NVIDIA GeForce GTX 690 vs NVIDIA Quadro RTX 4000 Comparison

NVIDIA Quadro RTX 4000 vs. NVIDIA GeForce GTX 690: a workstation-class Turing card against a dual-GPU Kepler flagship from 2012. The data shows a decisive generational leap, but the GTX 690’s legacy as a high-bandwidth dual-chip card still offers a point of comparison. The Quadro RTX 4000 leads in every shared benchmark, yet the GTX 690’s specifications reveal a different design philosophy—one built for raw multi-GPU throughput rather than modern feature sets.

FAQ

Q: Which card has the higher average benchmark score?

A: The NVIDIA Quadro RTX 4000 scores 17,789 on average, while the NVIDIA GeForce GTX 690 scores 17,037. The Quadro is 4.4% ahead in aggregate performance, placing it in the 61st percentile of all GPUs versus the GTX 690’s 60th percentile.

Q: How large is the performance gap in the two shared benchmark tests?

A: In Geekbench OpenCL, the Quadro RTX 4000 scores 74,540 versus 17,399 for the GTX 690, a 328.4% advantage. In Geekbench Vulkan, the Quadro scores 78,844 versus 16,675, a 372.8% lead. The Quadro wins both head-to-head tests.

Q: What are the memory specifications of each card?

A: The Quadro RTX 4000 has 8 GB of GDDR6 memory on a 256-bit bus, delivering 416.0 GB/s bandwidth. The GTX 690 has 2 GB of GDDR5 memory on a 256-bit bus, yielding 192.3 GB/s bandwidth.

Q: Do these cards support modern APIs like DirectX 12 Ultimate and Vulkan 1.4?

A: Only the Quadro RTX 4000 supports DirectX 12 Ultimate (12_2) and Vulkan 1.4. The GTX 690 is limited to DirectX 12 (11_0) and Vulkan 1.2.175, though both cards support OpenGL 4.6.

Q: Which card has ray tracing and tensor cores?

A: The Quadro RTX 4000 includes 36 RT cores and 288 tensor cores. The GTX 690 has none, as its Kepler architecture predates those features.

Q: What is the power and cooling footprint difference?

A: The Quadro RTX 4000 has a 160 W TDP, is single-slot, and uses one 8-pin power connector with a 450 W suggested PSU. The GTX 690 has a 300 W TDP, is dual-slot, uses two 8-pin connectors, and requires a 700 W suggested PSU.

The Verdict

The data points to one conclusion: the Quadro RTX 4000 is the superior card for nearly every modern workload. It wins both head-to-head benchmarks by margins exceeding 328%, offers four times the memory capacity, and supports contemporary APIs and features like ray tracing and tensor cores. The GTX 690, despite its higher launch MSRP of 999 USD (versus 899 USD for the Quadro), cannot compete in compute or graphics tests. Its only advantage is historical—it was a dual-GPU monster in its era, but that era has passed.

Choose the Quadro RTX 4000 if you need any of the following: DirectX 12 Ultimate support, Vulkan 1.4 compatibility, RT cores for ray tracing, tensor cores for AI workloads, or more than 2 GB of VRAM. Choose the GTX 690 only if you are building a retro system or need its specific dual-DVI outputs (3x DVI, 1x mini-DisplayPort 1.2) for legacy displays. For any serious computing, the GTX 690’s 3.130 TFLOPS FP32 performance and 2 GB memory are insufficient next to the Quadro’s 7.119 TFLOPS and 8 GB.

Head-to-Head Benchmarks

The only two tests where both cards have scores are Geekbench OpenCL and Geekbench Vulkan, and the results are not close. In OpenCL, the Quadro RTX 4000 scores 74,540 against the GTX 690’s 17,399, a 328.4% delta. In Vulkan, the gap widens slightly: 78,844 versus 16,675, a 372.8% delta. These are not incremental improvements; they represent a four-to-five-fold performance increase in compute workloads.

The Quadro’s wins are driven by its architecture. It has 2,304 shading units, 144 TMUs, and 64 ROPs, versus the GTX 690’s 1,536 shading units, 128 TMUs, and 32 ROPs. The pixel rate tells the story: 98.88 GPixel/s for the Quadro versus 32.61 GPixel/s for the GTX 690. Texture rate is similarly lopsided at 222.5 GTexel/s versus 130.4 GTexel/s. FP32 compute is 7.119 TFLOPS versus 3.130 TFLOPS. Every metric favors the Quadro, and the GTX 690’s dual-GPU design does not compensate for its older, less efficient Kepler cores.

The GTX 690 has no benchmark wins in this comparison. Its average score of 17,037 is within 1% of the AMD Radeon HD 7970M (17,019) and NVIDIA Tesla M4 (16,932), but it trails the Quadro by a wide margin. The nearest rival data for the GTX 690 shows it is effectively tied with the AMD Radeon RX 7600 XT (17,083, -0.3%), which underscores how dated it is—a modern midrange card matches it.

