Radeon 7900 XTX and XT review: Faster, hotter and cheaper than the RTX 4080

Zoom in / Radeon RX 7900 XTX’s triple-fan cooler.

Andrew Cunningham

Nvidia’s RTX 4080 And 4090 GPUs are amazing performers. Starting at $1,200 and $1,500, they are quite expensive the way Compatible with cards from partners such as MSI, Gigabyte and Asus. The 4080 is more than double the original $699 MSRP for the RTX 3080.

These price spikes are due in part to pandemic-era concerns like supply chain snarls and inflation, and partly to the cryptocurrency-fueled boom (which is now over, blessedly). And there was a lack of competition in the game Increasing cost and complexity Creating giant, monolithic chips in sophisticated manufacturing processes. Today, AMD is attempting to address the latter two issues with the release of its Radeon RX 7900 series of GPUs.

At $899 and $999, the RX 7900 XT and RX 7900 XTX are still objectively expensive—but since they don’t exceed the RX 6900 XT’s starting price, both cards offer a bargain in today’s GPU market. If you’re looking for cards that can consistently handle 4K gaming at 60 fps and above, these GPUs do it for less than Nvidia’s latest, and because they’re good enough and fast enough, they’ll start to undercut Nvidia’s prices a bit. , too.

But Nvidia still holds some key advantages that can complicate an easy David and Goliath story. These GPUs don’t feel like a raison d’être for AMD’s graphics division—a turning point that a scrappy AMD could take a big hit on the market share of an entrenched, complacent competitor. But if you can actually find them for their starting price, they’ll be the first sign that some relief is on the way for high-end-but-price-conscious PC gamers.

Great Navi

AMD's chiplet design is visible in this shot—one large central die with most of the compute resources, and six smaller dies that house the cache and memory controllers.
Zoom in / AMD’s chiplet design is visible in this shot—one large central die with most of the compute resources, and six smaller dies that house the cache and memory controllers.

AMD

The RX 7000 series is the third version of the RDNA GPU architecture, sometimes referred to as “Navi” after the codenames of the GPU chips. RDNA 3 doesn’t add anything that feels as significant as RDNA 2’s ray-tracing support, but AMD has added plenty of additional hardware and made important changes.

Most notable is the new chiplet-based approach, which is similar in concept AMD uses it for its Ryzen CPUs. Instead of building the entire GPU die in a single manufacturing process—increasing the footprint and therefore the chance that some or all of them are defective—AMD uses the 5 nm TSMC manufacturing process and builds a series of core Navi 3 GPU dies. 6 nm process miniature memory controller dies (MCDs). All of these chips are connected together with a high-speed interconnect that AMD says can transfer data at speeds of up to 5.3 terabytes per second.

The main graphics compute die (GCT) contains most of the hardware you think of when you think of GPU-computing units, shaders, ray-tracing hardware, media encoding and decoding blocks, and display output. Both the 7900 XTX and XT use the same Navi 31 GCD, but the XTX runs at higher clocks and has more CUs and stream processors enabled. XTX has 96 CUs and 6,144 stream processors, XT has 84 CUs and 5,376 stream processors. Both cards represent a step up from the Navi 21 die used in the RX 6900 series, which has a maximum of 80 CUs and 5,120 stream processors (and that’s before you account for performance-boosting improvements).

RDNA 3 brings a little more to AMD's high-end GPUs, from CU count to memory bandwidth.
Zoom in / RDNA 3 brings a little more to AMD’s high-end GPUs, from CU count to memory bandwidth.

AMD

The MCDs all include a single 64-bit memory controller and 16MB of AMD’s Infinity cache, and demonstrate the advantages of a chiplet-based approach. The 7900 XTX has a 384-bit sized memory bus and 96MB of infinity cache, where the 7900 XT has a 320-bit bus and 80MB of cache; To do this, all AMD has to do is remove an MCD. Identical MCDs can be reused up and down the stack across all of the different RDNA 3 GCDs that AMD chooses to release, from low-end products with a single MCD to midrange GPUs that use two to four. Defects in MCD dies do not require casting or pinning larger, more complex GCDs.

In the realm of all-new features for RDNA 3, there are three things worth noting. First, there are new AI accelerators on GPUs that are useful for both AI-assisted content creation and AI-assisted development that have sprouted up in the past year. FSR upscaling algorithm; Both DLSS and XESS use AI for optimization, but FSR 2.0 does not).

Second, the video encoding and decoding block supports hardware-accelerated encoding for the AV1 video codec, similar to the RTX 4000 series and Intel’s Arc GPUs. This is useful for both content creators and streamers who want to stream high-resolution video or video at the same resolution while using less bandwidth.

Third, the “Radiance Display Engine” adds DisplayPort 2.1 support for GPUs. Monitors that take full advantage of DisplayPort 2.1’s additional bandwidth aren’t actually covered in this article, but RDNA 3 GPUs can drive 4K displays up to 480 Hz and 8K displays up to 165 Hz. .

Leave a Reply

Your email address will not be published.