NVIDIA GeForce GTX 690 vs NVIDIA RTX PRO 6000 Blackwell Comparison

Head-to-Head Benchmarks

The data presents a direct comparison challenge: the two cards share no common benchmark tests in the FACT PACK. The GeForce GTX 690 has results for Geekbench OpenCL (17,399) and Geekbench Vulkan (16,675), averaging 17,037. The RTX PRO 6000 Blackwell has a single 3DMark Steel Nomad DX12 score of 16,408. Because these tests measure different workloads, a direct numeric comparison is not possible from the available facts. What the data does show is where each card sits relative to its own peer group.

The GTX 690's average score of 17,037 places it at the 60th percentile among all GPUs. Its nearest rivals provide context: it is 0.1% above the AMD Radeon HD 7970M (17,019), 0.3% below the AMD Radeon RX 7600 XT (17,083), 0.6% above the NVIDIA Tesla M4 (16,932), and 1.0% below the NVIDIA GeForce RTX 3070 (17,208). The RTX PRO 6000 Blackwell's 16,408 score sits at the 59th percentile, with rivals clustered tightly: 0.0% versus the AMD Radeon PRO W7500 (16,415), 0.3% above the AMD Radeon RX 5700 XT (16,361), 0.4% above the AMD Radeon Pro 5600M (16,351), and 0.6% below the NVIDIA GeForce RTX 5090 D V2 (16,504).

The key takeaway is not which card scores higher—the tests are incompatible—but that both cards land in nearly identical percentile territory. The GTX 690 edges ahead by raw average (17,037 vs 16,408), yet the RTX PRO 6000 Blackwell's result comes from a modern DX12 workload, while the GTX 690's scores stem from older compute and Vulkan tests. The deltaPct values against their respective rivals are all within ±1%, indicating both cards are tightly grouped with their contemporaries. Neither card dominates its benchmark pool; each is a mid-pack performer in its own era.

Architecture Differences

The architectural gap is generational and extreme. The GTX 690 uses the GK104 chip on a 28 nm TSMC process with 3,540 million transistors on a 294 mm² die, yielding a transistor density of 12.0M / mm². The RTX PRO 6000 Blackwell uses the GB202 chip on a 5 nm TSMC process with 92,200 million transistors on a 750 mm² die, achieving 122.9M / mm². That is a roughly 26x increase in transistor count and a 10x increase in density, all packed into a die that is only about 2.5x larger.

The GTX 690 is built on the Kepler architecture from the GeForce 600 generation. The RTX PRO 6000 Blackwell uses Blackwell 2.0 architecture from the Blackwell PRO W (x000) generation. The GTX 690 has 1,536 shading units, 128 TMUs, and 32 ROPs. The RTX PRO 6000 Blackwell has 24,064 shading units, 752 TMUs, and 192 ROPs—roughly 15.7x, 5.9x, and 6.0x more, respectively. The RTX PRO 6000 Blackwell also adds 188 ray tracing cores and 752 tensor cores; the GTX 690 has neither.

Memory architecture differs fundamentally. The GTX 690 uses 2 GB of GDDR5 on a 256-bit bus with 192.3 GB/s bandwidth. The RTX PRO 6000 Blackwell uses 96 GB of GDDR7 on a 512-bit bus with 1.79 TB/s bandwidth—48x the capacity and roughly 9.3x the bandwidth. Clock speeds also diverge: the GTX 690 runs at 915 MHz base and 1019 MHz boost, with 1502 MHz memory (6 Gbps effective); the RTX PRO 6000 Blackwell runs at 1590 MHz base and 2617 MHz boost, with 1750 MHz memory (28 Gbps effective).

The feature set reflects the era. The GTX 690 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The RTX PRO 6000 Blackwell supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The GTX 690 uses PCIe 3.0 x16; the RTX PRO 6000 Blackwell uses PCIe 5.0 x16. Display outputs differ: the GTX 690 has 3x DVI and 1x mini-DisplayPort 1.2; the RTX PRO 6000 Blackwell has 4x DisplayPort 2.1b. The RTX PRO 6000 Blackwell is also far denser in compute throughput, with 126.0 TFLOPS FP32 and 126.0 TFLOPS FP16 (1:1), versus the GTX 690's 3.130 TFLOPS FP32. Pixel rate is 502.5 GPixel/s versus 32.61 GPixel/s; texture rate is 1,968.0 GTexel/s versus 130.4 GTexel/s.

Where Each One Wins

The GTX 690 wins in historical context and legacy compatibility. Its DirectX 12 (11_0) support and Vulkan 1.2.175 make it usable in older software environments, and its 3x DVI outputs suit legacy multi-monitor setups. Its Geekbench Vulkan score of 16,675 and OpenCL score of 17,399 show it remains competitive in those specific synthetic workloads, even against modern cards in its percentile range. The 0.1% deltaPct against the AMD Radeon HD 7970M and 0.6% against the NVIDIA Tesla M4 indicate it still trades blows with near-contemporaries.

The RTX PRO 6000 Blackwell wins in every modern compute category. The 188 ray tracing cores and 752 tensor cores enable workloads the GTX 690 cannot handle at all. The 96 GB GDDR7 memory capacity is 48x larger, which is decisive for large dataset processing. The 1.79 TB/s bandwidth is roughly 9.3x higher, and the 126.0 TFLOPS FP32 is about 40x higher. The 1:1 FP16 ratio doubles the compute range for mixed-precision tasks. The PCIe 5.0 x16 interface doubles the bus bandwidth ceiling versus PCIe 3.0 x16. DirectX 12 Ultimate (12_2) and Vulkan 1.4 support cover current and near-future APIs.