Specification Differences

The core specifications diverge sharply. The Quadro RTX 4000 uses a TU104 chip on a 12 nm process, while the GTX 690 uses a GK104 chip on a 28 nm process. Transistor counts are 13,600 million versus 3,540 million, and die sizes are 545 mm² versus 294 mm². The Quadro packs 25.0M transistors per mm², double the GTX 690’s 12.0M.

Memory is another major split: 8 GB GDDR6 versus 2 GB GDDR5. Both use a 256-bit bus, but the Quadro’s memory clock is 1625 MHz (13 Gbps effective) versus 1502 MHz (6 Gbps effective), yielding 416.0 GB/s versus 192.3 GB/s bandwidth. The Quadro has 36 RT cores and 288 tensor cores; the GTX 690 has none.

Power and physical specs differ. The Quadro is single-slot, 241 mm long, and 111 mm high, with a 160 W TDP and one 8-pin connector. The GTX 690 is dual-slot, 279 mm long, 111 mm high, and 38 mm wide, with a 300 W TDP and two 8-pin connectors. The suggested PSU is 450 W for the Quadro and 700 W for the GTX 690. Display outputs also differ: the Quadro offers 3x DisplayPort 1.4a and 1x USB Type-C, while the GTX 690 offers 3x DVI and 1x mini-DisplayPort 1.2.

API support is a clear differentiator. The Quadro supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the GTX 690 only supports DirectX 12 (11_0) and Vulkan 1.2.175. Both support OpenGL 4.6. Release dates are 2018-11-12 for the Quadro and 2012-05-02 for the GTX 690, which explains the feature gap.

Architecture Differences

The Quadro RTX 4000 is built on the Turing architecture, a 12 nm design from TSMC. Turing introduces dedicated RT cores for ray tracing and tensor cores for AI acceleration, neither of which exist on the GTX 690’s Kepler architecture. Kepler, also from TSMC but on 28 nm, was designed for the GeForce 600 generation and lacks hardware-level support for modern compute features.

The transistor density difference is stark: 25.0M per mm² for Turing versus 12.0M per mm² for Kepler. This allows the Quadro to fit 2,304 shading units, 144 TMUs, and 64 ROPs into a 545 mm² die, while the GTX 690’s 294 mm² die holds 1,536 shading units, 128 TMUs, and 32 ROPs. The Quadro also has 36 RT cores and 288 tensor cores, which are absent on the GTX 690.

Memory architecture differs as well. The Quadro uses GDDR6 with a 256-bit bus and 416.0 GB/s bandwidth, while the GTX 690 uses GDDR5 with the same bus width but only 192.3 GB/s. The Quadro’s memory clock is 1625 MHz (13 Gbps effective) versus 1502 MHz (6 Gbps effective). FP16 support is present on the Quadro at 14.24 TFLOPS (2:1), while the GTX 690 has no FP16 specification listed.

The API ceiling reflects the architectural gap. Turing supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, enabling features like variable rate shading and mesh shaders. Kepler is capped at DirectX 12 (11_0) and Vulkan 1.2.175, which limits compatibility with newer games and applications.

Where Each One Wins

The Quadro RTX 4000 wins in every category that matters for modern computing. It dominates compute benchmarks, with 328.4% and 372.8% leads in OpenCL and Vulkan respectively. Its 8 GB memory is essential for large datasets, and its 416.0 GB/s bandwidth handles high-resolution textures and multi-tasking. RT cores enable ray-traced rendering, and tensor cores accelerate AI inference and DLSS-style workloads. Its single-slot design and 160 W TDP make it easier to install in dense workstation builds, and its DisplayPort 1.4a outputs support high refresh rate and HDR monitors.

The GTX 690 wins only in niche legacy scenarios. Its 3x DVI outputs are useful for old monitors that lack DisplayPort or HDMI. Its dual-GPU design, while not reflected in the benchmark scores, was historically for multi-monitor setups or SLI-based gaming. However, its 2 GB memory and 192.3 GB/s bandwidth are severe bottlenecks for modern assets. Its 300 W TDP and dual-slot footprint are drawbacks, and its lack of RT or tensor cores means it cannot accelerate modern graphics features.

For a practical builder, the choice is clear: the Quadro RTX 4000 is the only viable option for current software. The GTX 690 is a collector’s item or a paperweight for retro rigs. The data shows a 4.4% average score advantage for the Quadro, but the per-test deltas reveal the true scale—it is not a close race. If you need professional-grade compute, memory capacity, or API support, the Quadro is the answer. If you need DVI outputs and nothing else, the GTX 690 exists, but it will not satisfy any performance requirement.