NVIDIA GeForce GTX 680 vs NVIDIA GeForce GTX 770 Comparison

The NVIDIA GeForce GTX 770 and NVIDIA GeForce GTX 680 share the same GK104 chip and Kepler architecture, but the benchmark data separates them into distinct performance tiers. Across three Geekbench tests, the GTX 770 claims two wins, while the GTX 680 takes one, resulting in an average benchmark score of 14747 for the GTX 770 against 14150 for the GTX 680. The GTX 770 also holds a higher percentile ranking at 57 versus 55, placing it slightly higher among all GPUs.

Head-to-Head Benchmarks

The most decisive result appears in the Geekbench Metal test, where the GTX 770 scores 9369 against the GTX 680’s 7897. That is an 18.6% advantage, the largest margin in any head-to-head comparison between the two cards. This substantial lead suggests the GTX 770 is markedly stronger in Metal-accelerated workloads, likely due to its higher clock speeds and memory bandwidth.

The Geekbench Vulkan test also favors the GTX 770, though by a narrower margin. The GTX 770 posts 19166, while the GTX 680 reaches 17646, a difference of 8.6%. This result reinforces the pattern that the GTX 770 generally outperforms its predecessor in modern API benchmarks, but the gap is not as pronounced as in Metal.

The single victory for the GTX 680 comes in Geekbench OpenCL, where it scores 16906 against the GTX 770’s 15707. That is a 7.1% swing in favor of the older card. This is an interesting anomaly, as the GTX 680 trails in both other tests yet manages to overtake the GTX 770 in OpenCL compute performance. The data suggests that OpenCL workloads may be less sensitive to the clock and memory advantages of the GTX 770, or that the GTX 680’s specific configuration yields better results in this particular benchmark.

When placed against the broader field, the GTX 770’s average score of 14747 puts it 0.4% ahead of the AMD Radeon RX 5500 XT (14692) and 1.5% ahead of the NVIDIA Quadro M2000 (14532). The GTX 770 also outperforms the NVIDIA GeForce GTX 660 Ti by 2.1% (15063) and the AMD Radeon Pro 560X by 2.2% (15082). For the GTX 680, its average of 14150 places it 0.9% ahead of the NVIDIA Tesla K10 (14029), but 0.9% behind the NVIDIA GeForce GTX 1070 Ti (14277), 1.2% behind the Intel Iris Xe MAX Graphics (14315), and 1.4% behind the AMD Radeon Vega 11 (14352).

FAQ

Q: Which card has the higher average benchmark score?

A: The NVIDIA GeForce GTX 770 has an average benchmark score of 14747, while the NVIDIA GeForce GTX 680 scores 14150. The GTX 770 sits 4.2% higher in this aggregate metric.

Q: How do the two cards compare in Metal performance?

A: The GTX 770 wins the Geekbench Metal test with a score of 9369, which is 18.6% higher than the GTX 680’s 7897. This is the largest performance gap between the two cards in any benchmark.

Q: Is there any test where the GTX 680 beats the GTX 770?

A: Yes, in the Geekbench OpenCL test, the GTX 680 scores 16906, which is 7.1% higher than the GTX 770’s 15707. This is the only head-to-head test where the GTX 680 comes out ahead.

Q: What is the difference in their percentile rankings?

A: The GTX 770 ranks in the 57th percentile among all GPUs, while the GTX 680 ranks in the 55th percentile. This indicates the GTX 770 performs better relative to the entire GPU landscape.

Q: How does each card compare to the NVIDIA GeForce GTX 1070 Ti?

A: The GTX 770 does not have a direct comparison to the GTX 1070 Ti in the nearest rivals data. The GTX 680, however, is 0.9% behind the GTX 1070 Ti, which has an average score of 14277.

Q: What is the closest rival for each card?

A: The GTX 770’s closest rival is the AMD Radeon RX 5500 XT, with a 0.4% difference in average score. The GTX 680’s closest rival is the NVIDIA Tesla K10, with a 0.9% difference.

The Verdict

The data points to the NVIDIA GeForce GTX 770 as the stronger card overall. Its average benchmark score of 14747 exceeds the GTX 680’s 14150, and it wins two of the three head-to-head tests. The 18.6% lead in Metal is substantial, and the 8.6% lead in Vulkan confirms a general trend of superiority in these workloads. The GTX 770 also achieves a higher percentile rank, 57 versus 55, reflecting its better standing among all GPUs.

However, the GTX 680 is not without merit. Its 7.1% victory in OpenCL shows that it can handle certain compute tasks more effectively than the newer card. For users whose primary workload is OpenCL-based, the GTX 680 may be the better choice despite its lower overall average. The GTX 680 also has a lower TDP of 195 W compared to the GTX 770’s 230 W, which could be a consideration for power-constrained systems, though this is not a performance metric.

For most gaming and general-purpose scenarios, the GTX 770 is the clear pick based on the benchmark results. Its wins in Metal and Vulkan, combined with a higher average score, make it the more capable card across a broader range of applications. The GTX 680 remains relevant for those specifically targeting OpenCL performance, but the data favors the GTX 770 in the majority of cases.

