NVIDIA GeForce GT 745M vs NVIDIA GeForce GTX 460M Comparison

The NVIDIA GeForce GTX 460M and the NVIDIA GeForce GT 745M represent two distinct generations of mobile graphics, separated by a shift in process technology and architectural philosophy. The benchmark data presents a clear, if narrow, picture: the older Fermi-based GTX 460M holds a decisive lead in the single shared compute test, while the newer Kepler-based GT 745M offers a broader feature set and more memory. This analysis digs into what those differences mean for a potential buyer looking at end-of-life mobile GPUs.

Head-to-Head Benchmarks

The only common benchmark between these two GPUs is the Geekbench OpenCL test, and it paints a definitive picture. The NVIDIA GeForce GTX 460M scores 4282, while the NVIDIA GeForce GT 745M trails with a score of 3580. This results in a deltaPct of 19.6, meaning the GTX 460M is a substantial 19.6% faster than the GT 745M in this specific compute workload. The data shows that despite being from an older generation, the GTX 460M's raw compute throughput is significantly higher.

To contextualize this win, the GTX 460M's score places it in the 25th percentile of all GPUs, with an average benchmark score of 4282. Its nearest rivals include the AMD FirePro W2100 (4295, a -0.3% delta) and the AMD Radeon Vega 3 (4268, a +0.3% delta). This means it is essentially tied with those parts, but it is also interestingly close to the NVIDIA GeForce RTX 4070 GDDR6 (4335, a -1.2% delta), which is a much newer and more powerful card in other contexts. This suggests the GTX 460M holds its own in this specific synthetic test against a wide range of hardware.

On the other side, the GT 745M's score of 3580 in Geekbench OpenCL is its lowest result across its three recorded benchmarks. Its other scores are higher: 2777 in Geekbench Metal and 5502 in Geekbench Vulkan. The GT 745M's average score across all its benchmarks is 3953, and it sits in the 23rd percentile of all GPUs. Its nearest rivals are the AMD Radeon R5 M420 (3956, a -0.1% delta) and the NVIDIA GeForce 830M (3957, a -0.1% delta), showing it is directly competitive with those entry-level parts. The GTX 460M's 19.6% lead in OpenCL is a clear victory, but the GT 745M's ability to run Vulkan and Metal tests suggests a more modern software stack that the older card simply cannot access.

Architecture Differences

The core difference lies in their architectural generations. The GTX 460M is built on the Fermi architecture, using the GF106 chip, while the GT 745M uses the Kepler architecture with the GK107 chip. This is a fundamental shift in design philosophy. Fermi was known for its high compute throughput but also high power consumption, whereas Kepler was designed to be more efficient per watt.

The manufacturing process highlights this evolution. The GTX 460M is fabricated on a 40 nm process node at TSMC, while the GT 745M uses a smaller and more advanced 28 nm node, also from TSMC. This process shrink allows the GT 745M to pack more transistors into a smaller space. The GTX 460M has a die size of 238 mm² and houses 1,170 million transistors, resulting in a transistor density of 4.9M / mm². In contrast, the GT 745M has a smaller die size of 118 mm² but contains more transistors at 1,270 million, leading to a significantly higher transistor density of 10.8M / mm². The data shows the GT 745M is more than twice as dense, a direct result of the newer process node.

Despite having fewer transistors, the GTX 460M is configured with fewer but potentially more powerful cores. The GTX 460M has 192 shading units, 32 TMUs, and 24 ROPs. The GT 745M, while having more transistors, has 384 shading units but only 32 TMUs and 16 ROPs. This configuration is interesting: the GT 745M has double the shading units but fewer ROPs. This leads to different theoretical limits. The GTX 460M achieves a pixel rate of 5.400 GPixel/s and a texture rate of 21.60 GTexel/s, both higher than the GT 745M's 4.392 GPixel/s and 17.57 GTexel/s. In terms of raw floating-point performance (FP32), the GTX 460M is also ahead with 518.4 GFLOPS versus the GT 745M's 421.6 GFLOPS.

The Verdict

The data points to a clear, albeit nuanced, verdict. For pure compute performance as measured by Geekbench OpenCL, the NVIDIA GeForce GTX 460M is the superior part, offering a 19.6% performance advantage. Its higher FP32 throughput and pixel/texture rates indicate it is better suited for tasks that rely on raw mathematical processing and fill-rate. If your benchmark is the sole metric, the GTX 460M wins.

