NVIDIA GeForce GTX 690 vs NVIDIA TITAN Xp Comparison

Head-to-Head Benchmarks

The database contains two shared benchmark results for these cards, and both are decisive victories for the TITAN Xp. In Geekbench OpenCL, the TITAN Xp records 72,585 points against the GTX 690's 17,399 points. That is a 317.2% advantage, meaning the newer card delivers roughly four times the compute throughput in this workload. The gap is even larger in Geekbench Vulkan, where the TITAN Xp scores 87,180 versus 16,675, a 422.8% lead. In both tests, the margin is so wide that the GTX 690 is not merely behind; it is operating in a different performance class entirely.

The TITAN Xp's own benchmark portfolio shows why it sits at the 64th percentile among all GPUs in the database. Its average benchmark score is 19,177, and it has a Passmark G3D result of 18,750, a Passmark GPU Compute score of 9,430, and a Passmark G2D score of 883. The GTX 690, by contrast, has only two recorded benchmarks, both Geekbench tests, and its average score is 17,037. That average places it at the 60th percentile overall, a modest gap in percentile terms but a reflection of the fact that the GTX 690's remaining benchmark data is sparse. The TITAN Xp has ten recorded benchmark entries spanning DirectX 9, 10, 11, 12, compute, and 2D tests, while the GTX 690 has only the two OpenCL and Vulkan results in the database. This asymmetry limits direct comparison, but in the two tests where both cards appear, the TITAN Xp's dominance is unambiguous.

Looking at the TITAN Xp's nearest rivals in the database, it edges out the GTX 780 by 0.1% in average score, the Tesla K20m by 0.5%, the RTX 4050 Mobile by 0.7%, and the RX 6600 by 0.7%. These are narrow margins, indicating that the TITAN Xp's average score sits just above a cluster of cards in the 19,000 range. The GTX 690's nearest rivals tell a similar story but from a lower vantage point: it is 0.1% ahead of the HD 7970M, 0.3% behind the RX 7600 XT, 0.6% ahead of the Tesla M4, and 1% behind the RTX 3070. The GTX 690's average of 17,037 places it in a group where the surrounding cards are all within a percentage point, but that entire group sits roughly 2,000 points below the TITAN Xp's cluster.

The delta percentages in the head-to-head tests are the clearest signal. A 317.2% lead in OpenCL and a 422.8% lead in Vulkan are not incremental improvements; they represent generational leaps. The TITAN Xp's FP32 throughput of 12.15 TFLOPS compared to the GTX 690's 3.130 TFLOPS reinforces this, as does the memory bandwidth: 547.6 GB/s versus 192.3 GB/s. The texture rate tells the same story, 379.7 GTexel/s versus 130.4 GTexel/s, and the pixel rate follows, 151.9 GPixel/s versus 32.61 GPixel/s. Every raw throughput metric in the database favors the TITAN Xp by a factor of roughly three to four, which matches the benchmark deltas.

The Verdict

The recorded data points to a single conclusion: the TITAN Xp is the superior card by every measurable metric in the database. In the two shared benchmarks, it wins both, with deltas of 317.2% and 422.8%. Its average benchmark score is 19,177 versus 17,037 for the GTX 690, a difference of 2,140 points. Its percentile ranking among all GPUs is 64 versus 60. The TITAN Xp wins 2 out of 2 head-to-head tests, while the GTX 690 wins none.

Who should pick the TITAN Xp? Anyone whose workload involves OpenCL or Vulkan compute, based on the recorded data. The card also offers 12 GB of GDDR5X memory, 3,840 shading units, 240 texture mapping units, and 96 ROPs, all of which are documented in the database. The GTX 690, with 2 GB of GDDR5 memory, 1,536 shading units, 128 TMUs, and 32 ROPs, is at a structural disadvantage in memory capacity and shading throughput. The GTX 690's only advantage in the specification sheet is its TDP of 300 W versus 250 W, which is not a performance benefit. Its power connectors require 2x 8-pin inputs, while the TITAN Xp uses 1x 6-pin plus 1x 8-pin. The suggested PSU for the GTX 690 is 700 W, compared to 600 W for the TITAN Xp.

The GTX 690 remains a functional card for legacy DirectX workloads, but the database shows no scenario where it outperforms the TITAN Xp. Its launch MSRP was 999 USD, while the TITAN Xp launched at 1,199 USD. The price difference is recorded in the database, but the performance difference is far larger than the price difference. For any buyer choosing between these two based on the data, the TITAN Xp is the clear pick.

Architecture Differences

The TITAN Xp is built on the Pascal architecture, using the GP102 chip, and it belongs to the GeForce 10 generation. The GTX 690 uses the Kepler architecture, built on the GK104 chip, and belongs to the GeForce 600 generation. The process node differs substantially: the TITAN Xp is fabricated on a 16 nm process at TSMC, while the GTX 690 uses a 28 nm process at the same foundry. This node shrink is a primary reason for the TITAN Xp's higher transistor density of 25.1 million transistors per square millimeter, compared to 12.0 million for the GTX 690.

The transistor counts reflect the design differences. The TITAN Xp packs 11,800 million transistors on a 471 mm² die. The GTX 690 has 3,540 million transistors on a 294 mm² die. Notably, the GTX 690 is a dual-GPU card, meaning it contains two GK104 chips, but the database lists its transistor count and die size as single-chip figures. Even accounting for that, the TITAN Xp's single GP102 die contains more than three times the transistors of one GK104 die, and the Pascal architecture delivers far higher per-transistor efficiency.

