NVIDIA GeForce GTX 650 Ti vs NVIDIA GeForce GTX 980 Comparison

# Head-to-Head Benchmarks

The benchmark data reveals a decisive performance gap in the two shared tests between the NVIDIA GeForce GTX 980 and the NVIDIA GeForce GTX 650 Ti. In Geekbench OpenCL, the GTX 980 scores 34,676 against the GTX 650 Ti’s 7,877, a staggering 340.2% advantage. That delta dwarfs the usual generational leap; it is a full-class separation. The GTX 980 delivers over four times the raw compute throughput in this workload, which stresses general-purpose GPU acceleration.

In Geekbench Vulkan, the margin narrows but remains emphatic. The GTX 980 posts 22,543, while the GTX 650 Ti manages 8,229, giving the newer card a 173.9% lead. Vulkan’s lower-level API reduces driver overhead, yet the GTX 980 still more than doubles the older card’s output. Interesting that the gap in Vulkan is roughly half the OpenCL gap — suggesting that the GTX 650 Ti’s Kepler architecture handles Vulkan’s explicit command buffers relatively better than it handles OpenCL’s more abstract execution model. Still, 173.9% is a landslide.

The head-to-head record is 2 wins for the GTX 980, 0 for the GTX 650 Ti. No benchmark in the shared set favors the older card. The wider benchmark suite reinforces this: the GTX 980’s average benchmark score across all tests is 8,167, versus 8,053 for the GTX 650 Ti. That is only a 1.4% difference in aggregate, which seems counterintuitive given the massive OpenCL and Vulkan deltas. The explanation lies in the composition of the averages — the GTX 650 Ti only has two benchmark entries (both Geekbench tests), while the GTX 980 has eleven, including several Passmark tests where it scores relatively modestly (e.g., Passmark DirectX 9 at 164, Passmark G3D at 11,095). The averages are not directly comparable across unequal test sets.

The percentile rankings tell a more nuanced story. The GTX 980 sits in the 43rd percentile of all GPUs, while the GTX 650 Ti is in the 42nd. For two cards separated by two generations and a 340% delta in one test, that percentile difference is surprisingly small. This suggests that the overall GPU landscape has shifted dramatically — modern integrated graphics and entry-level discrete cards have pulled the median upward, making both of these older products comparable in relative standing despite their massive performance gap. The GTX 980’s nearest rivals are the GTX 950M (delta 0.4%), Radeon R9 M360 (delta 0.5%), and GeForce 945M (delta 0.8%) — all mobile or low-end parts. The GTX 650 Ti’s nearest rivals include the GTX 650 Ti Boost (delta -0.2%) and Quadro P5000 (delta 0.2%), showing it clusters with a different tier entirely.

# Architecture Differences

The architectural chasm between these two cards is vast. The GTX 980 uses the GM204 chip built on Maxwell 2.0 architecture, part of the GeForce 900 generation. The GTX 650 Ti uses GK106S, a Kepler architecture chip from the GeForce 600 generation. Both are fabricated on TSMC’s 28 nm process, but that is where the manufacturing similarity ends. The GTX 980 packs 5,200 million transistors onto a 398 mm² die, achieving a transistor density of 13.1M per mm². The GTX 650 Ti has 2,540 million transistors on a 221 mm² die, with a density of 11.5M per mm². The GTX 980 crams more than twice the transistors into less than twice the die area, reflecting Maxwell 2.0’s more efficient logic design.

Compute resources diverge sharply. The GTX 980 fields 2,048 shading units, 128 texture mapping units, and 64 ROPs. The GTX 650 Ti has 768 shading units, 64 TMUs, and just 16 ROPs. That 2.67x advantage in shading units and 4x advantage in ROPs explains the pixel rate disparity: 77.82 GPixel/s for the GTX 980 versus 14.85 GPixel/s for the GTX 650 Ti. Texture rate follows suit at 155.6 GTexel/s versus 59.39 GTexel/s. Floating-point performance (FP32) is 4.981 TFLOPS for the GTX 980, compared to 1,425.4 GFLOPS (approximately 1.425 TFLOPS) for the GTX 650 Ti — a 3.5x raw compute advantage.

