AMD Radeon Vega 11 vs NVIDIA GeForce GTX 660 Ti Comparison
AMD Radeon Vega 11
GeForce GTX 660 Ti
PERFORMANCE BENCHMARKS
Analysis: AMD Radeon Vega 11 vs NVIDIA GeForce GTX 660 Ti
The NVIDIA GeForce GTX 660 Ti and AMD Radeon Vega 11 represent two very different approaches to graphics processing: a dedicated, end-of-life discrete card from 2012 versus an integrated GPU (IGP) from 2019. The data shows that the GTX 660 Ti wins both head-to-head benchmark comparisons, but the Vega 11 has its own compelling strengths in API support and efficiency. The verdict is straightforward: the GTX 660 Ti is the clear choice for raw performance in existing applications, while the Vega 11 is the better pick for a low-power, modern-featured system where a discrete card isn't an option.
The Verdict
Benchmark results indicate that the NVIDIA GeForce GTX 660 Ti is the performance winner. In the Geekbench OpenCL test, it scores 15,113 against the Vega 11's 13,392, a 12.9% lead. The gap widens in the Geekbench Vulkan test, where the GTX 660 Ti scores 15,012 versus 12,273, a 22.3% advantage. If your priority is compute performance in these cross-platform APIs, the GTX 660 Ti is the only choice from this pair.
However, the AMD Radeon Vega 11 is not without merit. Its average benchmark score of 14,352 places it in the 56th percentile of all GPUs, just one percentile behind the GTX 660 Ti's 57th percentile. The Vega 11 also supports DirectX 12 (12_1) and Vulkan 1.3, both newer than the GTX 660 Ti's DirectX 12 (11_0) and Vulkan 1.2.175. For a builder assembling a system with no room for a discrete card, the Vega 11 offers a modern feature set at a fraction of the power draw. The GTX 660 Ti is for those who need the extra performance headroom; the Vega 11 is for those who value efficiency and API modernity above raw scores.
Architecture Differences
The two GPUs are built on fundamentally different architectures and processes. The GTX 660 Ti uses NVIDIA's Kepler architecture on a 28 nm TSMC process, while the Vega 11 uses AMD's GCN 5.0 architecture on a 12 nm GlobalFoundries process. The Vega 11's chip, codenamed Picasso, packs 4,940 million transistors into a 210 mm² die, resulting in a transistor density of 23.5M per mm². The GTX 660 Ti's GK104 chip contains 3,540 million transistors on a larger 294 mm² die, yielding a density of 12.0M per mm².
The compute resources differ significantly. The GTX 660 Ti has 1,344 shading units, 112 texture mapping units (TMUs), and 24 raster operation units (ROPs). The Vega 11 has 704 shading units, 44 TMUs, and only 8 ROPs. This explains the GTX 660 Ti's higher pixel rate of 27.44 GPixel/s versus 11.20 GPixel/s for the Vega 11, and its texture rate of 109.8 GTexel/s versus 61.60 GTexel/s. The GTX 660 Ti also produces 2.634 TFLOPS of FP32 compute, while the Vega 11 produces 1.971 TFLOPS. Notably, the Vega 11 does have an FP16 capability of 3.942 TFLOPS (2:1), a feature the GTX 660 Ti lacks entirely.
Memory architecture is another major divider. The GTX 660 Ti has dedicated 2 GB of GDDR5 memory on a 192-bit bus, delivering 144.2 GB/s of bandwidth. The Vega 11 uses System Shared memory, meaning its bandwidth is "System Dependent" and its bus width is not fixed. This makes the GTX 660 Ti's memory performance predictable and consistent, whereas the Vega 11's performance will vary with the host system's memory configuration.
Head-to-Head Benchmarks
The two Geekbench tests tell a clear story of NVIDIA's dominance in this pairing. In Geekbench OpenCL, the GTX 660 Ti scores 15,113 against the Vega 11's 13,392. That 12.9% delta is significant and reflects the GTX 660 Ti's higher shading unit count and dedicated memory bandwidth. In Geekbench Vulkan, the result is even more lopsided: 15,012 versus 12,273, a 22.3% delta. The GTX 660 Ti's lead in Vulkan suggests its architecture handles this modern API better relative to the Vega 11, despite the latter's newer Vulkan 1.3 support.
Looking at the broader benchmark landscape, the GTX 660 Ti's average score of 15,063 puts it in the 57th percentile of all GPUs. Its nearest rival, the AMD Radeon Pro 560X, scores 15,082, a mere 0.1% higher. The Vega 11's average of 14,352 places it in the 56th percentile, with its nearest rival being the NVIDIA GeForce GTX TITAN at 14,373, also a 0.1% difference. These percentile scores indicate that while the GTX 660 Ti is faster, both GPUs sit in a similar performance tier relative to the entire GPU landscape. The practical difference between the two is a 4.7% gap in average benchmark scores, which is noticeable but not transformative.
