Radeon RX 9070 XT Gaming OC: Gigabyte’s affordable choice

The AMD Radeon RX 9070 XT delivers better price-to-performance than competing GeForce cards and 4 GB more memory than the RTX 5070. The Radeon RX 9070 XT Gaming OC 16G is the more affordable of Gigabyte’s two RX 9070 XT models. While it doesn’t match premium variants in overall performance, it offers more capable power delivery, a higher power limit for factory overclocking, and greater overclocking headroom than base models.

We introduced the Radeon RX 9070 XT and RX 9070 in articles dedicated to the RDNA4 architecture, the specifications of both models, their features, and the new technologies they bring.

After testing the Sapphire Pulse AMD Radeon RX 9070 XT, we’ll now look at the cheaper of the competing models from Gigabyte—the Radeon RX 9070 XT Gaming OC.

Before diving into the more significantly overclocked model, let’s recap the basic parameters of both AMD cards launched last week.

Gigabyte offers two Radeon RX 9070 XT models—the premium AORUS Radeon RX 9070 XT Elite 16G and the smaller, cheaper Radeon RX 9070 XT Gaming OC 16G. Both are factory-overclocked.

The Gigabyte card (in front) is slightly smaller than the Pulse RX 9070 XT or Pure RX 9070 we tested.

Even the cheaper model doesn’t lack a dual BIOS, which again only differs in cooling regulation settings. Below are the parameters stored in the Performance BIOS. The boost clock is increased from the reference 2970 MHz to 3060 MHz. The advertised typical game clock is raised from 2400 MHz to 2520 MHz. In practice, the values are even higher, as you’ll see in the sections on operating characteristics, which include graphs of clock speed behavior.

Better cooling means lower power consumption and slightly more room for automatic overclocking to higher speeds. However, these are usually just tens of MHz, so the performance drop with the quiet BIOS is barely measurable and unnoticeable to the eye.

With the increased clock speeds, the power limit (TBP) has also been raised from the reference 304 W to 330 W.

To the values from GPU-Z monitoring, I’ll add data from HWiNFO, which provides plenty of details. The power limit can be increased by another ten percent.


The second set of images is from the Silent BIOS, and aside from its version, there’s no difference in the displayed data.

However, you’ll definitely hear that the card is noisier. Therefore, for performance tests, we’ll use the slightly slower, quieter BIOS, and we’ll measure operating characteristics on both BIOS versions.

 

On the back of the packaging, you’ll find the usual presentation of Gigabyte’s features and technologies utilized by the card. These include the Windforce cooling system with ARGB lighting, a reinforced metal backplate for card rigidity and PCB protection, a new generation of proprietary fans, and a cooler with a vapor chamber base that cools both the chip and memory. The last image highlights the freely accessible part of the heatsink, allowing air heated by the cooler to flow freely over the card.

In addition to the card, the package includes only a brief installation manual.


AMD has started mass production of Zen 6. Only for Epycs for now?

It’s been nearly two years since AMD’s Zen 5 architecture launched (in late July and early August 2024) and time is ripe for a successor—especially given how far competitors such as Qualcomm’s Snapdragon X2 Elite and Apple’s M5 have advanced in the meantime. AMD’s answer in the form of the Zen 6 architecture is now finally getting closer—the company has announced the ramp up of production, which could signal a launch later this year. Read more “AMD has started mass production of Zen 6. Only for Epycs for now?” »

Advanced Shader Delivery on AMD: No more waiting for compilation

Gamers definitely know the experience: You want to launch a game, but it starts to compile shaders, which can take several minutes. The reason is that shaders have to be built from code specifically for your hardware. And the compilation must be repeated after every GPU driver update, so you can end up seeing this waiting screen quite frequently. But Windows is now getting an improvement that should largely eliminate this waiting. Read more “Advanced Shader Delivery on AMD: No more waiting for compilation” »

FSR 4.1 AI upscaling finally coming to older Radeon GPU users

When AMD launched the FSR4 AI-based upscaling technology for games last year, it was exclusively available for the new Radeon RX 9000 generation GPUs using the RDNA 4 architecture. This was despite the fact that a version using INT8 compute compatible with older GPUs had leaked out, apparently by accident. But owners of older Radeon cards are finally in luck now: FSR 4.1 is coming to RDNA 2 and RDNA 3 as well. Read more “FSR 4.1 AI upscaling finally coming to older Radeon GPU users” »

Comments (2) Add comment

  1. I simply love your review format, now back to these cards…

    In this day and age is shameful that (both Ngreedia and Amd) charge this much for this level of performance but here we are. I won’t buy one until the price comes down a lot.
    Really makes me think about switching to a console or a Steam deck if they keep it up.
    In my opinion you really need to introduce an UE5 game in your reviews just to see how bad performance is in those.
    Robocop Rogue City, Stalker 2, Black myth Wukong etc.

    Thanks for your work, keep it up.

    1. Thanks. I am already testing Black Myth: Wukong and Satisfactory, I just need to retest some cards to add charts. More games will follow as soon as I will have time to play them for a few hours and find optimal locations for testing.

      I am able to test Stalker 2, Robocop, Hogwarts Legacy, Frostpunk 2, Fornite, God of War Ragnarok, Gray Zone: Warfare and some other games, the only thing I don’t have is the time to play it and prepare for testing. :(.

      When it comes to performance, I assume that developers already rely on TSR/FSR/DLSS and this shift is now irreversible.

Leave a Reply

Your email address will not be published. Required fields are marked *