Memory architecture tells a similar story. The GTX 980 has 4 GB of GDDR5 on a 256-bit bus, yielding 224.4 GB/s of bandwidth. The GTX 650 Ti has 1,024 MB (1 GB) of GDDR5 on a 128-bit bus, producing just 86.40 GB/s. The GTX 980’s memory clock is 1753 MHz (7 Gbps effective), while the GTX 650 Ti runs at 1350 MHz (5.4 Gbps effective). The GTX 980 also supports more display outputs — 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2 — versus the GTX 650 Ti’s 2x DVI and 1x mini-HDMI 1.4a.

API support differs meaningfully. The GTX 980 supports DirectX 12 (12_1), while the GTX 650 Ti only reaches DirectX 12 (11_0) — a feature-level gap that affects advanced rendering techniques. OpenGL is identical at 4.6, but Vulkan support is newer on the GTX 980 (1.4 versus 1.2.175). The GTX 980’s base clock is 1127 MHz with a 1216 MHz boost; the GTX 650 Ti has no listed base or boost clock, only its memory clock.

# FAQ

Q: How much faster is the GTX 980 in OpenCL?

A: The GTX 980 scores 34,676 in Geekbench OpenCL, which is 340.2% higher than the GTX 650 Ti’s 7,877. This represents a four-fold performance advantage in general-purpose compute workloads.

Q: Does the GTX 650 Ti win any benchmark against the GTX 980?

A: No. In the two shared benchmarks (Geekbench OpenCL and Geekbench Vulkan), the GTX 980 wins both. The head-to-head record is 2 wins for the GTX 980 and 0 for the GTX 650 Ti.

Q: Why are the average benchmark scores so close despite the huge per-test deltas?

A: The GTX 980 has eleven benchmark entries, while the GTX 650 Ti has only two. The GTX 980’s average (8,167) includes many Passmark tests where it scores relatively low (e.g., 53 in DirectX 10, 46 in DirectX 12), pulling its average down. The GTX 650 Ti’s average (8,053) is based solely on its two Geekbench scores.

Q: What are the transistor counts of these two GPUs?

A: The GTX 980 contains 5,200 million transistors on a 398 mm² die. The GTX 650 Ti has 2,540 million transistors on a 221 mm² die. Both use TSMC’s 28 nm process node.

Q: Which card has more memory bandwidth?

A: The GTX 980 offers 224.4 GB/s across a 256-bit bus, compared to the GTX 650 Ti’s 86.40 GB/s across a 128-bit bus. The GTX 980 also has 4 GB of memory versus 1 GB on the GTX 650 Ti.

Q: Do both cards support DirectX 12?

A: Yes, but at different feature levels. The GTX 980 supports DirectX 12 (12_1), while the GTX 650 Ti supports DirectX 12 (11_0). Both support OpenGL 4.6, but the GTX 980 has a newer Vulkan version (1.4 versus 1.2.175).

# Specification Differences

| Specification | NVIDIA GeForce GTX 980 | NVIDIA GeForce GTX 650 Ti |

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

| Architecture | Maxwell 2.0 | Kepler |

| Generation | GeForce 900 | GeForce 600 |

| Chip | GM204 | GK106S |

| Transistors | 5,200 million | 2,540 million |

| Die Size | 398 mm² | 221 mm² |

| Transistor Density | 13.1M / mm² | 11.5M / mm² |

| Base Clock | 1127 MHz | Not listed |

| Boost Clock | 1216 MHz | Not listed |

| Memory Clock | 1753 MHz (7 Gbps effective) | 1350 MHz (5.4 Gbps effective) |

| Memory Size | 4 GB | 1024 MB |

| Memory Bus Width | 256 bit | 128 bit |

| Memory Bandwidth | 224.4 GB/s | 86.40 GB/s |

| Shading Units | 2048 | 768 |

| TMUs | 128 | 64 |

| ROPs | 64 | 16 |

| Pixel Rate | 77.82 GPixel/s | 14.85 GPixel/s |

| Texture Rate | 155.6 GTexel/s | 59.39 GTexel/s |

| FP32 Performance | 4.981 TFLOPS | 1,425.4 GFLOPS |

| TDP | 165 W | 110 W |

| Slot Width | Dual-slot | Single-slot |

| Power Connectors | 2x 6-pin | 1x 6-pin |

| Suggested PSU | 450 W | 300 W |

| Display Outputs | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.2 | 2x DVI, 1x mini-HDMI 1.4a |