However, the GT 745M is not without its merits. It is a more modern part with a higher transistor count and density, and it features a broader set of supported APIs. The GT 745M supports Vulkan 1.2.175, a modern graphics API that the GTX 460M does not support at all. It also has a larger memory pool of 2 GB compared to the GTX 460M's 1536 MB. For a user concerned with software compatibility and running applications that leverage newer APIs, the GT 745M is the more future-proof choice, even if it is slower in the one test where they overlap. The GTX 460M is also a higher power draw, with a TDP of 50 W compared to the GT 745M's 45 W. The verdict depends on the priority: raw speed versus modern feature support.

Specification Differences

The following table outlines the key specifications where these two GPUs differ, based strictly on the provided data.

| Specification | NVIDIA GeForce GTX 460M | NVIDIA GeForce GT 745M |

| :--- | :--- | :--- |

| Architecture | Fermi | Kepler |

| Chip | GF106 | GK107 |

| Generation | GeForce 400M | GeForce 700M |

| Process Node | 40 nm | 28 nm |

| Transistors | 1,170 million | 1,270 million |

| Die Size | 238 mm² | 118 mm² |

| Transistor Density | 4.9M / mm² | 10.8M / mm² |

| Memory Speed | 2.5 Gbps effective | 4 Gbps effective |

| Memory Size | 1536 MB | 2 GB |

| Memory Bus Width | 192 bit | 128 bit |

| Memory Bandwidth | 60.00 GB/s | 64.00 GB/s |

| Shading Units | 192 | 384 |

| ROPs | 24 | 16 |

| Pixel Rate | 5.400 GPixel/s | 4.392 GPixel/s |

| Texture Rate | 21.60 GTexel/s | 17.57 GTexel/s |

| FP32 Performance | 518.4 GFLOPS | 421.6 GFLOPS |

| TDP | 50 W | 45 W |

| Slot Width | MXM Module | IGP |

| Bus Interface | PCIe 2.0 x16 | PCIe 3.0 x16 |

| Vulkan API | None | 1.2.175 |

| Release Date | 2010-09-02 | 2013-03-31 |

FAQ

Q: Which GPU is faster in the Geekbench OpenCL benchmark?

A: The NVIDIA GeForce GTX 460M is faster, scoring 4282 compared to the NVIDIA GeForce GT 745M's 3580, a difference of 19.6%.

Q: Does the newer GT 745M support modern graphics APIs?

A: Yes, the GT 745M supports Vulkan 1.2.175. The GTX 460M does not have a listed Vulkan API version, though both support DirectX 12 (11_0) and OpenGL 4.6.

Q: How does the memory configuration differ between the two?

A: The GT 745M has a larger memory capacity of 2 GB, while the GTX 460M has 1536 MB. The GT 745M also has a higher effective memory speed of 4 Gbps versus 2.5 Gbps, but uses a narrower 128-bit bus compared to the GTX 460M's 192-bit bus, resulting in slightly higher total bandwidth for the GT 745M (64.00 GB/s vs 60.00 GB/s).

Q: Which GPU has a higher transistor density?

A: The NVIDIA GeForce GT 745M has a significantly higher transistor density of 10.8M / mm², compared to the GTX 460M's 4.9M / mm². This is due to its smaller 28 nm process node versus the 40 nm node used by the GTX 460M.

Q: What are the top rival scores for each GPU?

A: The GTX 460M's nearest rival is the AMD FirePro W2100 with a score of 4295, a -0.3% delta. The GT 745M's nearest rival is the AMD Radeon R5 M420 with a score of 3956, a -0.1% delta.

Q: Is there a benchmark where the GT 745M scores higher than its OpenCL result?

A: Yes, the GT 745M scores 5502 in the Geekbench Vulkan benchmark, which is higher than its Geekbench OpenCL score of 3580. It also has a Geekbench Metal score of 2777.

Where Each One Wins

The NVIDIA GeForce GTX 460M is the clear winner in raw compute performance. Its 518.4 GFLOPS of FP32 power and higher pixel/texture rates give it a theoretical edge in tasks that are heavily reliant on parallel mathematical calculations. The 19.6% lead in the Geekbench OpenCL test confirms this strength in a real-world compute workload. This GPU is the better choice for applications that leverage OpenCL and demand maximum throughput from the GPU's shader cores, even if it means dealing with higher power consumption.

The NVIDIA GeForce GT 745M wins in the areas of modern software compatibility and memory capacity. Its support for the Vulkan API opens the door to a newer generation of games and applications that simply will not run on the Fermi-based GTX 460M. With 2 GB of memory, it is also better equipped to handle high-resolution textures and larger datasets than the GTX 460M's 1536 MB. Its higher memory bandwidth (64.00 GB/s) further helps in memory-intensive scenarios. For users who value software support and memory headroom over peak compute speed, the GT 745M is the more practical option, despite its lower benchmark scores.