NVIDIA GeForce GTX 675MX vs NVIDIA GeForce GTX 980 Comparison

The NVIDIA GeForce GTX 675MX and NVIDIA GeForce GTX 980 represent two distinct generations of NVIDIA mobile and desktop graphics, separated by a significant architectural leap. While both cards are now end-of-life products, their benchmark data reveals a clear performance hierarchy. The GTX 980, a desktop-oriented part from the Maxwell 2.0 generation, utterly dominates the older Kepler-based GTX 675MX in every measurable compute task, though the comparison highlights how far mobile graphics have come in a short span.

Head-to-Head Benchmarks

The head-to-head data contains two benchmark entries, and in both, the NVIDIA GeForce GTX 980 is the decisive winner. The most striking result is in the Geekbench OpenCL test, where the GTX 980 scores 34,676 points against the GTX 675MX’s 10,723 points. This represents a delta of -69.1% for the GTX 675MX, meaning the newer card delivers more than three times the raw compute throughput in this OpenCL workload. The margin is so large that it underscores a generational chasm in shader processing and memory bandwidth, not just a clock speed advantage.

The second benchmark, Geekbench Metal, shows a similar pattern. The GTX 980 scores 15,163, while the GTX 675MX manages 6,131. The delta here is -59.6%, indicating the GTX 980 is roughly 2.5 times faster. While Metal is an API more commonly associated with Apple platforms, the relative performance gap is consistent with the OpenCL result, suggesting that the GTX 980’s advantage is not workload-specific but rather a fundamental throughput superiority.

It is importantly the GTX 675MX’s average benchmark score across all its tests is 8,427, while the GTX 980’s average is 8,167. This seemingly reverses the head-to-head results, but it is a statistical artifact of different test suites. The GTX 675MX’s average is derived from only two Geekbench tests, whereas the GTX 980’s average includes a wider array of Passmark tests (DirectX 9, 10, 11, 12, G2D, G3D, and GPU compute) alongside the Geekbench and 3DMark entries. Many of those Passmark legacy tests score very low on modern hardware, pulling the GTX 980’s average down. In the direct head-to-head comparisons, the GTX 980 is unequivocally superior.

FAQ

Q: Which GPU has the higher raw compute performance in OpenCL?

A: The NVIDIA GeForce GTX 980 achieves a Geekbench OpenCL score of 34,676, which is 69.1% higher than the GTX 675MX’s score of 10,723. This indicates the GTX 980 has a massive lead in general-purpose GPU compute.

Q: Is the GTX 675MX competitive with the GTX 980 in any benchmark?

A: No. In the two head-to-head benchmarks available (Geekbench Metal and Geekbench OpenCL), the GTX 980 wins both. The GTX 675MX trails by 59.6% in Metal and 69.1% in OpenCL, showing no area of parity.

Q: How do the two cards compare in terms of memory bandwidth?

A: The GTX 980 has a memory bandwidth of 224.4 GB/s, which is nearly double the GTX 675MX’s 115.2 GB/s. This is a critical factor in the performance difference, as it allows the GTX 980 to feed its much larger shader array more effectively.

Q: Are these two GPUs from the same architectural family?

A: No. The GTX 675MX is based on the Kepler architecture (chip GK104), while the GTX 980 is based on Maxwell 2.0 (chip GM204). Both are built on a 28 nm process by TSMC, but Maxwell 2.0 introduces significant microarchitectural improvements over Kepler.

Q: What is the transistor count difference between the two chips?

A: The GTX 980’s GM204 chip contains 5,200 million transistors on a 398 mm² die, while the GTX 675MX’s GK104 contains 3,540 million transistors on a 294 mm² die. The GTX 980 packs roughly 47% more transistors.

Q: Which card has a higher pixel fill rate?

A: The GTX 980 has a pixel rate of 77.82 GPixel/s, compared to the GTX 675MX’s 13.08 GPixel/s. This is a 5.9x difference, making the GTX 980 vastly superior for high-resolution rendering and heavy fragment shading.

Architecture Differences

The architectural gap between these two GPUs is the primary driver of their performance differential. The GTX 675MX uses the GK104 chip, which is the flagship Kepler design for the GeForce 600M series. Kepler architecture focused on efficiency through a balanced design of shader and texture units, but it lacked some of the advanced scheduling and compression techniques found in later generations. The GTX 675MX features 960 shading units, 80 texture mapping units (TMUs), and 32 ROPs. This configuration is respectable for its era but is dwarfed by the GTX 980.

