NVIDIA GeForce GTX 670 vs NVIDIA GeForce GTX 960A Comparison

The NVIDIA GeForce GTX 670 and the NVIDIA GeForce GTX 960A represent two distinct generations of NVIDIA mobile and desktop GPU design, separated by nearly three years of architectural evolution. The GTX 670 is a large, power-hungry desktop part from the Kepler generation, while the GTX 960A is a compact, efficient MXM module from the Maxwell generation. Benchmark data shows a clear performance hierarchy, with the older desktop card delivering substantially higher raw compute throughput, despite the newer card's superior architectural efficiency.

Head-to-Head Benchmarks

The only direct comparison available between these two cards is the Geekbench OpenCL test, and the results are decisive. The GTX 670 scores 15,341 points, while the GTX 960A scores 11,998 points. This translates to a 27.9% advantage for the GTX 670, a significant margin that dwarfs the differences seen in either card's nearest rival comparisons. For context, the GTX 670's average benchmark score is 12,773, which places it within 0.5% of the AMD Radeon RX 7600M XT (12,710) and 0.4% behind the NVIDIA GeForce GTX 590 (12,830). Meanwhile, the GTX 960A's average score of 11,998 puts it a nominal 0.3% ahead of the NVIDIA GeForce GTX 1080 (11,960) and 1.3% ahead of the AMD Radeon RX 6500 XT (11,842) in its own rival grouping.

The 27.9% delta between these two cards is far larger than any of the rival deltas listed for either card. This indicates that the performance gap is not a subtle margin but a fundamental tier difference. The GTX 670's win in this single OpenCL workload is consistent with its overall hardware specifications: it carries more than double the shading units (1,344 vs. 640) and far higher memory bandwidth. In practical terms, the GTX 670 delivers a level of compute throughput that the GTX 960A cannot approach, regardless of the latter's architectural improvements. The GTX 960A does not win any benchmark in the head-to-head comparison, making the GTX 670 a clean sweep winner with 1 win and 0 losses.

FAQ

Q: How much faster is the GTX 670 than the GTX 960A in OpenCL?

A: The GTX 670 scores 15,341 in Geekbench OpenCL, which is 27.9% higher than the GTX 960A's score of 11,998.

Q: Which card has a higher average benchmark score?

A: The GTX 670 has an average benchmark score of 12,773, compared to the GTX 960A's average of 11,998.

Q: How do these cards compare to their nearest rivals?

A: The GTX 670 is essentially tied with the AMD Radeon RX 7600M XT (0.5% ahead) and the NVIDIA GeForce GTX 590 (0.4% behind). The GTX 960A is 0.3% ahead of the NVIDIA GeForce GTX 1080 and 2.7% ahead of the NVIDIA GeForce GTX 1660.

Q: What is the memory bandwidth difference?

A: The GTX 670 provides 192.3 GB/s of bandwidth, while the GTX 960A provides only 80.19 GB/s, a difference of roughly 2.4x in favor of the GTX 670.

Q: What are the power requirements for each card?

A: The GTX 670 has a TDP of 170 W and requires two 6-pin power connectors, with a suggested PSU of 450 W. The GTX 960A has a TDP of 75 W and requires no external power connectors.

Q: Which card has a higher transistor density?

A: The GTX 960A has a slightly higher transistor density at 12.6M / mm², compared to the GTX 670's 12.0M / mm².

Architecture Differences

The architectural divide between these two GPUs is stark. The GTX 670 uses the GK104 chip on the Kepler architecture, fabricated on a 28 nm process at TSMC. It packs 3,540 million transistors onto a 294 mm² die. The GTX 960A uses the GM107 chip on the Maxwell architecture, also on a 28 nm process at TSMC, but with only 1,870 million transistors on a 148 mm² die. This makes the Maxwell chip much smaller and denser, achieving 12.6M transistors per mm² versus the Kepler chip's 12.0M / mm².

The compute configuration differs dramatically. The GTX 670 has 1,344 shading units, 112 texture mapping units (TMUs), and 32 raster output units (ROPs). The GTX 960A has 640 shading units, 40 TMUs, and 16 ROPs. Consequently, the GTX 670's peak pixel rate is 27.44 GPixel/s and its texture rate is 109.8 GTexel/s, while the GTX 960A manages only 18.82 GPixel/s and 47.04 GTexel/s. Floating-point performance follows suit: the GTX 670 delivers 2.634 TFLOPS of FP32 compute, while the GTX 960A delivers 1.505 TFLOPS.

