NVIDIA GeForce GTX 1080 Ti vs NVIDIA GeForce GTX 690 Comparison

The NVIDIA GeForce GTX 690 and the NVIDIA GeForce GTX 1080 Ti represent two distinct eras of NVIDIA’s flagship strategy, separated by nearly five years of architectural evolution. The GTX 690 is a dual-GPU Kepler monster from the GeForce 600 generation, while the GTX 1080 Ti is a single-chip Pascal powerhouse from the GeForce 10-series. Benchmark data shows a clear, dominant performance shift toward the newer card, but the older card still holds a specific place in the hardware landscape.

Where Each One Wins

The benchmark results are unambiguous: the GTX 1080 Ti wins every single head-to-head test recorded in the data. The GTX 690 does not win any of the two comparative benchmarks. In the Geekbench OpenCL test, the GTX 1080 Ti scores 67929, while the GTX 690 scores 17399. In the Geekbench Vulkan test, the GTX 1080 Ti scores 40511, while the GTX 690 scores 16675. There are no victories for the GTX 690 in any shared benchmark.

This does not mean the GTX 690 is without merit. Its average benchmark score of 17037 places it in the 60th percentile of all GPUs, which is actually two percentage points higher than the GTX 1080 Ti’s 58th percentile. The GTX 690’s nearest rivals include the AMD Radeon RX 7600 XT (average score 17083, delta -0.3%) and the NVIDIA GeForce RTX 3070 (average score 17208, delta -1%). This indicates that the GTX 690 remains competitive with much newer mid-range and even high-end cards in raw average compute benchmarks. The GTX 1080 Ti, despite its massive wins in specific tests, sits in a lower percentile due to its broader benchmark suite, which includes DirectX 9, 10, and 11 tests where it scores relatively low (231, 118, and 151 respectively) compared to its modern DirectX 12 and compute scores.

The GTX 1080 Ti’s wins are in modern compute and graphics APIs. Its Vulkan score of 40511 is 58.8% higher than the GTX 690’s 16675. Its OpenCL score of 67929 is 74.4% higher than the GTX 690’s 17399. The GTX 690’s performance is more evenly spread across older APIs, as evidenced by its Geekbench scores being relatively close (17399 OpenCL vs 16675 Vulkan), while the GTX 1080 Ti shows a massive gap between its OpenCL and Vulkan performance.

Architecture Differences

The two cards are built on fundamentally different architectures, nodes, and configurations. The GTX 690 uses the GK104 chip on the Kepler architecture, fabricated on a 28 nm process at TSMC. It contains 3,540 million transistors on a 294 mm² die, resulting in a transistor density of 12.0M / mm². The GTX 1080 Ti uses the GP102 chip on the Pascal architecture, also from TSMC, but on a 16 nm process. It packs 11,800 million transistors onto a 471 mm² die, achieving a density of 25.1M / mm². The newer card more than triples the transistor count and doubles the density.

Memory configurations differ drastically. The GTX 690 has 2 GB of GDDR5 on a 256-bit bus, delivering 192.3 GB/s of bandwidth. The GTX 1080 Ti has 11 GB of GDDR5X on a 352-bit bus, delivering 484.4 GB/s. This is more than 2.5 times the bandwidth and over 5 times the capacity. Clock speeds also favor the newer card: the GTX 690 runs at a base of 915 MHz with a boost of 1019 MHz, while the GTX 1080 Ti runs at 1481 MHz base and 1582 MHz boost. Memory clocks are 1502 MHz (6 Gbps effective) for the GTX 690 versus 1376 MHz (11 Gbps effective) for the GTX 1080 Ti.

Compute resources show the generational leap. The GTX 690 has 1536 shading units, 128 TMUs, and 32 ROPs. The GTX 1080 Ti has 3584 shading units, 224 TMUs, and 88 ROPs. Pixel rate jumps from 32.61 GPixel/s to 139.2 GPixel/s, and texture rate from 130.4 GTexel/s to 354.4 GTexel/s. FP32 performance is 3.130 TFLOPS for the GTX 690 versus 11.34 TFLOPS for the GTX 1080 Ti. The GTX 1080 Ti also has an FP16 rate of 177.2 GFLOPS (1:64), while the GTX 690 has no listed FP16 capability. Neither card has RT or Tensor cores.

Power and physical specifications differ as well. The GTX 690 has a TDP of 300 W, requires a 700 W PSU, and uses dual 8-pin power connectors. The GTX 1080 Ti has a lower TDP of 250 W, requires a 600 W PSU, and uses one 6-pin and one 8-pin connector. Both are dual-slot cards, with the GTX 690 measuring 279 mm in length, 111 mm in height, and 38 mm in width. The GTX 1080 Ti is slightly shorter at 267 mm, similar in height at 112 mm, and slightly thicker at 40 mm. API support also differs: the GTX 690 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175, while the GTX 1080 Ti supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4.

