NVIDIA GeForce GTX 460 SE vs NVIDIA GeForce GTX 670MX Comparison

NVIDIA GeForce GTX 460 SE and NVIDIA GeForce GTX 670MX represent two distinct generations of NVIDIA mobile and desktop GPUs, with the former built on the Fermi architecture and the latter on Kepler. The recorded data shows a single head-to-head OpenCL benchmark, where the GTX 460 SE scores 6389 versus the GTX 670MX’s 6125, a 4.3% advantage. However, the GTX 670MX has additional Vulkan benchmark data and a different feature set that shifts its suitability.

Where Each One Wins

The GTX 460 SE wins the only directly comparable benchmark in the database: Geekbench OpenCL. Its score of 6389 beats the GTX 670MX’s 6125, a 4.3% lead. This indicates that in raw compute workloads that rely on OpenCL, the older Fermi card holds a measurable edge. The GTX 460 SE also sits at the 37th percentile among all GPUs, while the GTX 670MX sits lower at the 33rd percentile. For tasks like OpenCL-accelerated physics, image processing, or general-purpose GPU compute, the GTX 460 SE is the stronger choice based on this metric.

The GTX 670MX, however, has a Vulkan benchmark score of 5316, which the GTX 460 SE lacks entirely. This means the GTX 670MX supports Vulkan 1.2.175, whereas the GTX 460 SE has no recorded Vulkan API support. In modern games or applications that use Vulkan for rendering or compute, the GTX 670MX is the only option between the two. Its 75 W TDP also makes it far more power-efficient, which is critical for laptop use, as its display outputs are listed as "Portable Device Dependent." The GTX 460 SE, in contrast, is a dual-slot desktop card with 2x 6-pin power connectors and a 150 W TDP.

So the split is clear: if the workload is OpenCL-heavy and you can tolerate desktop power and space, the GTX 460 SE wins. If you need Vulkan support, a mobile form factor, or lower power draw, the GTX 670MX is the only viable pick.

Architecture Differences

The two GPUs come from different NVIDIA architectures. The GTX 460 SE uses the GF104 chip on the Fermi architecture, manufactured on a 40 nm process at TSMC. The GTX 670MX uses the GK104 chip on the Kepler architecture, also from TSMC but on a 28 nm process. The process shrink from 40 nm to 28 nm is significant, but the transistor counts tell a more nuanced story. The GTX 460 SE packs 1,950 million transistors on a 332 mm² die, giving a transistor density of 5.9M per mm². The GTX 670MX has 3,540 million transistors on a smaller 294 mm² die, resulting in a density of 12.0M per mm². That is more than double the density, showing Kepler’s efficiency gains in packing logic into a smaller area.

Core counts differ substantially. The GTX 460 SE has 288 shading units, 48 texture mapping units, and 32 raster operation units. The GTX 670MX has 960 shading units, 80 TMUs, and 24 ROPs. The GTX 670MX has over three times the shading units and nearly double the TMUs, but fewer ROPs. This means the Kepler card is designed for higher shader throughput, while the Fermi card has more ROP throughput relative to its core count. Pixel rates reflect this: the GTX 460 SE achieves 7.800 GPixel/s, while the GTX 670MX achieves 12.02 GPixel/s. Texture rates are 31.20 GTexel/s for the GTX 460 SE and 48.08 GTexel/s for the GTX 670MX. FP32 compute is 748.8 GFLOPS versus 1,153.9 GFLOPS, favoring the GTX 670MX by a wide margin. Despite this, the OpenCL benchmark shows the GTX 460 SE winning, which suggests that architecture-specific drivers or memory bandwidth play a larger role in that particular test.

Memory configurations also differ. The GTX 460 SE has 1024 MB of GDDR5 on a 256-bit bus, yielding 108.8 GB/s of bandwidth. The GTX 670MX has 3 GB of GDDR5 on a 192-bit bus, yielding only 67.20 GB/s. The GTX 460 SE’s wider bus provides significantly more bandwidth, which likely explains its OpenCL lead. The GTX 670MX compensates with triple the capacity, which is better for holding larger textures or datasets, but at the cost of speed.

The Verdict

From the recorded data, the GTX 460 SE is the better card for raw OpenCL compute performance. Its 6389 score is 4.3% ahead of the GTX 670MX’s 6125, and it ranks higher in the overall percentile distribution (37th versus 33rd). If your application is primarily OpenCL-based and you are building a desktop system with the power budget for a 150 W card, the GTX 460 SE is the logical pick.

The GTX 670MX, however, is the only option for Vulkan workloads, as it has a recorded Vulkan score of 5316 and supports Vulkan 1.2.175. It also consumes half the power (75 W versus 150 W) and lacks external power connectors, making it suitable for laptops or low-power systems. Its 3 GB memory capacity is triple that of the GTX 460 SE, which is beneficial for memory-hungry tasks like larger render targets or modern game assets. For a mobile user or someone needing Vulkan API support, the GTX 670MX wins by default.

Neither card is current, as both are end-of-life, but the data shows distinct use cases. Pick the GTX 460 SE for OpenCL compute and higher memory bandwidth. Pick the GTX 670MX for Vulkan compatibility, lower power, and larger memory capacity.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The GTX 460 SE has an average benchmark score of 6389, while the GTX 670MX averages 5721 across its two recorded tests (OpenCL and Vulkan).

