Gaming PC

AMD Boasts About RDNA 3 Efficiency as RTX 40-Series Looms

According to AMD’s latest blog, Better performance per watt (opens in new tab) Across the past few generations of Radeon GPUs. AMD’s Sam Naffziger, who spoke to us about RDNA 3’s chiplet design and efficiency gains back in June, said AMD made some “big bets” on GPU design and efficiency and is now very He argues that current and future AMD GPUs are meeting the needs of demanding gaming and doing so without a very high power budget. I’m here. This is, if necessary, a timely reminder that energy prices are skyrocketing and Nvidia plans to launch a very expensive graphics card at his GeForce Beyond event tomorrow.

AMD SVP, Corporate Fellow and Product Technology Architect Sam Naffzinger says GPUs are starting to push beyond 400W as PC gamers demand better and faster visuals at ever higher resolutions. is not surprising. Naffzinger says AMD has tweaked his new RDNA architecture to break away from the ever-increasing power consumption cycle using “thoughtful and innovative engineering and refinements.” Improved energy efficiency means lower-power graphics cards run cooler and quieter, and are more readily available in small or mobile form factors.

AMD has provided some evidence of progress with its RDNA GPU architecture, a series that will launch its 3rd generation later this year. For example, according to Naffzinger’s stats, moving from GCN to his RDNA “has improved performance per watt by an average of 50%.” Moving from the Radeon RX 5000 (RDNA) generation to the Radeon RX 6000 (RDNA 2) generation, AMD claims “a whopping 65% increase in performance per watt.”

To spice up the comparison, AMD provided an interesting chart showing the “competitive advantage” of RDNA 2 and Ampere. Obviously, these numbers were handpicked to make AMD look good, and note the word “maximum.” Check out our GPU benchmark hierarchy and our list of the best graphics cards to independently see how things stack up.

(Image credit: AMD)

This slide is definitely misleading at best. If you ignore the table and just look at the headline, it seems like AMD’s value has improved significantly. However, that comparison uses a recent Radeon RX 6400, compared to a six-year-old GTX 1050 Ti. It also happens to be more expensive than the new GTX 1650. Hmm. At least AMD’s GPU price list looks pretty accurate (aside from including the GTX 1050 Ti). And even omitting the ray-traced game, the performance isn’t far from our own results.

For example, the RX 6950 XT averaged 115.4 fps at 1440p ultra in our tests, while the RTX 3090 averaged 106.5 fps. This is in line with the 8% performance difference AMD showed in their slides, but we tested the reference RTX 3090 Founders Edition against an overclocked Sapphire RX 6950 XT Nitro+ Pure. A comparison of the RX 6600 and the RTX 3050. As we recently pointed out, the RTX 2060 is Nvidia’s best value option, being both faster and cheaper than the RTX 3050. But let’s move on.

AMD’s blog explains that RDNA 2’s advancement over RDNA is primarily due to optimized switching, high frequency design, and smarter power management. Now let’s talk about how we’re moving forward with RDNA 3. Key technology improvements and introductions in RDNA 3 include a 5nm process, chiplet package technology, adaptive power management, and next-generation Infinity cache.

Related Articles

Back to top button