NVIDIA GeForce GTX 690 vs NVIDIA GeForce GTX 780 Comparison

The Verdict

The data shows a clear split between these two Kepler-era NVIDIA cards. The GeForce GTX 780 wins every recorded head-to-head benchmark, leading by 31.4% in OpenCL and 47% in Vulkan. The GTX 780 also holds a higher average benchmark score of 19164 versus 17037 for the GTX 690, and it sits at the 64th percentile among all GPUs compared to the GTX 690's 60th percentile. For anyone choosing strictly from these measurements, the GTX 780 is the stronger option in raw compute performance.

However, the GTX 690 has its own identity. It is a dual-GPU card built around two GK104 chips, and its architecture suggests a different intent. The GTX 690 was launched with a higher MSRP of 999 USD, while the GTX 780 launched at 649 USD. The recorded data does not show any benchmark where the GTX 690 wins, so the higher launch price does not translate into a performance advantage in the tests available. The GTX 780 is the pick for users who prioritize compute workloads and modern API performance. The GTX 690 remains relevant only in scenarios where its dual-chip design might have been designed for multi-monitor setups or specific multi-GPU scaling, but the database does not contain evidence of such an advantage.

Architecture Differences

Both cards use NVIDIA's Kepler architecture and are fabricated on TSMC's 28 nm process, but they are fundamentally different designs. The GTX 780 uses the GK110 chip, a large monolithic die measuring 561 mm² with 7,080 million transistors. The GTX 690 uses two GK104 chips, each measuring 294 mm² with 3,540 million transistors. This means the GTX 690 packs two smaller chips, totaling 7,080 million transistors across both dies, but each die is far smaller than the GK110. The transistor density is nearly identical: 12.6M per mm² for the GTX 780 and 12.0M per mm² for the GTX 690.

The GTX 780 has 2,304 shading units, 192 texture mapping units, and 48 raster operation units. The GTX 690, despite having two chips, lists a combined 1,536 shading units, 128 TMUs, and 32 ROPs. This is a critical difference: the GTX 690's dual-chip design does not double the shader count; instead, it offers fewer total shaders than the single-chip GTX 780. The GTX 780's GK110 is a larger, more complex chip, while the GTX 690's two GK104 dies are each smaller and more modest in raw resource counts.

Clock speeds differ as well. The GTX 780 runs at a base clock of 863 MHz with a boost of 902 MHz. The GTX 690 runs higher at 915 MHz base and 1019 MHz boost, but this higher clock does not compensate for the lower shader count. Memory clocks are identical at 1502 MHz with 6 Gbps effective speed, but the memory configuration diverges: the GTX 780 has 3 GB of GDDR5 on a 384-bit bus, delivering 288.4 GB/s of bandwidth, while the GTX 690 has 2 GB on a 256-bit bus per chip, delivering 192.3 GB/s. The GTX 690's dual-chip nature means each GK104 has its own 2 GB allocation, but the effective bandwidth per chip is lower than the GTX 780's single larger pool.

Both cards support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175, so there is no API feature gap. The GTX 780 uses a 1x 6-pin plus 1x 8-pin power connector configuration, while the GTX 690 requires 2x 8-pin connectors. The GTX 690 has a higher TDP at 300 W versus 250 W for the GTX 780, and it suggests a 700 W power supply versus 600 W for the GTX 780.

Head-to-Head Benchmarks

The database records two direct comparisons between these cards, and the GTX 780 wins both decisively. In the Geekbench OpenCL test, the GTX 780 scores 22863 against the GTX 690's 17399, a delta of 31.4%. This is a substantial margin in compute performance, likely reflecting the GTX 780's higher shader count and larger memory bandwidth.

The Vulkan test shows an even larger gap. The GTX 780 scores 24514, while the GTX 690 manages 16675, a delta of 47%. This near-50% advantage suggests that the GTX 780's single-chip design handles the modern API more efficiently than the GTX 690's dual-GPU configuration. Vulkan workloads often require explicit multi-GPU management, and the data implies that the GTX 690 does not scale well in this context.

The GTX 780 also has a third benchmark result, Geekbench Metal, scoring 10114, which has no corresponding result for the GTX 690. This absence itself is informative: the GTX 690 lacks a recorded Metal score, possibly indicating it was not tested or not supported in that environment. The GTX 780's average benchmark score of 19164 versus 17037 for the GTX 690 reinforces the head-to-head results, showing a consistent overall advantage of roughly 12.5%.

The nearest rivals for the GTX 780 include the NVIDIA TITAN Xp with an average score of 19177 (0.1% higher), the Tesla K20m at 19089 (0.4% lower), the RTX 4050 Mobile at 19049 (0.6% lower), and the AMD Radeon RX 6600 at 19036 (0.7% lower). This places the GTX 780 in a very tight cluster of performance, sitting almost exactly between the TITAN Xp and the K20m. The GTX 690's nearest rivals are the AMD Radeon HD 7970M at 17019 (0.1% lower), the RX 7600 XT at 17083 (0.3% higher), the Tesla M4 at 16932 (0.6% lower), and the RTX 3070 at 17208 (1.0% higher). The GTX 690's position among these rivals is less flattering, as it sits below the RX 7600 XT and RTX 3070.

