NVIDIA GeForce GTX 650 Ti Boost vs NVIDIA GeForce GTX 680M Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 650 Ti Boost records an average benchmark score of 8067, placing it in the 42nd percentile of all GPUs. The NVIDIA GeForce GTX 680M scores 7023 on average, sitting in the 39th percentile.

Q: How large is the performance gap between the two cards in the head-to-head tests?

A: The GTX 650 Ti Boost wins both recorded head-to-head tests. In Geekbench Metal it scores 4858 versus 4815 for the GTX 680M, a 0.9% advantage. In Geekbench OpenCL it scores 9302 versus 9230, a 0.8% lead.

Q: What are the nearest rivals to each card in the database?

A: For the GTX 650 Ti Boost, the closest competitors are the NVIDIA GRID K2 (avg 8080, -0.2%), GeForce GTX 650 Ti (8053, +0.2%), GeForce GTX 880M (8040, +0.3%), and Quadro P5000 (8039, +0.3%). For the GTX 680M, the nearest rivals are the NVIDIA T600 (7035, -0.2%), AMD Radeon R5 M240 (6975, +0.7%), GeForce GTX 675M (6946, +1.1%), and GeForce GTX 970 (7157, -1.9%).

Q: Which card has the higher transistor count and larger die?

A: The GTX 680M uses the GK104 chip with 3,540 million transistors on a 294 mm² die. The GTX 650 Ti Boost uses GK106 with 2,540 million transistors on a 221 mm² die. Both are built on TSMC's 28 nm process, but the GTX 680M has a slightly higher transistor density at 12.0M per mm² versus 11.5M per mm².

Q: Do both cards support the same graphics APIs?

A: Yes, both support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175, indicating identical API feature sets despite their different hardware configurations.

Q: What is the production status and release timeline?

A: Both are end-of-life products. The GTX 650 Ti Boost launched on 2013-03-25 with a launch MSRP of 169 USD, while the GTX 680M launched earlier on 2012-06-03. The GTX 650 Ti Boost succeeds the GeForce 500 series and precedes GeForce 700, while the GTX 680M follows GeForce 500M and precedes GeForce 700M.

Where Each One Wins

The GTX 650 Ti Boost wins in raw computational throughput per the benchmark data, taking both head-to-head tests. Its Geekbench OpenCL score of 9302 exceeds the GTX 680M's 9230, and its Metal score of 4858 beats 4815. This desktop-oriented card also holds a significant advantage in clock speeds, with a base clock of 980 MHz and boost up to 1032 MHz, compared to the GTX 680M's 719 MHz base and 758 MHz boost. The higher clocks allow the GTX 650 Ti Boost to achieve superior texture and pixel rates in specific workloads: it delivers 66.05 GTexel/s versus 84.90 GTexel/s for the GTX 680M? Actually, the GTX 680M has the higher texture rate. The GTX 650 Ti Boost wins in memory bandwidth at 144.2 GB/s over the GTX 680M's 115.2 GB/s, despite the latter having a wider 256-bit bus.

The GTX 680M wins in sheer shader and geometry resources. It packs 1344 shading units, 112 texture mapping units, and 32 ROPs, versus 768 shading units, 64 TMUs, and 24 ROPs for the GTX 650 Ti Boost. This translates to a higher FP32 compute peak of 2.038 TFLOPS compared to 1.585 TFLOPS. The GTX 680M also offers double the memory capacity at 4 GB versus 2 GB, which becomes relevant for large textures and high-resolution workloads. Its pixel rate of 21.22 GPixel/s exceeds the 16.51 GPixel/s of the GTX 650 Ti Boost.

The GTX 680M is designed for portability with a 100 W TDP and MXM module form factor, whereas the GTX 650 Ti Boost is a dual-slot desktop card requiring a 300 W suggested PSU and a 6-pin power connector. The GTX 680M has no power connectors and relies on the host system, making it the preferred option for mobile or compact systems.

Architecture Differences

Both GPUs share the Kepler architecture, but they use different chips. The GTX 650 Ti Boost is built on GK106, while the GTX 680M uses GK104. The GK104 is a larger, more complex chip with 3,540 million transistors on a 294 mm² die, compared to GK106's 2,540 million transistors on a 221 mm² die. The transistor density is slightly higher on GK104 at 12.0M per mm² versus 11.5M per mm², indicating a more efficient packing of logic units.

