NVIDIA GeForce GTX 460

NVIDIA’s GeForce GTX 460 occupies a specific slice of the Fermi generation, built on the GF104 chip at TSMC’s 40 nm process with 1,950 million transistors on a 332 mm² die. As an end-of-life product, its benchmark presence is defined by a single OpenCL score of 7,858, placing it at the 40th percentile of all GPUs in the database. The data shows a card that sits in a tightly contested mid-range cluster, where the differences between it and its nearest rivals are measured in single-digit percentage points rather than dramatic gulfs in performance.

Benchmark Performance

The GeForce GTX 460’s average benchmark score of 7,858 frames it as a modest performer by contemporary standards. The nearestRivals data reveals a striking compression at this performance tier: the AMD Radeon R9 M360 leads with an average score of 8,006, which is only 1.8% higher than the GTX 460’s result. That margin is statistically negligible in real-world terms, suggesting that the two cards are effectively interchangeable in OpenCL workloads. The NVIDIA GeForce GTX 650 Ti posts an average score of 8,018, a 2% advantage over the GTX 460, again a margin that falls within run-to-run variance for most applications. More surprisingly, the NVIDIA GeForce GTX 970 — a card from a later generation with a substantially different feature set — scores just 8,024, only 2.1% ahead of the GTX 460 in this particular test. The NVIDIA GRID K2 rounds out the rival set with an average score of 8,075, a 2.7% lead.

What these deltas indicate is that the GTX 460, despite its age and architectural lineage, maintains competitive parity with a range of cards that span multiple product generations. The benchmark results do not show the GTX 460 being outpaced by any rival by more than 2.7%, which suggests that the OpenCL workload here is not particularly sensitive to the architectural differences between Fermi and later designs. The card’s raw compute throughput — 907.2 GFLOPS of FP32 performance — is evidently sufficient to keep it within striking distance of cards that carry higher transistor counts and more modern memory subsystems. The 40th percentile ranking reinforces this: the GTX 460 is not an outlier at the bottom of the distribution, but rather sits in the lower-middle quartile, surrounded by a dense cluster of similarly-scoring hardware.

Memory Subsystem

The GTX 460 ships with 768 MB of GDDR5 memory on a 192-bit bus, yielding a bandwidth of 86.40 GB/s. The memory clock runs at 900 MHz, which translates to 3.6 Gbps effective data rate. For high-resolution workloads, this configuration presents a clear bottleneck. The 768 MB capacity is small even by the standards of its release era, and the 192-bit bus width limits the amount of data that can be moved per clock cycle. The 86.40 GB/s bandwidth is sufficient for 1080p gaming at medium settings, but the data suggests that higher resolutions — 1440p or beyond — would strain the memory subsystem severely. Texture-heavy scenes with large framebuffers would quickly exhaust the 768 MB pool, forcing the card to spill into system memory over the PCIe 2.0 x16 interface, which would incur a significant performance penalty.

The pixel rate of 9.450 GPixel/s and texture rate of 37.80 GTexel/s further contextualize the memory limitations. With 24 ROPs and 56 TMUs, the card can generate pixels and sample textures at rates that outpace its ability to store and retrieve the resulting data. This imbalance means that in practice, the GTX 460’s memory bandwidth becomes the limiting factor in most modern workloads, particularly those that use high-resolution textures or anti-aliasing. The 192-bit bus is a deliberate cost-cutting measure relative to wider 256-bit designs, and the benchmark data reflects this: the card’s overall score is competitive, but it does not excel in memory-bound scenarios. For users targeting 1080p with conservative settings, the 86.40 GB/s is adequate; for anything more demanding, the memory subsystem would need to be the primary concern.

Ray Tracing and Feature Set

The GTX 460 is built on the Fermi architecture, which predates dedicated ray tracing hardware. The fact pack lists no RT cores and no tensor cores, meaning the card relies entirely on its 336 shading units for any compute tasks. The FP32 throughput of 907.2 GFLOPS is the sole compute metric available, and it is this raw shading power that drives the OpenCL benchmark score. DirectX 12 support is listed as version 12 (11_0), which indicates feature-level 11_0 compatibility rather than full DirectX 12 Ultimate capabilities. OpenGL 4.6 is supported, but Vulkan is not listed, which limits the card’s compatibility with modern cross-platform APIs that rely on Vulkan for low-overhead rendering.

