NVIDIA GeForce GTX 980 Ti vs NVIDIA GeForce RTX 3090 Comparison

The NVIDIA GeForce GTX 980 Ti and the NVIDIA GeForce RTX 3090 represent two distinct eras of GPU design, separated by five years of architectural evolution. The data positions the RTX 3090 as the clear performance victor, winning all three head-to-head benchmark comparisons, yet both cards occupy the same 73rd percentile among all GPUs. This parity in percentile is a curious data point, suggesting that while the RTX 3090 is far faster in absolute terms, the benchmark database’s aggregate scoring places them in a similar tier of overall capability. The GTX 980 Ti, a Maxwell 2.0 part from 2015, holds its own in legacy and compute-focused workloads, while the RTX 3090, built on Ampere, dominates in modern DirectX 12 and OpenCL scenarios. The data implies that the choice between them is not about raw speed alone, but about which workload generation you are targeting.

The Verdict

For any user prioritizing modern gaming or compute performance, the data unambiguously points to the RTX 3090. In the 3DMark Steel Nomad DX12 test, the RTX 3090 scores 5118 against the GTX 980 Ti’s 1321, a delta of -74.2% for the older card. This is not a marginal improvement; it is a generational leap. The RTX 3090 also crushes the GTX 980 Ti in Geekbench OpenCL, posting 172758 versus 43513, a -74.8% delta. If your primary concern is current DirectX 12 titles or OpenCL-based compute tasks, the RTX 3090 is the only rational choice from this data.

However, the GTX 980 Ti is not without a niche. In Geekbench Vulkan, the gap narrows dramatically. The RTX 3090 scores 53927, while the GTX 980 Ti scores 47724, a delta of only -11.5%. This suggests that for Vulkan-based workloads, the older card remains surprisingly competitive, potentially due to lower driver overhead or architectural efficiencies in that API. The data shows the GTX 980 Ti’s average benchmark score is 28020, slightly higher than the RTX 3090’s 27565, despite losing every head-to-head test. This counterintuitive result implies that the GTX 980 Ti performs relatively better across a broader range of older or less demanding tests that are not captured in the head-to-head list.

The verdict from the data is clear: the RTX 3090 is the superior card for those who need maximum performance in modern, demanding workloads. The GTX 980 Ti is a viable option only for users heavily invested in Vulkan-based applications or those who rely on a wide spectrum of legacy benchmarks, where its aggregate score slightly edges out the newer card. The RTX 3090’s 24 GB of GDDR6X memory and 936.2 GB/s bandwidth dwarf the GTX 980 Ti’s 6 GB GDDR5 and 336.6 GB/s, making the newer card the obvious choice for high-resolution textures and large datasets. The data does not support choosing the GTX 980 Ti for any modern AAA gaming scenario.

FAQ

Q: Which card wins the 3DMark Steel Nomad DX12 benchmark?

A: The NVIDIA GeForce RTX 3090 wins with a score of 5118, compared to the GTX 980 Ti’s 1321. This represents a -74.2% delta for the GTX 980 Ti, indicating the RTX 3090 is significantly faster in this DirectX 12 test.

Q: Is the GTX 980 Ti competitive in any modern benchmark against the RTX 3090?

A: Yes, in the Geekbench Vulkan test, the gap is much smaller. The RTX 3090 scores 53927, while the GTX 980 Ti scores 47724. The delta is only -11.5%, suggesting the older card is reasonably close in Vulkan performance.

Q: How do the average benchmark scores compare between the two cards?

A: The GTX 980 Ti has a higher average benchmark score of 28020, while the RTX 3090 has an average of 27565. Despite losing all head-to-head tests, the GTX 980 Ti’s aggregate score is slightly better, indicating it performs well across a broader set of benchmarks.

Q: What are the memory specifications for each card?

A: The GTX 980 Ti has 6 GB of GDDR5 memory on a 384-bit bus with 336.6 GB/s bandwidth. The RTX 3090 has 24 GB of GDDR6X memory on a 384-bit bus with 936.2 GB/s bandwidth.

Q: Do both cards have the same percentile ranking among all GPUs?

A: Yes, both the GTX 980 Ti and the RTX 3090 are ranked in the 73rd percentile of all GPUs. This is notable because it shows they are in the same overall tier in the database, despite the RTX 3090’s superior raw performance in modern tests.

Q: Which card has more shading units and texture mapping units?

A: The RTX 3090 has 10496 shading units and 328 TMUs. The GTX 980 Ti has 2816 shading units and 176 TMUs. The RTX 3090 also has 82 RT cores and 328 tensor cores, which the GTX 980 Ti lacks entirely.

Architecture Differences

The GTX 980 Ti is built on the Maxwell 2.0 architecture, using the GM200 chip manufactured on a 28 nm process at TSMC. It contains 8,000 million transistors on a 601 mm² die, resulting in a transistor density of 13.3M per mm². In contrast, the RTX 3090 uses the Ampere architecture with the GA102 chip, fabricated by Samsung on an 8 nm process. This newer chip packs 28,300 million transistors into a 628 mm² die, achieving a density of 45.1M per mm². The data shows a massive jump in transistor count and density, which directly enables the RTX 3090’s higher core counts and feature set.

The most significant architectural difference is the inclusion of dedicated hardware in the RTX 3090. It features 82 RT cores for ray tracing and 328 tensor cores for AI acceleration, neither of which exists in the GTX 980 Ti. This means the RTX 3090 can handle real-time ray tracing and DLSS-style workloads, while the GTX 980 Ti relies purely on traditional rasterization. The GTX 980 Ti’s FP32 performance is 6.060 TFLOPS, while the RTX 3090 reaches 35.58 TFLOPS, a nearly six-fold increase. The RTX 3090 also offers FP16 performance at 35.58 TFLOPS (1:1), while the GTX 980 Ti has no listed FP16 capability, indicating a fundamental difference in compute flexibility.