Specification Differences

The two cards share several core specifications, but key differences emerge in clocks, memory, and power. The GTX 770 has a base clock of 1046 MHz and a boost clock of 1085 MHz, while the GTX 680 operates at 1006 MHz base and 1058 MHz boost. This gives the GTX 770 a 40 MHz advantage at base and a 27 MHz advantage at boost.

Memory configuration also differs. The GTX 770 runs its GDDR5 memory at 1753 MHz, translating to 7 Gbps effective, while the GTX 680 runs at 1502 MHz, or 6 Gbps effective. Both cards have 2 GB of memory on a 256-bit bus, but the GTX 770’s higher memory clock yields a bandwidth of 224.4 GB/s compared to the GTX 680’s 192.3 GB/s. This is a 32.1 GB/s difference, which contributes to the GTX 770’s superior performance in memory-intensive tests.

Power requirements also differ. The GTX 770 has a TDP of 230 W and requires a 550 W power supply, while the GTX 680 has a TDP of 195 W and a 450 W suggested PSU. The GTX 770 uses a 1x 6-pin plus 1x 8-pin power connector configuration, whereas the GTX 680 uses 2x 6-pin connectors. The GTX 770 is slightly longer at 267 mm (10.5 inches) versus the GTX 680’s 256 mm (10.1 inches), though height and width are identical at 111 mm and 38 mm.

Launch MSRP differs as well, with the GTX 770 at 399 USD and the GTX 680 at 499 USD. The GTX 770 was released on 2013-05-29, while the GTX 680 came earlier on 2012-03-21. Both cards are end-of-life, and both support PCIe 3.0 x16 with identical display outputs of 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2.

Architecture Differences

Both the GTX 770 and GTX 680 are built on the same GK104 chip using the Kepler architecture, manufactured by TSMC on a 28 nm process. They share identical transistor counts of 3,540 million and the same die size of 294 mm², resulting in a transistor density of 12.0M per mm² for both. The shading units, TMUs, and ROPs are also identical, with 1536 shading units, 128 TMUs, and 32 ROPs.

The pixel rate and texture rate are nearly the same, with the GTX 770 achieving 34.72 GPixel/s and 138.9 GTexel/s, while the GTX 680 achieves 33.86 GPixel/s and 135.4 GTexel/s. FP32 performance is similarly close, with the GTX 770 at 3.333 TFLOPS and the GTX 680 at 3.250 TFLOPS. Neither card has RT cores or tensor cores, and neither supports FP16.

The API support is identical for both cards: DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The GTX 770 belongs to the GeForce 700 generation with a predecessor in GeForce 600 and a successor in GeForce 900. The GTX 680 belongs to the GeForce 600 generation with a predecessor in GeForce 500 and a successor in GeForce 700.

The architectural differences are minimal since they share the same chip and design. The performance gaps observed in benchmarks stem primarily from the GTX 770’s higher clock speeds and faster memory, not from any fundamental architectural changes. The GTX 770 is essentially a refined version of the GTX 680 with enhanced clocks and memory bandwidth, while maintaining the same core structure.

Where Each One Wins

The GTX 770 dominates in Metal and Vulkan workloads. The 18.6% lead in Metal and the 8.6% lead in Vulkan make it the preferred choice for applications that leverage these APIs. Its higher memory bandwidth of 224.4 GB/s likely plays a role in these wins, as does its higher boost clock of 1085 MHz. For gaming, modern titles that utilize Vulkan will see a clear benefit from the GTX 770.

The GTX 680 wins in OpenCL, posting a 7.1% advantage over the GTX 770. This makes it the better option for compute tasks that rely on OpenCL, such as certain professional applications or GPU-accelerated processing. The GTX 680 achieves this despite lower clocks and bandwidth, suggesting that its specific configuration is better optimized for this workload.

For users prioritizing overall performance, the GTX 770’s higher average score of 14747 and its wins in two of three tests make it the broader performer. The GTX 680’s single win in OpenCL is notable but narrow in scope. The GTX 770 also offers a higher pixel rate of 34.72 GPixel/s and texture rate of 138.9 GTexel/s, which may benefit certain rendering tasks.

Power consumption is a differentiator if that matters to a build. The GTX 680’s 195 W TDP is lower than the GTX 770’s 230 W, and it requires a smaller power supply at 450 W versus 550 W. For compact or lower-wattage systems, the GTX 680 could be the more practical choice, even if its raw performance is lower. The GTX 770’s additional power draw may necessitate a more robust PSU and cooling solution.

In summary, the GTX 770 is the winner for most users, especially those focused on Metal or Vulkan performance. The GTX 680 remains a viable option for OpenCL-heavy workloads or for systems where lower power consumption is critical. The benchmark data clearly favors the GTX 770 in overall terms, but the GTX 680’s specific strengths keep it relevant in targeted use cases.