In the benchmark percentile context, the GTX 690 holds a 1-percentile advantage (60th vs 59th), but this is likely an artifact of the test selection. The RTX PRO 6000 Blackwell's single 3DMark Steel Nomad DX12 score reflects a modern gaming/professional workload, whereas the GTX 690's scores come from compute and Vulkan synthetics. For professional rendering, AI inference, or large-scale visualization, the RTX PRO 6000 Blackwell's architectural advantages are decisive. For retro gaming or legacy CAD software that predates Blackwell, the GTX 690 might still serve, though its 2 GB memory limits are severe.

Specification Differences

The two cards differ in nearly every measurable specification. Process node: 28 nm vs 5 nm. Transistors: 3,540 million vs 92,200 million. Die size: 294 mm² vs 750 mm². Transistor density: 12.0M / mm² vs 122.9M / mm². Base clock: 915 MHz vs 1590 MHz. Boost clock: 1019 MHz vs 2617 MHz. Memory clock: 1502 MHz (6 Gbps effective) vs 1750 MHz (28 Gbps effective). Memory size: 2 GB vs 96 GB. Memory type: GDDR5 vs GDDR7. Bus width: 256 bit vs 512 bit. Bandwidth: 192.3 GB/s vs 1.79 TB/s.

Shading units: 1,536 vs 24,064. TMUs: 128 vs 752. ROPs: 32 vs 192. RT cores: none vs 188. Tensor cores: none vs 752. Pixel rate: 32.61 GPixel/s vs 502.5 GPixel/s. Texture rate: 130.4 GTexel/s vs 1,968.0 GTexel/s. FP32: 3.130 TFLOPS vs 126.0 TFLOPS. FP16: not specified vs 126.0 TFLOPS (1:1). TDP: 300 W vs 600 W. Power connectors: 2x 8-pin vs 1x 16-pin. Suggested PSU: 700 W vs 1000 W. Bus interface: PCIe 3.0 x16 vs PCIe 5.0 x16.

Physical dimensions: length 279 mm (11 inches) vs 304 mm (12 inches); height 111 mm (4.4 inches) vs 137 mm (5.4 inches); width 38 mm (1.5 inches) vs 40 mm (1.6 inches). Both are dual-slot. Display outputs: 3x DVI + 1x mini-DisplayPort 1.2 vs 4x DisplayPort 2.1b. DirectX: 12 (11_0) vs 12 Ultimate (12_2). OpenGL is identical at 4.6. Vulkan: 1.2.175 vs 1.4. Production status: End-of-life vs Active. Release date: 2012-05-02 vs 2025-03-17. Predecessor: GeForce 500 vs Workstation Ada. Successor: GeForce 700 vs none. Launch MSRP: 999 USD vs 8,565 USD.

Identical fields: manufacturer (NVIDIA), foundry (TSMC), slot width (dual-slot), and OpenGL version (4.6). Everything else differs.

FAQ

Q: Which card has more memory bandwidth?

A: The RTX PRO 6000 Blackwell has 1.79 TB/s bandwidth on a 512-bit bus with GDDR7, versus the GTX 690's 192.3 GB/s on a 256-bit bus with GDDR5. That is roughly 9.3x higher bandwidth.

Q: Does the GTX 690 support ray tracing?

A: No. The GTX 690 has no ray tracing cores listed. The RTX PRO 6000 Blackwell has 188 ray tracing cores.

Q: Which card has a higher boost clock?

A: The RTX PRO 6000 Blackwell boosts to 2617 MHz, while the GTX 690 boosts to 1019 MHz.

Q: What is the transistor count difference?

A: The RTX PRO 6000 Blackwell has 92,200 million transistors, versus 3,540 million on the GTX 690. That is approximately 26x more.

Q: Are the benchmark scores comparable?

A: No direct comparison is possible. The GTX 690 has Geekbench OpenCL (17,399) and Vulkan (16,675) scores, while the RTX PRO 6000 Blackwell has a 3DMark Steel Nomad DX12 score (16,408). They measure different workloads.

Q: Which card has a higher percentile ranking?

A: The GTX 690 sits at the 60th percentile among all GPUs, while the RTX PRO 6000 Blackwell sits at the 59th percentile. The difference is one percentile point.

The Verdict

The data dictates a clear split. For anyone running modern professional workloads—AI inference, ray-traced rendering, large-scale visualization—the RTX PRO 6000 Blackwell is the only defensible choice. Its 96 GB GDDR7 memory, 188 RT cores, 752 tensor cores, and 126.0 TFLOPS FP32 are not incremental improvements; they are categorical leaps over the GTX 690. The 1.79 TB/s bandwidth and PCIe 5.0 x16 interface ensure data movement keeps pace with compute. The 5 nm process and 122.9M / mm² density deliver efficiency that the 28 nm Kepler chip cannot approach.

For anyone running legacy software or hardware-specific benchmarks, the GTX 690 still holds a narrow edge in percentile ranking (60th vs 59th) and shows competitive scores in Geekbench OpenCL and Vulkan. Its 3x DVI outputs and older API support (DirectX 12 11_0, Vulkan 1.2.175) make it usable in vintage environments. However, the 2 GB GDDR5 memory and 32 ROPs are hard constraints that will bottleneck any modern workload.

The RTX PRO 6000 Blackwell is the superior card in every architectural metric except percentile rank, where it trails by a single point. The GTX 690 is a historical artifact—end-of-life, with a successor (GeForce 700) already released in 2013. The RTX PRO 6000 Blackwell is active and current, with no successor listed. The 8,565 USD launch MSRP versus the GTX 690's 999 USD reflects the massive capability gap. Choose the RTX PRO 6000 Blackwell for any task that demands current or future hardware support. Choose the GTX 690 only if legacy compatibility is the sole requirement.