AMD & NVIDIA GPU Silicon Performance, Efficiency & Cost Progress From The Last 10 Years Visualized

July 4, 2021 OSTN Staff

10 years of GPU Silicon advancements from AMD and NVIDIA have been visualized in a series of charts made by Reddit user, Vito_ponfe_Andariel. The charts show all the progress that has been achieved by the two main GPU makers in the industry starting and each generation of GPU shows both, its strength and weaknesses.

AMD & NVIDIA GPU Silicon Progress From The Past 10 Years Visualized

The data was gathered from the Techpowerup GPU database and the top parts were used to measure the performance, efficiency & cost advancements witnessed over several generations. The charts start off with the VLIW5 based AMD Radeon HD 5000 & the Fermi-based GeForce GTX 400 series graphics cards.

AMD Radeon HD 5000 vs NVIDIA GeForce GTX 400 Series

So starting off with the first GPUs in comparison, the Terascale 2 (VLIW5) based Radeon HD 5000 and Fermi-based GeForce GTX 400 series. Both GPU families were similar in performance & price but Fermi consumed more power and hence its energy efficiency wasn’t on par with the AMD GPU lineup at the time. Radeon GPUs also offered higher transistor performance.

The main issue with Fermi was that the top-tier chips were suited for large-scale HPC workloads and as such, featured unnecessary hardware units which led to a hot and power-hungry design. This was rectified with the release of more mainstream options and the next lineup in the GeForce family.

AMD Radeon HD 6000 vs NVIDIA GeForce GTX 500 Series

Moving on, we have the last iteration of Terascale and VLIW architecture as a whole from AMD, the Radeon HD 6000 series versus the second incarnation of Fermi, the GeForce GTX 500 series. Both GPU families saw a little gain in performance as they were widely regarded as optimized versions of what came before them but NVIDIA made some wise decisions in optimizing its broken Fermi GPU architecture for gaming and ended up delivering a slightly better product in general.

AMD did cut the costs of its Radeon HD 6000 series which were widely affordable for the masses but this generation was mostly regarded as an intermediate launch before the next generation arrived from both parties.

AMD Radeon HD 7000 vs NVIDIA GeForce GTX 600 Series

The GCN and Kepler cards were both based on a 28nm architecture and accelerated into the next gen. AMD launched its Radeon HD 7970 first which was well received with great performance and pricing while the same happened with NVIDIA which brought a range of new features and offered a huge increase in efficiency over the GCN architecture.

The Kepler architecture could be seen as the start of NVIDIA’s efficiency journey which would come into fruition in the coming generations.

AMD Radeon R200 vs NVIDIA GeForce GTX 700 Series

A year later, NVIDIA unveiled its GeForce GTX 780 Ti graphics card as a response to the AMD R9 200 series which was aimed at the enthusiast segment. AMD Hawaii cards were both a step forward & a step back, offering huge performance increases to match NVIDIA’s Titan cards but also sacrificed overall efficiency.

The AMD Radeon R9 290X as a result was hot but a strong competitor to NVIDIA’s GeForce GTX 780 and Titan but came short of the GTX 780 Ti. With that said, both cards had similar pricing and while performance was close, it was NVIDIA that was offering higher efficiency again.

AMD Radeon R300 vs NVIDIA GeForce GTX 900 Series

The NVIDIA GeForce GTX 900 series based on the Maxwell architecture was the biggest uplift in performance and efficiency the green team had delivered yet. The Maxwell lineup really pushed the energy that came from NVIDIA, but it was still just a teaser for what was to come.

AMD didn’t learn from their Hawaii lineup and moved to offer an even more power-hungry Fiji chip which was not only limited in terms of memory but also featured a pricing similar to the GTX 980 Ti while offering performance lower than that. AMD did become the first vendor to offer HBM memory on its consumer chips but that was the only good thing about the card aside from the liquid cooling variation. The GTX 980 Ti launched, first, delivered a massive blow to AMD in the high-end market and things started looking grim for AMD’s next-gen architecture.

AMD Radeon RX Vega vs NVIDIA GeForce GTX 10 Series

Everyone had huge hopes for AMD’s Radeon RX Vega architecture which was designed by the new head of RTG, Raja Koduri. But NVIDIA also wasn’t keeping silent. The company launched its first 10 series cards months ahead of AMD’s Vega lineup which pushed the efficiency & performance chart even further.

But it was the GTX 1080 Ti that truly crushed AMD’s high-end GPU efforts for the next several years. The GeForce GTX 1080 Ti not only offered a stunning price point but its performance was the biggest leap forward for any Ti graphics card. One that remains unmatched to this day. The Vega 64 managed to offer performance close to the GTX 1080 but it was hot, loud, and consumed way more power than what AMD had promised.

AMD Radeon RX 5000 vs NVIDIA GeForce RTX 20 Series

Enter the David Wang era, the new head of RTG had different and more constructive plans for the red team. Focus on both efficiency and performance. AMD’s goal changed from offering the best value to the best in all categories. This was seen in the first iteration of the Navi architecture with the Radeon RX 5700 XT which managed to offer great mainstream performance at an attractive price.

NVIDIA on the other hand, knowing that they had the lead in the performance category, went on to expand upon their feature set with DLSS and raytracing technologies. The Turing generation saw marginal gains over the last generation but the prices saw a huge increase due to increased wafer costs and the addition of premium GDDR6 memory modules.

AMD Radeon RX 6000 vs NVIDIA GeForce RTX 30 Series

We finally come to the current generation where NVIDIA tried its best to reach new performance heights with its Ampere architecture but AMD really blew them away by offering a GPU lineup that tackles NVIDIA across the board, even the flagship RTX 3090. The RTX 3080 and RX 6800 XT are close in pricing and while the former offers a great feature set & superior raytracing performance, the AMD Radeon RX 6800 XT leads in efficiency and even manages to come up with its own feature set such as FSR which was recently revealed. Both graphics cards are great in their own rights but AMD really outdid themselves with the 2nd iteration of Navi and things are looking great with the next-generation RDNA 3 family from what we have heard in rumors.

Of course, this is just a small analysis of all the generations of AMD and NVIDIA GPUs that have come in the last 10 years and we look forward to what innovations and architectural upgrades are headed our way in the next 10 years.

Which GPU are you using inside your current PC?