NVIDIA GeForce GTX 690
NVIDIA GeForce GTX 690 in 2025: A Retrospective on the Legend of the Dual-GPU Era
Introduction
Released in 2012, the NVIDIA GeForce GTX 690 became a symbol of an era when multi-chip solutions dominated the battle for maximum performance. Thirteen years later, this model still holds cult status among enthusiasts, but its place in the modern world requires reassessment. In this article, we will explore what the GTX 690 can still surprise us with in 2025 and where its limitations become critical.
Architecture and Key Features
Kepler Architecture: Double Trouble
The GTX 690 is built on the Kepler architecture (GK104 model) and is unique because it combines two GPUs on a single board. The manufacturing process is 28 nm, which was cutting-edge in 2012. Each chip contains 1536 CUDA cores, totaling 3072 stream processors.
Lack of Modern Technologies
The GTX 690 was created before the era of ray tracing and AI algorithms. It does not support RTX, DLSS, or FidelityFX. However, at its time, NVIDIA introduced TXAA (Temporal Anti-Aliasing) and Adaptive VSync—features that enhance anti-aliasing and frame synchronization.
Dual-GPU System Characteristics
The card uses SLI technology to combine the two chips. However, performance scaling rarely reached 100%, and support for multi-GPU setups in games remains unstable even in 2025. Many modern titles are not optimized for such solutions at all.
Memory: Potential and Limitations
GDDR5 and 4 GB: Insufficient for Modern Tasks
Each GPU is equipped with 2 GB of GDDR5 memory with a 256-bit bus (totaling 4 GB). The bandwidth is 384 GB/s (192 GB/s per chip). For games from the 2010s, this was sufficient, but in 2025, even 8 GB is the minimum standard. For instance, in "Cyberpunk 2077" at 1080p, 6-8 GB of VRAM is required.
Buffer Issues
The limited memory capacity and its division between chips create a bottleneck in modern games. High-resolution textures and complex scenes lead to frame rate drops due to buffer overflow.
Gaming Performance: Nostalgia with Caveats
1080p: Acceptable Only for Older Titles
In less demanding games like "CS:GO" or "Dota 2," the GTX 690 achieves 100-150 FPS on medium settings. However, in "Elden Ring" or "Starfield," even on low presets, the frame rate barely reaches 30 FPS.
1440p and 4K: Not for the Faint of Heart
The card is not designed for resolutions above Full HD. Attempting to run "Hogwarts Legacy" at 1440p results in 15-20 FPS, while 4K turns into a slideshow.
Ray Tracing: Not Available
The absence of hardware support for RT cores makes ray tracing impossible, even through third-party mods.
Professional Tasks: Time Has Taken Its Toll
Video Editing and Rendering
Thanks to CUDA, the GTX 690 can handle basic editing in DaVinci Resolve or Premiere Pro, but rendering 4K video takes 3-4 times longer compared to modern GPUs (e.g., RTX 4060).
3D Modeling
In Blender or Maya, the card shows modest results. Projects with high-polygon models (>1 million polygons) cause lag.
Scientific Calculations
Support for CUDA and OpenCL allows the GTX 690 to be used for simple simulations, but energy efficiency is extremely low. For comparison, a single RTX 4090 chip outperforms the GTX 690 in FP32 calculations by 20 times.
Power Consumption and Heat Output
TDP 300 W: Prepare for Electric Bills
The GTX 690 requires powerful cooling and a high-quality PSU. Its TDP is 300 W, but peak load can reach 350 W.
Cooling Recommendations
- A case with at least 3 fans: 2 for intake and 1 for exhaust.
- Replacing thermal paste is essential for units that haven't been serviced for years.
- An ideal room temperature is below 25°C. At 30°C, the GPU can heat up to 85°C.
Comparison with Competitors
AMD Radeon HD 7990: The Main Rival
The dual-chip HD 7990 (2x Tahiti XT) competed with the GTX 690 in 2013. Today, both cards are equally outdated, but the AMD solution suffers more from frame drops due to less efficient drivers.
Modern Counterparts: RTX 3050
Even the budget RTX 3050 (8 GB GDDR6) outperforms the GTX 690 in performance, consuming only 130 W.
Practical Advice
Power Supply: Minimum 600 W
Even if your system is modest, choose a PSU with some margin. Recommended models include Corsair CX650M or Be Quiet! Pure Power 12 M 600W.
Platform Compatibility
- PCIe 3.0 x16: the card works in PCIe 4.0/5.0 slots, but with no speed increase.
- Windows 10/11: NVIDIA ceased support for the GTX 600 series in 2021. The last stable driver version is 472.12.
Driver Nuances
- Modern games may not run due to outdated API (DirectX 12 Ultimate is not supported).
- The enthusiast community releases modified drivers, but their stability is questionable.
Pros and Cons
Pros:
- Historical Value: Iconic design with lighting and aluminum casing.
- Uniqueness: One of the last dual-GPU solutions from NVIDIA.
- SLI Support for compatibility with other Kepler cards.
Cons:
- Outdated Architecture: No RTX, no DLSS, low memory capacity.
- High Power Consumption.
- Limited game and driver support.
Final Conclusion: Who Is the GTX 690 For in 2025?
This graphics card is suitable for:
1. Collectors and retro enthusiasts building PCs in the 2010s style.
2. Owners of older systems where upgrading is impossible (e.g., LGA 1155 platform).
3. Experimenters willing to struggle with drivers to run classics like "Crysis 3" on ultra settings.
For modern gaming and professional tasks, the GTX 690 is not suitable. Its domain is nostalgia and niche applications. If you're looking for power around $150, consider a used GTX 1660 Super or RX 6600.
Postscript
The NVIDIA GeForce GTX 690 stands as a monument to an era when engineers competed in multi-chip solutions. Today, it reminds us how quickly technology evolves, allowing us to delve into the past where every frame was hard-won. But for real work and gaming in 2025, choose something modern—like the RTX 4060 or RX 7600.