| DirectX | 12 (12_1) | 12 (11_0) |

| Vulkan | 1.4 | 1.2.175 |

| Length | 267 mm (10.5 inches) | 145 mm (5.7 inches) |

| Height | 111 mm (4.4 inches) | Not listed |

| Width | 40 mm (1.6 inches) | Not listed |

| Release Date | 2014-09-18 | 2012-10-08 |

| Predecessor | GeForce 700 | GeForce 500 |

| Successor | GeForce 10 | GeForce 700 |

# Where Each One Wins

The GTX 980 wins decisively in every shared benchmark category. Its 340.2% OpenCL advantage speaks to compute-heavy workloads — GPGPU tasks, scientific simulation, and machine learning inference. The 173.9% Vulkan lead covers modern game engines that leverage this API, as well as any Vulkan-based compute applications. The GTX 980’s 4 GB memory frame buffer versus 1 GB means it can handle larger textures, higher resolution assets, and multi-monitor setups without exhausting VRAM. Its 64 ROPs deliver 77.82 GPixel/s of fill rate, which matters for high-resolution rendering and heavy post-processing effects. The Maxwell 2.0 architecture’s DirectX 12 (12_1) feature level enables hardware features like conservative rasterization and rasterizer-ordered views that the Kepler card cannot access.

The GTX 650 Ti wins in precisely one measurable category: physical footprint and power draw. Its single-slot design and 145 mm length (5.7 inches) allow installation in compact cases where the GTX 980’s 267 mm (10.5 inches) dual-slot body will not fit. Its 110 W TDP versus 165 W means less heat output and a lower power supply requirement — the suggested PSU drops from 450 W to 300 W. The GTX 650 Ti also uses just one 6-pin power connector versus the GTX 980’s two. For a small form factor build or a legacy system with limited PSU capacity, the GTX 650 Ti is the only physically viable option. Its 1 GB memory, while small, is sufficient for older DirectX 9/10-era games at 1080p with reduced settings. The GTX 650 Ti’s 42nd percentile ranking versus the GTX 980’s 43rd shows that in the broader GPU ecosystem, these cards are not as far apart as raw specs suggest — but in any direct comparison, the GTX 980 dominates.

# The Verdict

The data is unambiguous: the NVIDIA GeForce GTX 980 is the superior card in every performance metric measured. A 340.2% lead in OpenCL and a 173.9% lead in Vulkan are not marginal improvements — they represent a generational chasm. Anyone requiring compute performance, modern API support, or high-resolution gaming should select the GTX 980 without hesitation. Its 4 GB memory, 224.4 GB/s bandwidth, and 4.981 TFLOPS FP32 throughput place it in a different performance tier entirely.

The GTX 650 Ti has a narrow but real constituency. Its single-slot, 145 mm profile and 110 W TDP make it suitable for legacy systems, HTPC builds, or any chassis where the GTX 980’s dual-slot, 267 mm length will not physically fit. The 300 W suggested PSU requirement and single 6-pin connector also ease installation in older power supplies. For users constrained by physical space and power delivery rather than performance, the GTX 650 Ti remains a functional option — but it is a card for specific constraints, not for general-purpose use.

The percentile data complicates the picture. Both cards sit within one percentile point of each other (43rd vs 42nd), and their nearest rivals include mobile and low-end parts. This indicates that neither card is competitive in the modern GPU landscape — the GTX 980’s enormous lead over the GTX 650 Ti does not translate into relevance against contemporary hardware. The benchmark averages (8,167 vs 8,053) are deceptively close, but the unequal test counts make direct comparison unreliable. The verdict: choose the GTX 980 for performance, choose the GTX 650 Ti only if physical constraints force your hand. The GTX 980 wins 2-0 in head-to-head tests, and no data point suggests the GTX 650 Ti can close that gap.