Specification Differences

The two cards differ in nearly every hardware specification except for the process node, foundry, memory type, memory clock, and API support. The GTX 780 uses the GK110 chip, while the GTX 690 uses two GK104 chips. Transistor count is 7,080 million for the GTX 780 versus 3,540 million for the GTX 690 (per die), and die size is 561 mm² versus 294 mm². Transistor density is 12.6M per mm² versus 12.0M per mm².

Clock speeds favor the GTX 690: 915 MHz base and 1019 MHz boost versus 863 MHz and 902 MHz for the GTX 780. However, the GTX 780 has more execution resources: 2304 shading units versus 1536, 192 TMUs versus 128, and 48 ROPs versus 32. Pixel rate is 43.30 GPixel/s for the GTX 780 versus 32.61 GPixel/s for the GTX 690. Texture rate is 173.2 GTexel/s versus 130.4 GTexel/s. FP32 compute is 4.156 TFLOPS versus 3.130 TFLOPS.

Memory size differs: 3 GB versus 2 GB. Bus width is 384-bit versus 256-bit. Bandwidth is 288.4 GB/s versus 192.3 GB/s. TDP is 250 W versus 300 W. Power connectors are 1x 6-pin + 1x 8-pin versus 2x 8-pin. Suggested PSU is 600 W versus 700 W.

Physical dimensions are close but not identical: the GTX 780 is 267 mm long (10.5 inches), while the GTX 690 is 279 mm (11 inches). Height and width are the same at 111 mm (4.4 inches) and 38 mm (1.5 inches). Display outputs differ: the GTX 780 has 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the GTX 690 has 3x DVI and 1x mini-DisplayPort 1.2.

Release dates differ by about a year: the GTX 690 launched on 2012-05-02, and the GTX 780 launched on 2013-05-22. The GTX 690 is part of the GeForce 600 generation with a predecessor in the GeForce 500 series and successor in GeForce 700. The GTX 780 is part of the GeForce 700 generation with a predecessor in GeForce 600 and successor in GeForce 900. Both are end-of-life production status.

FAQ

Q: Which card has a higher average benchmark score?

A: The GTX 780 has an average benchmark score of 19164, which is significantly higher than the GTX 690's 17037. The gap is roughly 12.5%.

Q: Does the GTX 690's dual-chip design double the performance?

A: No. The GTX 690 has 1,536 shading units, 128 TMUs, and 32 ROPs, which is less than the GTX 780's 2,304 shading units, 192 TMUs, and 48 ROPs. The GTX 780 wins both recorded benchmarks.

Q: What is the largest performance difference in the head-to-head tests?

A: The Vulkan benchmark shows the largest gap, with the GTX 780 scoring 24514 versus 16675 for the GTX 690, a delta of 47%.

Q: Which card has higher clock speeds?

A: The GTX 690 runs at 915 MHz base and 1019 MHz boost, while the GTX 780 runs at 863 MHz base and 902 MHz boost. The GTX 690 is faster in clock speed but loses on execution resources.

Q: Are the cards compatible with the same APIs?

A: Yes, both support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175.

Q: Which card requires more power connectors?

A: The GTX 690 requires 2x 8-pin connectors, while the GTX 780 uses 1x 6-pin plus 1x 8-pin. The GTX 690 also has a higher TDP of 300 W versus 250 W.

Where Each One Wins

The GTX 780 wins in every measured category. In OpenCL, it is 31.4% ahead. In Vulkan, it is 47% ahead. Its average benchmark score is higher by over 2,100 points. It has more shading units, TMUs, ROPs, higher pixel rate, higher texture rate, higher FP32 throughput, more memory, wider bus, and more bandwidth. It also has a lower TDP at 250 W versus 300 W and a lower suggested PSU of 600 W versus 700 W. For compute-heavy tasks like OpenCL workloads, the GTX 780's 4.156 TFLOPS versus 3.130 TFLOPS gives it a clear edge. For modern API workloads like Vulkan, the single-chip design appears to avoid the scaling penalties that likely affect the dual-GPU GTX 690.

The GTX 690's advantages are narrower. It runs at higher clock speeds, with a base of 915 MHz and boost of 1019 MHz, which could benefit lightly threaded tasks that do not fully utilize the GTX 780's larger shader array. It has a dual-chip layout that may have been intended for multi-monitor output, featuring 3x DVI and 1x mini-DisplayPort 1.2, while the GTX 780 offers 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. The GTX 690 also has a higher launch MSRP of 999 USD, though this does not correspond to any performance advantage in the recorded data.

The GTX 690's nearest rival cluster includes the RTX 3070 at 17208 and the RX 7600 XT at 17083, both of which score higher than the GTX 690. The GTX 780, by contrast, sits within 0.7% of the RX 6600 and within 0.1% of the TITAN Xp, showing it is competitive with much newer hardware. For users with older systems that already have high-wattage power supplies, the GTX 690 might still function, but the data does not support choosing it over the GTX 780 for any benchmark-verified workload.

In practical terms, the GTX 780 is the card for compute, modern API compatibility, and overall efficiency. The GTX 690 is the card for those who value its dual-chip heritage or require its specific display output configuration, but the benchmark record shows no scenario where it outperforms the GTX 780. The GTX 780's higher percentile rank at 64 versus 60 further confirms its superior standing in the broader GPU landscape.