Clock speeds tell a different story. The GTX 960A runs at a base clock of 1097 MHz and a boost of 1176 MHz, higher than the GTX 670's 915 MHz base and 980 MHz boost. Memory clocks also differ: the GTX 670 uses 1502 MHz (6 Gbps effective) GDDR5, while the GTX 960A uses 1253 MHz (5 Gbps effective). The GTX 670 has a 256-bit memory bus, versus the GTX 960A's 128-bit bus, which explains the massive bandwidth gap. Both cards support DirectX 12 (11_0) and OpenGL 4.6, but the GTX 960A supports Vulkan 1.4, while the GTX 670 supports Vulkan 1.2.175.

The Verdict

The benchmark data is unambiguous: the GTX 670 is the superior performer. Its 27.9% lead in OpenCL and higher average score of 12,773 versus 11,998 make it the clear choice for any workload where raw compute matters. The GTX 670's advantage is rooted in its much larger GPU: more shading units, more TMUs, more ROPs, and over twice the memory bandwidth. While the GTX 960A is a more modern architecture with higher clocks and better transistor density, it simply does not have the hardware resources to compete.

However, the verdict is not purely about performance. The GTX 670 is a dual-slot desktop card that is 241 mm long, requires a 450 W PSU, and draws 170 W. The GTX 960A is an MXM module with a 75 W TDP, no power connectors, and no fixed dimensions — designed for portable devices. If the use case is a desktop tower with room for a large card, the GTX 670 is the obvious pick. If the use case is a slim laptop or mobile workstation, the GTX 960A is the only viable option of the two. The data shows that the GTX 670 wins on performance, but the GTX 960A wins on form factor and power efficiency.

Specification Differences

The two cards differ across nearly every measurable specification. The GTX 670 has a base clock of 915 MHz and boost of 980 MHz, while the GTX 960A runs at 1097 MHz base and 1176 MHz boost. The GTX 670's memory is clocked at 1502 MHz (6 Gbps effective), versus 1253 MHz (5 Gbps effective) for the GTX 960A. The GTX 670 has 2 GB of GDDR5 on a 256-bit bus, giving 192.3 GB/s of bandwidth; the GTX 960A has 2 GB of GDDR5 on a 128-bit bus, giving 80.19 GB/s.

Shader units: 1,344 for the GTX 670, 640 for the GTX 960A. TMUs: 112 versus 40. ROPs: 32 versus 16. Pixel rate: 27.44 GPixel/s versus 18.82 GPixel/s. Texture rate: 109.8 GTexel/s versus 47.04 GTexel/s. FP32 compute: 2.634 TFLOPS versus 1.505 TFLOPS. TDP: 170 W versus 75 W. The GTX 670 uses a PCIe 3.0 x16 interface and dual-slot cooling with two 6-pin connectors; the GTX 960A uses an MXM-B (3.0) interface with no power connectors. The GTX 670 has 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2 outputs; the GTX 960A's outputs are portable device dependent. The GTX 670 has a launch MSRP of 399 USD; the GTX 960A has no listed launch MSRP.

Where Each One Wins

The GTX 670 wins decisively in any compute-heavy scenario. Its 27.9% OpenCL lead and higher average benchmark score make it the better choice for tasks like GPU-accelerated rendering, scientific computing, or heavy gaming at high resolutions — provided the system can supply 170 W and accommodate a 241 mm dual-slot card. The 192.3 GB/s of memory bandwidth is critical for texture-heavy workloads and large frame buffers, giving it a clear edge over the GTX 960A's 80.19 GB/s.

The GTX 960A wins in scenarios where power draw and physical size are the primary constraints. With a 75 W TDP and no external power connectors, it can be dropped into systems with minimal power delivery, such as thin laptops or compact MXM-based workstations. Its higher clock speeds (1097 MHz base vs. 915 MHz) and better transistor density (12.6M / mm² vs. 12.0M / mm²) indicate a more efficient design per transistor, though this efficiency does not translate into a performance win. The GTX 960A's only benchmark listing is OpenCL, where it scores 11,998 — respectable for its class, but 27.9% behind the GTX 670.

For gaming, the GTX 670's superior texture rate (109.8 vs. 47.04 GTexel/s) and pixel rate (27.44 vs. 18.82 GPixel/s) will deliver higher frame rates in most titles, especially at higher detail settings. The GTX 960A is best suited for less demanding games or as a mobile solution where battery life and thermals take priority over raw performance. In short, the GTX 670 is a performance card; the GTX 960A is an efficiency card.