The GTX 980’s GM204 chip is based on Maxwell 2.0, a major revision that improved instruction scheduling, increased L2 cache efficiency, and introduced advanced color compression. Maxwell 2.0 also brought support for DirectX 12 (feature level 12_1), whereas the GTX 675MX only supports DirectX 12 (feature level 11_0). This is a key differentiator for modern API compatibility. The GTX 980 doubles the shader count to 2,048 units, has 128 TMUs, and 64 ROPs. It also has a higher transistor density at 13.1M / mm² compared to the GTX 675MX’s 12.0M / mm², indicating a more efficient layout despite the larger absolute die size.

The memory subsystems also differ architecturally. While both use a 256-bit GDDR5 bus, the GTX 980 runs its memory at 7 Gbps effective, yielding 224.4 GB/s of bandwidth. The GTX 675MX runs at 3.6 Gbps effective, yielding 115.2 GB/s. This doubling of bandwidth is critical for the GTX 980’s higher resolution texture workloads and compute tasks. Additionally, the GTX 980 supports Vulkan 1.4, while the GTX 675MX is limited to Vulkan 1.2.175, showing a newer driver stack and feature set.

Specification Differences

The specification tables reveal stark contrasts in nearly every measurable category. The GTX 980 has a base clock of 1127 MHz and a boost clock of 1216 MHz, while the GTX 675MX does not list base or boost clocks in the data. The GTX 980’s memory is clocked at 1753 MHz (7 Gbps effective), while the GTX 675MX runs at 900 MHz (3.6 Gbps effective). Memory capacity differs as well: the GTX 980 has 4 GB, the GTX 675MX has 2 GB. Both use GDDR5, but the bandwidth differential is 224.4 GB/s versus 115.2 GB/s.

The compute specifications are heavily skewed toward the GTX 980. Its FP32 performance is 4.981 TFLOPS, while the GTX 675MX is rated at 1,255.7 GFLOPS (1.256 TFLOPS). The GTX 980’s texture rate is 155.6 GTexel/s versus 52.32 GTexel/s for the older card. Pixel rates are 77.82 GPixel/s versus 13.08 GPixel/s. Power consumption is also higher on the GTX 980 at 165 W TDP, compared to the GTX 675MX’s 100 W TDP, reflecting its larger silicon and higher clock speeds.

Physical and interface differences are significant. The GTX 675MX is an MXM module (MXM-B 3.0 interface), designed for laptops, with no dedicated power connectors and display outputs dependent on the portable device. The GTX 980 is a dual-slot desktop card, 267 mm (10.5 inches) long, 111 mm (4.4 inches) tall, and 40 mm (1.6 inches) wide, requiring a 450 W power supply and two 6-pin power connectors. Its display outputs include 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2. The GTX 675MX has no comparable discrete display outputs listed.

Where Each One Wins

The NVIDIA GeForce GTX 980 wins in every head-to-head benchmark and in every speculative use case derived from the specification data. For gaming at high resolutions and detail settings, the GTX 980’s 4 GB of VRAM, 224.4 GB/s bandwidth, and 64 ROPs provide a commanding lead. Its 4.981 TFLOPS FP32 performance makes it suitable for demanding compute workloads, such as OpenCL-based rendering or physics simulations, where it is over three times faster than the GTX 675MX. The GTX 980’s support for DirectX 12 (12_1) and Vulkan 1.4 ensures better compatibility with modern game engines and graphics APIs.

The GTX 675MX, by contrast, is a legacy mobile part. Its only advantage lies in its form factor and power efficiency. At 100 W TDP, it draws significantly less power than the GTX 980’s 165 W, making it suitable for thin-and-light gaming laptops where thermal and power budgets are constrained. Its MXM module design allows for upgradeability in certain laptop chassis, but its performance is strictly last-generation. In terms of benchmark wins, the GTX 675MX has zero wins against the GTX 980. Its percentile rank (43rd) is equal to the GTX 980’s, but that is misleading because the percentile is relative to all GPUs, and the GTX 675MX benefits from a smaller, less demanding benchmark suite.

For users prioritizing raw performance, desktop gaming, or compute acceleration, the GTX 980 is the only choice. The GTX 675MX is relegated to legacy laptop use cases where its lower power draw and compact MXM form factor are the primary considerations. The data is unambiguous: the GTX 980 is a significantly more capable processor in every measurable aspect.