The memory subsystems are architecturally different as well. The TITAN Xp uses GDDR5X memory with a 384-bit bus, while the GTX 690 uses GDDR5 with a 256-bit bus. The effective memory clock is 11.4 Gbps on the TITAN Xp versus 6 Gbps on the GTX 690. This combination yields 547.6 GB/s of bandwidth on the TITAN Xp versus 192.3 GB/s on the GTX 690. The TITAN Xp's memory capacity of 12 GB is six times the GTX 690's 2 GB. For modern workloads that require large framebuffers or in-memory compute, this is a decisive architectural advantage.

The API support also differs. The TITAN Xp supports DirectX 12 with feature level 12_1, while the GTX 690 supports DirectX 12 with feature level 11_0. Both support OpenGL 4.6, but the Vulkan versions differ: the TITAN Xp supports Vulkan 1.4, while the GTX 690 supports Vulkan 1.2.175. The TITAN Xp's display outputs include HDMI 2.0 and three DisplayPort 1.4a connections, while the GTX 690 provides three DVI ports and one mini-DisplayPort 1.2. The TITAN Xp also has a higher pixel rate of 151.9 GPixel/s versus 32.61 GPixel/s, and a higher texture rate of 379.7 GTexel/s versus 130.4 GTexel/s.

Specification Differences

The two cards differ in nearly every specification field in the database. The TITAN Xp has 3,840 shading units, 240 TMUs, and 96 ROPs. The GTX 690 has 1,536 shading units, 128 TMUs, and 32 ROPs. The TITAN Xp's base clock is 1405 MHz with a boost clock of 1582 MHz. The GTX 690 runs at 915 MHz base and 1019 MHz boost. The memory configuration differs as described: 12 GB GDDR5X on a 384-bit bus versus 2 GB GDDR5 on a 256-bit bus, with effective speeds of 11.4 Gbps versus 6 Gbps.

The FP32 compute rating is 12.15 TFLOPS for the TITAN Xp and 3.130 TFLOPS for the GTX 690. The TITAN Xp has an FP16 rating of 189.8 GFLOPS with a 1:64 ratio, while the GTX 690 has no recorded FP16 value. The TDP is 250 W for the TITAN Xp and 300 W for the GTX 690. Power connectors are 1x 6-pin plus 1x 8-pin for the TITAN Xp, versus 2x 8-pin for the GTX 690. The suggested PSU is 600 W versus 700 W. Physical dimensions differ slightly: the TITAN Xp is 267 mm long, 112 mm tall, and 40 mm wide, while the GTX 690 is 279 mm long, 111 mm tall, and 38 mm wide. Both are dual-slot cards.

The release dates are far apart. The GTX 690 was released on May 2, 2012, while the TITAN Xp was released on April 5, 2017. The GTX 690's predecessor is the GeForce 500 series and its successor is the GeForce 700 series. The TITAN Xp's predecessor is the GeForce 900 series and its successor is the GeForce 20 series. Both cards are marked as end-of-life in the database. The launch MSRP for the TITAN Xp is 1,199 USD, and for the GTX 690 it is 999 USD.

FAQ

Q: Which card has higher memory bandwidth?

A: The TITAN Xp has 547.6 GB/s of bandwidth, while the GTX 690 has 192.3 GB/s. This is a 285% advantage for the TITAN Xp.

Q: How do the two cards compare in Geekbench Vulkan?

A: The TITAN Xp scores 87,180 versus 16,675 for the GTX 690, a 422.8% lead for the TITAN Xp.

Q: What is the average benchmark score for each card?

A: The TITAN Xp has an average benchmark score of 19,177, while the GTX 690 has an average of 17,037.

Q: Which card has more shading units?

A: The TITAN Xp has 3,840 shading units, compared to 1,536 for the GTX 690.

Q: What is the process node for each card?

A: The TITAN Xp is fabricated on a 16 nm process, while the GTX 690 uses a 28 nm process. Both are made by TSMC.

Q: Does the GTX 690 win any head-to-head benchmark?

A: No. In the database, the TITAN Xp wins 2 out of 2 head-to-head tests, and the GTX 690 wins 0.

Where Each One Wins

The TITAN Xp wins in every recorded benchmark category where both cards appear. In Geekbench OpenCL, it leads by 317.2%. In Geekbench Vulkan, it leads by 422.8%. Beyond those shared tests, the TITAN Xp has additional recorded results in Passmark DirectX 9, 10, 11, 12, G2D, G3D, and GPU Compute, while the GTX 690 has no entries in those categories. For workloads involving OpenCL compute, Vulkan rendering, or any DirectX variant, the data favors the TITAN Xp.

The GTX 690's only recorded benchmark scores are the two Geekbench tests, and it loses both by wide margins. There is no test in the database where the GTX 690 comes out ahead. Its higher TDP of 300 W does not translate into any performance benefit in the recorded data. Its lower transistor count, smaller memory bus, and older architecture all contribute to its lower scores.

For users with legacy software that relies on DVI outputs, the GTX 690 provides three DVI ports, while the TITAN Xp offers HDMI 2.0 and DisplayPort 1.4a outputs. For users with older displays that lack HDMI or DisplayPort, the GTX 690's DVI connectivity might be more convenient, though adapters exist. The GTX 690 also has a shorter release-to-end-of-life history, having been succeeded by the GeForce 700 series.

The TITAN Xp wins in raw compute, memory capacity, memory bandwidth, shading throughput, texture throughput, pixel throughput, API support, and every shared benchmark. The GTX 690 wins only in the sense of having a lower launch MSRP by 200 USD, but the performance gap in the database is so large that this price difference does not offset the measured results. For any application that can use the TITAN Xp's features, the data shows a decisive advantage. The GTX 690 remains a historical artifact of the Kepler era, while the TITAN Xp continues to hold a respectable position at the 64th percentile of all GPUs in the database.