The process node shrink from 28 nm to 8 nm is not just about density; it enables higher clock speeds. The GTX 980 Ti has a base clock of 1000 MHz and a boost of 1076 MHz, while the RTX 3090 runs at 1395 MHz base and 1695 MHz boost. The memory architecture also diverges: the GTX 980 Ti uses GDDR5 at 7 Gbps effective, while the RTX 3090 uses GDDR6X at 19.5 Gbps effective. This contributes to the RTX 3090’s 936.2 GB/s bandwidth versus 336.6 GB/s for the GTX 980 Ti. The API support also differs, with the RTX 3090 supporting DirectX 12 Ultimate (12_2) versus DirectX 12 (12_1) on the GTX 980 Ti, though both support OpenGL 4.6 and Vulkan 1.4.

Specification Differences

The two cards differ across nearly every measurable specification. The GTX 980 Ti has 2816 shading units, 176 TMUs, and 96 ROPs, while the RTX 3090 has 10496 shading units, 328 TMUs, and 112 ROPs. The pixel rate is 103.3 GPixel/s for the GTX 980 Ti and 189.8 GPixel/s for the RTX 3090. Texture rate sees a similar jump, from 189.4 GTexel/s to 556.0 GTexel/s. Memory capacity is a major differentiator: 6 GB versus 24 GB, with bandwidth increasing from 336.6 GB/s to 936.2 GB/s, both on a 384-bit bus.

Power and physical requirements also differ significantly. The GTX 980 Ti has a TDP of 250 W and uses a 1x 6-pin plus 1x 8-pin power connector, with a suggested PSU of 600 W. The RTX 3090 draws 350 W, uses a single 12-pin connector, and requires a 750 W PSU. The cards also differ in size: the GTX 980 Ti is 267 mm long, 111 mm tall, and 40 mm wide, occupying a dual-slot. The RTX 3090 is 336 mm long, 140 mm tall, and 61 mm wide, taking up a triple-slot. The bus interface advances from PCIe 3.0 x16 to PCIe 4.0 x16. Display outputs change from 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2 on the GTX 980 Ti to 1x HDMI 2.1 and 3x DisplayPort 1.4a on the RTX 3090. The launch MSRP for the GTX 980 Ti was 649 USD, while the RTX 3090 launched at 1,499 USD.

Head-to-Head Benchmarks

The head-to-head data shows a decisive victory for the RTX 3090 across all three tests. In 3DMark Steel Nomad DX12, the RTX 3090 scores 5118 versus the GTX 980 Ti’s 1321. This is a -74.2% delta for the GTX 980 Ti, meaning the RTX 3090 is roughly 3.9 times faster in this modern DirectX 12 workload. This test likely stresses the RTX 3090’s higher shading unit count, faster clocks, and larger memory bandwidth, all of which are significantly superior.

In Geekbench OpenCL, the RTX 3090 again dominates with a score of 172758 against 43513 for the GTX 980 Ti, a -74.8% delta. OpenCL is a compute-heavy workload that benefits from the RTX 3090’s massive FP32 throughput of 35.58 TFLOPS and its 328 tensor cores, even if they are not directly used for this API. The GTX 980 Ti’s 6.060 TFLOPS is simply outclassed.

The closest contest is in Geekbench Vulkan. The RTX 3090 wins with 53927 points, but the GTX 980 Ti trails by just 6203 points with a score of 47724, a -11.5% delta. This is interesting because Vulkan is a low-overhead API that can sometimes favor older architectures with simpler driver stacks. The GTX 980 Ti’s Maxwell architecture, despite being older, appears to handle Vulkan tasks relatively efficiently. The data shows that while the RTX 3090 is faster, the GTX 980 Ti is not embarrassed in this specific test, making it a potential consideration for Vulkan-centric workloads.

Where Each One Wins

The RTX 3090 wins decisively in modern DirectX 12 and OpenCL compute workloads. The 3DMark Steel Nomad result and the Geekbench OpenCL score are absolute blowouts, with deltas exceeding 74% in favor of the RTX 3090. For users running contemporary AAA games that use DirectX 12, or engaging in GPU compute tasks like rendering, machine learning inference, or scientific simulations via OpenCL, the RTX 3090 is the clear winner. Its 24 GB of GDDR6X memory also makes it far more suitable for large datasets, high-resolution textures, and multi-tasking workloads that would exhaust the GTX 980 Ti’s 6 GB frame buffer.

The GTX 980 Ti’s only notable win is in the aggregate average benchmark score. With an average of 28020 versus the RTX 3090’s 27565, the GTX 980 Ti scores 455 points higher across all benchmarks in the database. This suggests that in a wide range of legacy or less demanding tests—likely older DirectX 9, 10, and 11 titles, or 2D workloads—the GTX 980 Ti remains very capable. Its lower power draw of 250 W and smaller physical footprint (267 mm versus 336 mm) also make it easier to integrate into older or more compact systems. The data implies that for users with a library of older games or who value compatibility with a broader range of software, the GTX 980 Ti offers a balanced profile that the RTX 3090 does not match in aggregate. The Vulkan result, while a loss, shows the GTX 980 Ti is only 11.5% behind, making it a reasonable choice for Vulkan-based indie titles or emulators where that API is common.