Specification Differences
The specification sheets reveal stark contrasts beyond performance. The GTX 660 Ti is a dual-slot card requiring 2x 6-pin power connectors and a 450 W suggested PSU, with a TDP of 150 W. The Vega 11 is an IGP with no slot width, no power connectors, and a TDP of just 15 W. This makes the Vega 11 an order of magnitude more power-efficient in terms of rated TDP.
Clock speeds differ as well: the GTX 660 Ti runs at a base of 915 MHz with a boost of 980 MHz, while the Vega 11 has a base of 300 MHz and a boost of 1,400 MHz. The Vega 11's higher boost clock partially compensates for its fewer shading units. Memory clocks are not comparable, as the GTX 660 Ti has a fixed 1502 MHz (6 Gbps effective) GDDR5 speed, while the Vega 11's memory speed is system-dependent.
The GTX 660 Ti offers fixed display outputs of 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. The Vega 11's display outputs are motherboard-dependent, meaning flexibility depends on the host system. The GTX 660 Ti uses a PCIe 3.0 x16 interface, while the Vega 11 uses an IGP bus interface. The GTX 660 Ti has a launch MSRP of 299 USD; the Vega 11 has no launch MSRP listed.
FAQ
Q: Which GPU has higher raw compute performance?
A: The GTX 660 Ti produces 2.634 TFLOPS of FP32 compute, versus 1.971 TFLOPS for the Vega 11. The GTX 660 Ti also has higher pixel and texture rates: 27.44 GPixel/s and 109.8 GTexel/s, respectively, compared to 11.20 GPixel/s and 61.60 GTexel/s for the Vega 11.
Q: How do their benchmark scores compare?
A: The GTX 660 Ti wins both head-to-head tests. In Geekbench OpenCL, it scores 15,113 versus 13,392, a 12.9% lead. In Geekbench Vulkan, it scores 15,012 versus 12,273, a 22.3% lead. Average benchmark scores are 15,063 for the GTX 660 Ti and 14,352 for the Vega 11.
Q: What are the power requirements for each?
A: The GTX 660 Ti has a TDP of 150 W, requires 2x 6-pin power connectors, and needs a suggested 450 W PSU. The Vega 11 has a TDP of only 15 W, requires no power connectors, and has no suggested PSU listed because it is an integrated GPU.
Q: Which GPU has better DirectX and Vulkan support?
A: The Vega 11 supports DirectX 12 (12_1) and Vulkan 1.3, which are newer than the GTX 660 Ti's DirectX 12 (11_0) and Vulkan 1.2.175. Both support OpenGL 4.6.
Q: How do their memory configurations differ?
A: The GTX 660 Ti has 2 GB of dedicated GDDR5 memory on a 192-bit bus with 144.2 GB/s bandwidth. The Vega 11 uses System Shared memory, with its bandwidth described as "System Dependent" and its bus width not fixed.
Q: What are the physical size considerations?
A: The GTX 660 Ti is a dual-slot card measuring 241 mm (9.5 inches) in length. The Vega 11 is an IGP with no physical dimensions listed, as it integrates into the motherboard.
Where Each One Wins
The GTX 660 Ti wins in every measured benchmark category. It is the obvious pick for any workload that relies on OpenCL or Vulkan compute, as shown by its 12.9% and 22.3% leads respectively. Its dedicated GDDR5 memory with 144.2 GB/s bandwidth ensures consistent performance regardless of system RAM speed. The 1,344 shading units and 112 TMUs give it a significant edge in texture-heavy and shader-bound tasks. If you have a motherboard with a PCIe 3.0 x16 slot and can accommodate a dual-slot, 150 W card, the GTX 660 Ti is the stronger performer.
The AMD Radeon Vega 11 wins on efficiency and platform integration. Its 15 W TDP is one-tenth that of the GTX 660 Ti, and it requires no power connectors or dedicated PSU headroom. It supports newer API versions with DirectX 12 (12_1) and Vulkan 1.3, which may matter for future software compatibility. The Vega 11 also has FP16 compute capability at 3.942 TFLOPS, a feature absent from the GTX 660 Ti. In a small form factor or office PC where a discrete card is impractical, the Vega 11 provides a usable graphics solution with modern features. Its 56th percentile standing, just one point below the GTX 660 Ti, shows it is not far behind in overall performance despite its integrated nature.