Q: Does the GTX 670MX support Vulkan?

A: Yes, the GTX 670MX has a Vulkan API version of 1.2.175 and a recorded Geekbench Vulkan score of 5316. The GTX 460 SE has no Vulkan support listed.

Q: How does memory bandwidth compare between the two?

A: The GTX 460 SE has a 256-bit bus with 108.8 GB/s bandwidth, while the GTX 670MX has a 192-bit bus with 67.20 GB/s. The GTX 460 SE offers 41.6 GB/s more bandwidth.

Q: Which card has more shading units?

A: The GTX 670MX has 960 shading units, compared to the GTX 460 SE’s 288, a ratio of over 3 to 1.

Q: What is the power consumption difference?

A: The GTX 460 SE has a 150 W TDP and requires 2x 6-pin power connectors. The GTX 670MX has a 75 W TDP and uses no external power connectors.

Q: Which card has a larger memory capacity?

A: The GTX 670MX has 3 GB, while the GTX 460 SE has 1024 MB (1 GB). The GTX 670MX holds triple the memory.

Head-to-Head Benchmarks

The database contains one direct head-to-head benchmark: Geekbench OpenCL. The GTX 460 SE scores 6389, and the GTX 670MX scores 6125. The delta is 4.3% in favor of the GTX 460 SE. This is a modest lead, but it is consistent with the GTX 460 SE’s higher memory bandwidth (108.8 GB/s versus 67.20 GB/s) and its wider 256-bit bus. OpenCL workloads often scale with bandwidth, especially when moving large data blocks. The GTX 670MX has more shading units (960 versus 288) and higher FP32 compute (1,153.9 GFLOPS versus 748.8 GFLOPS), yet it still loses this test. This suggests that the GTX 460 SE’s Fermi architecture is better optimized for the specific OpenCL operations in this benchmark, or that the memory subsystem is the limiting factor for the GTX 670MX.

The GTX 670MX’s only other benchmark is Geekbench Vulkan, where it scores 5316. There is no Vulkan score for the GTX 460 SE, so no direct comparison is possible. However, the GTX 670MX’s Vulkan support is a clear functional advantage. Its average benchmark score across both tests is 5721, which places it at the 33rd percentile, while the GTX 460 SE’s single score places it at the 37th percentile. In terms of nearest rivals, the GTX 460 SE matches the NVIDIA GeForce GTX 580M exactly (both 6389, 0% delta) and trails the NVIDIA RTX PRO 5000 72 GB Blackwell by only 0.3%. The GTX 670MX is within 0.2% of the Intel Iris Pro Graphics P6300 and 0.2% behind the NVIDIA GeForce GTX 550 Ti.

The biggest win for the GTX 460 SE is the OpenCL test, where it is 4.3% ahead. The biggest win for the GTX 670MX is its Vulkan capability, which is not a score comparison but a feature that the GTX 460 SE lacks. If you only look at the numbers, the GTX 460 SE is the stronger performer in the shared test. If you look at API coverage, the GTX 670MX is the more versatile card.

Specification Differences

The following fields differ between the two GPUs:

• Chip: GF104 (GTX 460 SE) versus GK104 (GTX 670MX)
• Architecture: Fermi versus Kepler
• Generation: GeForce 400 versus GeForce 600M
• Process Node: 40 nm versus 28 nm
• Transistors: 1,950 million versus 3,540 million
• Die Size: 332 mm² versus 294 mm²
• Transistor Density: 5.9M per mm² versus 12.0M per mm²
• Base Clock: Not listed for the GTX 460 SE versus 601 MHz for the GTX 670MX (boost is also 601 MHz)
• Memory Clock: 850 MHz (3.4 Gbps effective) versus 700 MHz (2.8 Gbps effective)
• Memory Size: 1024 MB versus 3 GB
• Memory Bus Width: 256 bit versus 192 bit
• Memory Bandwidth: 108.8 GB/s versus 67.20 GB/s
• Shading Units: 288 versus 960
• TMUs: 48 versus 80
• ROPs: 32 versus 24
• Pixel Rate: 7.800 GPixel/s versus 12.02 GPixel/s
• Texture Rate: 31.20 GTexel/s versus 48.08 GTexel/s
• FP32: 748.8 GFLOPS versus 1,153.9 GFLOPS
• TDP: 150 W versus 75 W
• Slot Width: Dual-slot versus not listed
• Power Connectors: 2x 6-pin versus None
• Suggested PSU: 450 W versus not listed
• Bus Interface: PCIe 2.0 x16 versus PCIe 3.0 x16
• Display Outputs: 2x DVI, 1x mini-HDMI 1.3a versus Portable Device Dependent
• Vulkan API: Not listed versus 1.2.175
• Release Date: 2010-11-14 versus 2012-09-30
• Predecessor: GeForce 200 versus GeForce 500M
• Successor: GeForce 500 versus GeForce 700M
• Launch MSRP: 160 USD versus not listed

These differences highlight a generational shift: Kepler brought more cores, lower power, and a smaller die, but Fermi retained a wider memory bus and higher bandwidth. The GTX 460 SE’s launch MSRP was 160 USD; the GTX 670MX has no recorded launch price.