For ray tracing, the absence of RT cores means that any ray-traced effects would have to be computed via shader-based methods, which is prohibitively slow on a card with 907.2 GFLOPS of FP32 performance. The data does not include any ray tracing benchmarks, and given the architectural constraints, it is reasonable to infer that the GTX 460 is not a viable option for hardware-accelerated ray tracing in any capacity. The feature set is firmly rooted in the DirectX 11 era, with the 2x DVI and 1x mini-HDMI 1.3a outputs reflecting the display connectivity standards of its time. The card’s 160 W TDP and dual-slot cooler with 2x 6-pin power connectors indicate that it was designed for mainstream desktop use, with a suggested PSU of 450 W being a modest requirement by modern standards.

How It Compares

AMD Radeon R9 M360: The R9 M360 leads the GTX 460 by 1.8% in average benchmark score, with 8,006 versus 7,858. This is a negligible margin that places the two cards on equal footing in the OpenCL test. The R9 M360 is a mobile-oriented part, which makes its slight lead notable given the GTX 460’s desktop power envelope of 160 W versus the likely lower TDP of the mobile card. In practice, the data suggests no meaningful performance difference between these two.

NVIDIA GeForce GTX 650 Ti: The GTX 650 Ti scores 8,018, a 2% advantage over the GTX 460. This is the smallest lead among the NVIDIA rivals, and it comes from a card that was released two years later. The GTX 650 Ti benefits from a newer architecture, but the benchmark results show that the GTX 460’s higher shading unit count (336 versus the 650 Ti’s configuration) partially compensates. The 2% delta is within the noise of most benchmarking suites.

NVIDIA GeForce GTX 970: The GTX 970 posts an average score of 8,024, which is 2.1% higher than the GTX 460. This is the most surprising comparison in the rival set. The GTX 970 is a high-end card from a later generation with a 256-bit memory bus and 4 GB of VRAM, yet the OpenCL score shows it only marginally ahead of the GTX 460. This suggests that the specific workload measured in the geekbench_opencl test does not emphasize the GTX 970’s strengths, such as memory bandwidth or shader count. The data does not support a claim that the GTX 460 outperforms the GTX 970, but it does indicate that in this narrow benchmark, the gap is far smaller than the generational difference would imply.

NVIDIA GRID K2: The GRID K2, a virtualized GPU for cloud workloads, scores 8,075, leading the GTX 460 by 2.7%. This is the largest delta in the rival group, yet still a modest margin. The GRID K2’s enterprise orientation means it is optimized for multi-user scenarios, and its 2.7% lead in this single-threaded OpenCL test does not necessarily translate to real-world virtual desktop performance. The data shows that the GTX 460 holds its own against a professional-grade virtualized GPU.

FAQ

Q: What is the GeForce GTX 460’s average benchmark score?

A: The average benchmark score is 7,858, based on the geekbench_opencl test, which places the card at the 40th percentile of all GPUs in the database.

Q: How does the GTX 460 compare to the NVIDIA GeForce GTX 970?

A: The GTX 970 has an average score of 8,024, which is 2.1% higher than the GTX 460’s 7,858. Despite being a later-generation high-end card, the GTX 970’s lead in this specific OpenCL test is minimal.

Q: What memory configuration does the GTX 460 use?

A: It uses 768 MB of GDDR5 memory on a 192-bit bus, with a memory clock of 900 MHz (3.6 Gbps effective) and a bandwidth of 86.40 GB/s.

Q: Does the GTX 460 support hardware ray tracing?

A: No. The fact pack lists no RT cores or tensor cores, and the Fermi architecture predates dedicated ray tracing hardware. Any ray tracing would require shader-based computation on the 336 shading units.

Q: What are the power requirements for the GTX 460?

A: The card has a TDP of 160 W, requires 2x 6-pin power connectors, and has a suggested PSU rating of 450 W. It is a dual-slot card.

Q: What is the launch MSRP for the GTX 460?

A: The launch MSRP is 199 USD. The card was released on July 11, 2010, and is now end-of-life, with the GeForce 500 series as its successor.