The GTX 680M's GK104 features a 256-bit memory bus, which is wider than the 192-bit bus of the GTX 650 Ti Boost's GK106. However, the GTX 650 Ti Boost compensates with faster memory clocks: 1502 MHz (6 Gbps effective) versus 900 MHz (3.6 Gbps effective) for the GTX 680M. This results in the GTX 650 Ti Boost having higher memory bandwidth (144.2 GB/s) despite the narrower bus.

The core configuration differs significantly. The GTX 680M has 1344 shading units, 112 TMUs, and 32 ROPs, while the GTX 650 Ti Boost has 768 shading units, 64 TMUs, and 24 ROPs. This nearly doubles the shader count and TMU count on the GTX 680M, which explains its higher theoretical compute figures. The GTX 650 Ti Boost's advantage lies in its clock speed, which partially offsets the raw resource deficit.

Both support identical API versions: DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. Neither has dedicated ray tracing or tensor cores, consistent with the Kepler generation.

Specification Differences

| Specification | GTX 650 Ti Boost | GTX 680M |

|---|---|---|

| Chip | GK106 | GK104 |

| Generation | GeForce 600 | GeForce 600M |

| Process Node | 28 nm | 28 nm |

| Transistors | 2,540 million | 3,540 million |

| Die Size | 221 mm² | 294 mm² |

| Base Clock | 980 MHz | 719 MHz |

| Boost Clock | 1032 MHz | 758 MHz |

| Memory Clock | 1502 MHz (6 Gbps effective) | 900 MHz (3.6 Gbps effective) |

| Memory Size | 2 GB | 4 GB |

| Memory Type | GDDR5 | GDDR5 |

| Memory Bus | 192 bit | 256 bit |

| Memory Bandwidth | 144.2 GB/s | 115.2 GB/s |

| Shading Units | 768 | 1344 |

| TMUs | 64 | 112 |

| ROPs | 24 | 32 |

| Pixel Rate | 16.51 GPixel/s | 21.22 GPixel/s |

| Texture Rate | 66.05 GTexel/s | 84.90 GTexel/s |

| FP32 | 1.585 TFLOPS | 2.038 TFLOPS |

| TDP | 134 W | 100 W |

| Slot Width | Dual-slot | MXM Module |

| Power Connectors | 1x 6-pin | None |

| Suggested PSU | 300 W | None |

| Bus Interface | PCIe 3.0 x16 | MXM-B (3.0) |

| Display Outputs | 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 | Portable Device Dependent |

| Dimensions | 241 mm (9.5 inches) | Not specified |

| Launch MSRP | 169 USD | None |

Head-to-Head Benchmarks

The database records two direct comparisons between these GPUs, and the GTX 650 Ti Boost emerges victorious in both. In Geekbench Metal, the GTX 650 Ti Boost scores 4858 against 4815 for the GTX 680M, a margin of 0.9%. This is a narrow but consistent lead. In Geekbench OpenCL, the gap widens slightly: the GTX 650 Ti Boost scores 9302 versus 9230, a 0.8% advantage.

These results are notable because the GTX 680M has substantially more shader units and a higher theoretical FP32 peak (2.038 TFLOPS versus 1.585 TFLOPS). The GTX 650 Ti Boost compensates with much higher clock speeds, running at 980 MHz base and 1032 MHz boost compared to 719 MHz and 758 MHz. The higher memory bandwidth of the GTX 650 Ti Boost (144.2 GB/s) also contributes to its lead, as memory-bound workloads benefit from the faster GDDR5 at 6 Gbps effective.

The GTX 680M's larger memory pool (4 GB versus 2 GB) and wider bus (256 bit versus 192 bit) do not translate into benchmark wins in these tests. Its lower memory clock (3.6 Gbps effective) reduces bandwidth to 115.2 GB/s, which may bottleneck its otherwise faster compute hardware.

When placed against their respective nearest rivals, both cards sit in a similar competitive position relative to their peer groups. The GTX 650 Ti Boost is within 0.3% of three competitors (GRID K2, GTX 650 Ti, GTX 880M, Quadro P5000), while the GTX 680M is within 1.1% of the T600, R5 M240, and GTX 675M, but trails the GTX 970 by 1.9%. The average scores differ by about 13% between the two cards (8067 versus 7023), which aligns with the GTX 650 Ti Boost's higher percentile ranking (42nd versus 39th).

The data suggests that for applications sensitive to clock speed and memory bandwidth, the GTX 650 Ti Boost is the stronger choice. For workloads that scale with shader count and memory capacity, the GTX 680M has theoretical advantages, but the measured benchmarks do not reflect this, likely due to the significant clock and bandwidth differences. Both cards remain end-of-life products, so these results reflect their historical performance rather than current relevance.