Head-to-Head Benchmarks

The two head-to-head benchmarks available in the data show a consistent and overwhelming advantage for the GTX 1080 Ti. In Geekbench OpenCL, the GTX 1080 Ti scores 67929, which is 74.4% higher than the GTX 690’s 17399. This is a massive gap, indicating that the GTX 1080 Ti is nearly four times faster in this compute workload. The delta percentage of -74.4% is listed from the perspective of the GTX 690, meaning the GTX 690 is 74.4% slower than the winner.

In Geekbench Vulkan, the GTX 1080 Ti scores 40511, which is 58.8% higher than the GTX 690’s 16675. While still a decisive victory, the margin is smaller than in OpenCL. This suggests the GTX 690’s relative performance is better in Vulkan than in OpenCL. The GTX 690’s Vulkan score (16675) is very close to its OpenCL score (17399), differing by only about 4%. In contrast, the GTX 1080 Ti’s Vulkan score (40511) is only about 60% of its OpenCL score (67929). This shows the GTX 1080 Ti excels particularly in OpenCL compute workloads, while the GTX 690 has more balanced performance across these two APIs.

Looking at the broader benchmark suite for the GTX 1080 Ti, its Passmark scores provide context. It scores 18600 in G3D, 9632 in GPU Compute, and 939 in G2D. In DirectX tests, it scores 231 in DirectX 9, 118 in DirectX 10, 151 in DirectX 11, and 66 in DirectX 12. These scores show that the GTX 1080 Ti’s performance is heavily skewed toward compute and modern workloads, with older DirectX versions scoring much lower. The GTX 690 has no such breakdown, but its average benchmark score of 17037 is within 0.1% of the AMD Radeon HD 7970M (17019) and 0.6% of the NVIDIA Tesla M4 (16932), placing it in a similar performance tier to those cards.

FAQ

Q: Which card is faster in OpenCL compute?

A: The GTX 1080 Ti is dramatically faster, scoring 67929 compared to the GTX 690’s 17399, a 74.4% advantage.

Q: Does the GTX 690 beat the GTX 1080 Ti in any benchmark?

A: No. In the head-to-head benchmarks available, the GTX 1080 Ti wins both Geekbench OpenCL and Geekbench Vulkan. The GTX 690 records zero wins.

Q: How do the memory bandwidths compare?

A: The GTX 1080 Ti has 484.4 GB/s of bandwidth from 11 GB of GDDR5X on a 352-bit bus. The GTX 690 has 192.3 GB/s from 2 GB of GDDR5 on a 256-bit bus.

Q: What is the average benchmark score difference?

A: The GTX 690 has an average benchmark score of 17037, while the GTX 1080 Ti has an average score of 15548. Despite losing head-to-head, the GTX 690 has a higher average score due to the different test suites.

Q: Which card has a higher transistor count?

A: The GTX 1080 Ti has 11,800 million transistors on the GP102 chip, while the GTX 690 has 3,540 million on the GK104 chip.

Q: Are both cards still in production?

A: No. Both the GTX 690 and the GTX 1080 Ti have a production status of "End-of-life."

The Verdict

The data is clear for modern workloads: the GTX 1080 Ti is the superior card. Its OpenCL score of 67929 is 74.4% higher than the GTX 690’s 17399, and its Vulkan score of 40511 is 58.8% higher than the GTX 690’s 16675. The GTX 1080 Ti also offers significantly more memory (11 GB vs 2 GB), higher bandwidth (484.4 GB/s vs 192.3 GB/s), and superior compute resources (3584 shading units vs 1536). For any user running compute-heavy tasks, modern DirectX 12 workloads, or Vulkan applications, the GTX 1080 Ti is the only logical choice from this data.

However, the GTX 690 is not entirely obsolete. Its average benchmark score of 17037 exceeds the GTX 1080 Ti’s 15548, and it sits in the 60th percentile of all GPUs, two points higher than the GTX 1080 Ti’s 58th percentile. This suggests that in a broader set of legacy benchmarks, the GTX 690 holds its own against much newer hardware. Its nearest rival is the AMD Radeon RX 7600 XT (17083, delta -0.3%) and the NVIDIA GeForce RTX 3070 (17208, delta -1%), which are both modern cards. This indicates that the GTX 690, despite its age and dual-GPU complexity, can still deliver competitive average performance in certain test suites.

The choice depends on workload. Users prioritizing modern API performance, high-resolution textures (due to 11 GB VRAM), and raw compute power should choose the GTX 1080 Ti. Users running legacy applications or those who value average benchmark consistency across a wide range of tests might find the GTX 690 adequate, but they must accept its higher TDP (300 W vs 250 W) and dual-GPU configuration. The GTX 1080 Ti wins every direct comparison in the data, making it the definitive pick for current and future software.