving it 432 GB/s of memory bandwidth.


Much of the architectural innovation is this generation is with the RDNA 3 Dual-Compute Unit (or Compute Unit pair). The "Navi 32" GPU physically features 60 compute units spread across three Shader Engines. AMD claims that at the same engine clocks, the RDNA 3 CU offers a 17.4% IPC increase over the RDNA2 CU.


The new RDNA 3 CU introduces multi-precision capability for the 64 stream processors per CU: operating either as 1x SIMD64 or 2x SIMD32 units. The Vector Unit that houses these SIMD units can either function as a SIMD execution mechanism, or as a Matrix execution unit, thanks to the new AI Matrix Accelerator, which provides a 2.7x matrix multiplication performance uplift versus conventional SIMD execution. Also added are support for the Bfloat16 instruction-set, and SIMD8 execution. The GPU hence enjoys AI hardware-acceleration that can be leveraged in future feature-additions relevant to gamers. Game developers will also look for ways to exploit accelerated AI, now that all three brands feature it (NVIDIA Tensor cores and Intel XMX cores).


AMD's first-generation Ray Accelerator, introduced with the RDNA2 architecture, was the result of a hasty effort to catch up to NVIDIA with a DirectX 12 Ultimate GPU, where they developed fixed-function hardware to calculate ray intersections, and offloaded a large chunk of RT processing to the generationally-doubled SIMD resources. With RDNA 3, they've refined the Ray Accelerator to achieve an 80% ray tracing performance uplift over the previous generation, when you add up the Ray Accelerator count, their higher engine clocks, and other hardware-level optimizations, such as early sub-tree culling, specialized box sorting modes, and reduced traversal iterations.


There is a 50% ray intersection capacity improvement for RDNA 3 thanks to these optimizations, and cycles-per-ray reduction. Besides these, AMD has also made several improvements to the geometry- and pixel-pipes, with the introduction of the new multi-draw indirect accelerator (MDIA), which reduces CPU API and driver-level overheads by gathering and parsing of multi-draw command data. At the hardware-level 12 primitives per clock is now supported compared to 8 per clock on RDNA2, thanks to culling. The core-configuration overall enables 50% more rasterized performance per clock.


AMD has significantly improved the Display Engine of "Navi 32" over the previous-generation in terms of connectivity. The new Radiance Display Engine comes with native support for DisplayPort 2.1, which enables 8K output at up to 165 Hz refresh-rate, or 4K at up to 480 Hz, with a single cable. AMD has refined its FSR 2 algorithm to support 8K (i.e. render at a lower resolution with FSR-enhanced upscaling), to make it possible to enjoy the latest AAA titles at playable frame-rates on 8K displays. The RX 7800/7700 XT gets two full-size DP 2.1 connectors, besides an HDMI 2.1b, and a USB-C with DP 1.2 passthrough. The "Navi 32" silicon receives full hardware-accelerated AV1 encode and decode capabilities. With this generation, AMD is also introducing SmartAccess Video, a feature that lets the AMD driver leverage the hardware encoders of the RDNA2 iGPU of Ryzen 7000 desktop processors, for additional encoding performance.

FidelityFX SuperResolution 3 Fluid Motion Frames (FSR 3 and FMF)

As part of the Radeon RX 7800 XT and RX 7700 XT announcements, AMD finally announced the much awaited FidelityFX Super Resolution 3 and Fluid Motion Frames. FSR 3 is being announced as a technological rival to NVIDIA DLSS 3 Frame Generation. The premise with both technologies is the same—to effectively double frame-rates by generating alternate frames without running them through the entire graphics rendering pipeline, it's just that the two technologies differ in their approach to this goal.

FSR 3 builds on FSR 2 with its updated super resolution upscaler promising generational quality improvements at a every given rendering resolution. Fluid Motion Frames (FMF) isn't the entirety of the FSR 3 feature-set, but is its most important feature-addition. FMF is a frame interpolation technology much like the one consumer televisions come with. Alternate frames are generated as an approximate of two frames. Where FMF differs from DLSS 3 Frame Generation is that while NVIDIA uses a hardware component called optical flow accelerator and the GPU's AI acceleration to generate an intermediate frame without involving the graphics rendering pipeline, FMF uses a certain amount of the graphics rendering pipeline. At a hardware level, FMF uses the main SIMD machinery of the GPU, leveraging asynchronous compute. As with DLSS 3 FG, FSR 3 FMF comes with added latency. NVIDIA counteracts this with Reflex, while AMD uses Radeon AntiLag+. Both technologies try to keep the frame queue short to reduce whole system latencies.

One major advantage FSR 3 FMF enjoys over DLSS 3 FG is that it works on any modern DirectX 12 GPU that supports async compute, since it doesn't require a specific hardware component the way DLSS 3 FG requires the Optical Flow Accelerator on NVIDIA "Ada" GPUs. The only limiting factor here is the performance. To be more specific, AMD says that all Radeon GPUs from RX 5700 series onward; and all GeForce GPUs from RTX 20-series onward, should support FSR 3 FMF. Also, FSR 3 FMF is as easy to integrate with games as FSR 2 is. The first games implementing FSR 3 FMF should arrive in Fall 2023. AMD is also working to extend FMF to Radeon Super Resolution, the driver-level technology that enables performance upscaling to even games that don't support FSR.

HYPR-RX

HYPR-RX is an interesting new feature AMD plans to integrate with the AMD Software (control center) application. It is a one-click performance boosting technology that works with any DirectX 11 or DirectX 12 game. The software is a cocktail of Radeon Boost, Radeon AntiLag+, and Radeon Super Resolution, and applies the three features on any running game as needed, automatically. Radeon Boost improves performance by dynamically reducing the render resolution of a game when there's too much motion on the screen (and hence not enough detail needed). Radeon Super Resolution improves frame-rates as it applies FSR on the output of a game rendered at a lower resolution (including the frames lowered in resolution by Radeon Boost. AntiLag+ counteracts the latency added by these two, by shortening the frame queue. AMD said that it is working to integrate FMF into the HYPR-RX feature-set.

Packaging