NVIDIA GeForce GTX 980 Ti

The NVIDIA GeForce GTX 980 Ti is a Maxwell 2.0 architecture GPU built on TSMC's 28 nm process, packing 8,000 million transistors on a 601 mm² die. It runs at a base clock of 1000 MHz and a boost of 1076 MHz, with 2816 shading units, 176 texture units, and 96 ROPs. The card carries 6 GB of GDDR5 memory on a 384-bit bus, yielding 336.6 GB/s of bandwidth. Its launch MSRP was 649 USD. Released on June 1, 2015, it belongs to the GeForce 900 generation, succeeding the GeForce 700 series and preceding the GeForce 10 series. The GPU is now end-of-life.

How It Compares

The GTX 980 Ti's average benchmark score is 27,956, placing it in the 71st percentile of all GPUs. Against the AMD Radeon RX 6600 XT, which averages 27,985, the 980 Ti trails by 0.1% — a negligible margin that puts the two effectively at parity. In real-world terms, this means the older Maxwell card trades blows with a much newer RDNA2 product, though the delta is so small it falls within typical run-to-run variance.

The AMD Radeon R9 M395X, a mobile GPU, averages 27,883. Here the 980 Ti leads by 0.3%. This is a narrow but consistent advantage, suggesting the desktop 980 Ti holds a slight edge over this high-end laptop part, despite the M395X being a later release.

Against the NVIDIA GeForce RTX 3070, the 980 Ti averages 27,956 versus 28,238 — a 1% deficit. The RTX 3070 is a much newer Ampere card, yet the 980 Ti's performance gap is surprisingly small, likely due to the 3070's lower memory bandwidth relative to its compute capabilities, or the specific benchmark mix. Still, the 1% difference means the 980 Ti is essentially one step behind in aggregate throughput.

The AMD FirePro S7150, a workstation-oriented GPU, scores 28,409 on average. The 980 Ti is 1.6% slower. This is the largest gap among the listed rivals, but it remains modest. The FirePro's higher average likely reflects its optimized drivers for professional compute workloads, whereas the 980 Ti targets gaming.

Ray Tracing and Feature Set

The GTX 980 Ti has no dedicated ray tracing cores and no tensor cores — both fields are null in the specification. This is expected for a Maxwell 2.0 GPU, which predates NVIDIA's RT and tensor hardware. Consequently, hardware-accelerated ray tracing is not supported. However, the card does support DirectX 12 with feature level 12_1, OpenGL 4.6, and Vulkan 1.4. These APIs allow for modern rendering techniques, including asynchronous compute and explicit multi-GPU, though without RT-specific extensions.

The display outputs comprise one DVI, one HDMI 2.0, and three DisplayPort 1.2 connectors, enabling multi-monitor setups. The card uses a PCIe 3.0 x16 interface, which remains compatible with contemporary motherboards. Its dual-slot design measures 267 mm in length, 111 mm in height, and 40 mm in width — a typical size for a high-end card of its era.

Benchmark Performance

The GTX 980 Ti's benchmark results show a consistent pattern. In 3DMark Steel Nomad (DX12), it scores 1,321. This test stresses modern features like mesh shaders and ray tracing, but the 980 Ti lacks RT hardware, so the score reflects its raw rasterization and compute capabilities. In Geekbench Metal, it achieves 19,520, indicating strong performance on Apple's graphics API. The OpenCL score of 43,259 and Vulkan score of 47,724 demonstrate solid compute throughput across cross-platform APIs.

The average benchmark score of 27,956 positions the card in the 71st percentile of all GPUs. Relative to its nearest rivals, the deltas are tight: -0.1% versus the RX 6600 XT, +0.3% versus the R9 M395X, -1% versus the RTX 3070, and -1.6% versus the FirePro S7150. These margins suggest that the 980 Ti, despite its age, remains competitive with a range of newer and older GPUs. The 1% gap to the RTX 3070 is particularly telling — it implies that in many gaming workloads, the 980 Ti can hold its own, though the 3070's ray tracing and DLSS features would tip the balance in supported titles.

The FP32 compute throughput is 6.060 TFLOPS, which is substantial for a 2015 GPU. Pixel and texture rates are 103.3 GPixel/s and 189.4 GTexel/s, respectively. These numbers indicate that the card is well balanced for traditional rendering, with enough fill rate to drive high-resolution displays.

Power and Cooling

The GTX 980 Ti has a TDP of 250 W. NVIDIA recommends a 600 W power supply, which is a common guideline for high-end cards of this generation. The card requires one 6-pin and one 8-pin PCIe power connector, totaling 14 pins. This is a standard configuration for a 250 W GPU. The dual-slot cooler is designed to dissipate that heat within a typical chassis. The card's length of 267 mm (10.5 inches) means it fits in most mid-tower cases, though users with compact builds should verify clearance. The 111 mm height and 40 mm width are within typical expansion slot limits.

FAQ

Q: Does the GTX 980 Ti support hardware ray tracing?

A: No. The card has no RT cores, so hardware-accelerated ray tracing is not available.

Q: What is the memory size and type?

A: It has 6 GB of GDDR5 memory on a 384-bit bus.

Q: What is the memory bandwidth?

A: The bandwidth is 336.6 GB/s.

Q: Which APIs does the card support?

A: It supports DirectX 12 (feature level 12_1), OpenGL 4.6, and Vulkan 1.4.

Q: What power supply is recommended?

A: NVIDIA suggests a 600 W PSU, and the card uses one 6-pin and one 8-pin power connector.

Q: How does it compare to the RTX 3070 in average score?

A: The 980 Ti averages 27,956, while the RTX 3070 averages 28,238, making the 980 Ti 1% slower.

Memory Subsystem

The GTX 980 Ti's memory subsystem is a defining feature. It uses 6 GB of GDDR5 on a 384-bit bus, delivering 336.6 GB/s of bandwidth. The memory clock is 1753 MHz, which translates to 7 Gbps effective. This configuration was top-tier in 2015 and remains relevant for high-resolution gaming. The 384-bit bus provides a wide data path, which is particularly beneficial for large texture datasets and high-resolution rendering. At 4K, the 6 GB capacity can hold a substantial amount of texture data, though newer games may exceed that limit at maximum settings. The bandwidth of 336.6 GB/s ensures that the shading units and ROPs are fed efficiently, reducing potential bottlenecks. The pixel rate of 103.3 GPixel/s and texture rate of 189.4 GTexel/s complement the memory throughput, allowing the card to maintain high frame rates in demanding scenes.

Who Should Consider It

The GTX 980 Ti's average score of 27,956 and 71st percentile ranking indicate it sits comfortably in the mid-to-upper tier of GPUs, even today. Its performance is within 1.6% of several modern and professional cards, making it a viable option for users who already own one or find it at a reasonable second-hand price. The 6 GB VRAM and 336.6 GB/s bandwidth make it suitable for high-resolution textures and multi-monitor setups. Its FP32 throughput of 6.060 TFLOPS ensures solid compute performance for non-gaming tasks like video editing or 3D rendering. However, the lack of ray tracing and tensor cores means it cannot accelerate those specific workloads. For gamers who prioritize traditional rasterization performance at 1080p or 1440p, the 980 Ti remains a capable choice, provided the power supply meets the 600 W recommendation. Those targeting 4K should consider the memory capacity and bandwidth as adequate for most titles, but may need to lower settings in the most demanding releases. The card's end-of-life status means no new driver optimizations are expected, but the existing driver set supports current APIs. In summary, the GTX 980 Ti is a balanced performer that holds its ground against much newer hardware, though it lacks the modern feature set of RTX and RDNA2 GPUs.