NVIDIA GeForce GTX 675M vs NVIDIA GeForce GTX 970 Comparison

The NVIDIA GeForce GTX 970 and NVIDIA GeForce GTX 675M represent two very different eras of mobile and desktop graphics. The GTX 970 is a desktop-oriented card built on the Maxwell 2.0 architecture, while the GTX 675M is a mobile MXM module based on the older Fermi 2.0 design. The data shows a decisive performance gap, but each card has a specific role depending on the system it is installed in and the workloads it faces.

Where Each One Wins

The GTX 970 wins in every measurable benchmark category where both have data. In the only direct head-to-head test available, Geekbench OpenCL, the GTX 970 scores 29,982 points against the GTX 675M’s 6,946 points, a delta of 331.6%. This is not a close contest. The GTX 970’s advantage stems from a much larger GPU: it has 1,664 shading units versus 384, 104 TMUs versus 64, and 56 ROPs versus 32. The GTX 970 also has a significantly higher pixel rate at 65.97 GPixel/s compared to 9.92 GPixel/s, and a texture rate of 122.5 GTexel/s versus 39.68 GTexel/s. In raw throughput, the GTX 970 delivers 3.920 TFLOPS of FP32 compute, while the GTX 675M manages only 952.3 GFLOPS.

Where the GTX 675M could be considered a "winner" is in its physical form factor and power requirements. It is an MXM module with no power connectors, drawing 100 W, and its display outputs are "Portable Device Dependent." This makes it a drop-in upgrade for older gaming laptops that use the MXM-B (3.0) interface. The GTX 970, by contrast, is a dual-slot desktop card that is 267 mm long, requires 2x 6-pin power connectors, and needs a 300 W power supply. So, the GTX 675M wins in the niche of repairability and laptop compatibility, but it loses every performance comparison.

For compute workloads, the GTX 970 also holds a clear edge. In the Passmark GPU Compute test, it scores 4,073. The GTX 675M has no equivalent score in the data, but its overall average benchmark score is 6,946, which is far below the GTX 970’s average of 7,157. The GTX 970 also supports Vulkan 1.4, while the GTX 675M has no Vulkan support listed, making the newer card more suitable for modern cross-platform graphics APIs.

The Verdict

The data is unambiguous: the GTX 970 is the superior performer. It wins the only head-to-head benchmark by a factor of roughly 4.3x, and its architecture is two generations newer, with a smaller 28 nm process node versus 40 nm, and nearly three times the transistor count (5,200 million versus 1,950 million). If you are building a desktop system or upgrading a desktop GPU, the GTX 970 is the only rational choice from this pair. It offers 4 GB of GDDR5 memory on a 256-bit bus with 224.4 GB/s bandwidth, compared to the GTX 675M’s 2 GB GDDR5 on a 256-bit bus with 96.00 GB/s bandwidth.

The GTX 675M is strictly for a legacy laptop scenario where the MXM form factor is mandatory. It is end-of-life, as is the GTX 970, but the GTX 675M’s performance is closer to integrated graphics. Its nearest rival, the AMD Radeon R5 M240, has an average score of 6,975, which is only 0.4% higher than the GTX 675M’s 6,946. That puts the GTX 675M in the same performance class as entry-level mobile GPUs from the same era. Do not pick the GTX 675M for any new build or any workload requiring modern API support; its DirectX support is limited to 12 (11_0), while the GTX 970 supports 12 (12_1).

For anyone choosing between these two, the verdict is simple: the GTX 970 wins on every technical specification that matters for gaming and compute, and it does so by a wide margin. The GTX 675M only makes sense if you are constrained to a specific laptop chassis that uses MXM modules and you cannot replace the entire system.

Head-to-Head Benchmarks

The only direct comparison in the data is the Geekbench OpenCL test. The GTX 970 scores 29,982, and the GTX 675M scores 6,946. The delta is 331.6%, meaning the GTX 970 is roughly 4.3 times faster. This is a compute-oriented benchmark, but it is indicative of the overall gap. For context, the GTX 970’s average benchmark score across all tests is 7,157, while the GTX 675M’s average is 6,946. The average scores are closer than the OpenCL result, but that is because the GTX 970 has many more benchmark entries, including lower-scoring DirectX tests. For example, the GTX 970 scores 46 in Passmark DirectX 10, 71 in DirectX 11, and 41 in DirectX 12, which drags down its average. The GTX 675M has only one benchmark entry (Geekbench OpenCL), so its average is exactly that score.

Looking at the GTX 970’s individual results, it shows strength in different areas. Its Passmark G3D score is 9,638, which is high, and its Passmark G2D score is 764. Its Geekbench Metal score is 13,595, and its Geekbench Vulkan score is 20,005. None of these have direct counterparts for the GTX 675M, but they demonstrate that the GTX 970 can handle a variety of modern APIs. The GTX 675M’s lack of Vulkan support and its older DirectX 11_0 feature level mean it cannot fully utilize newer games or applications that rely on these APIs.

The GTX 970’s average score places it at the 39th percentile of all GPUs, with its nearest rival being the Intel Iris Pro Graphics P580, which scores 7,170 and is 0.2% faster. The GTX 675M is also at the 39th percentile, with its nearest rival being the AMD Radeon R5 M240, which scores 6,975 and is 0.4% faster. Both cards are in the lower half of the performance distribution, but the GTX 970’s raw scores in specific tests are vastly higher than the GTX 675M’s only score.

FAQ

Q: Which GPU is faster in compute workloads?

A: The GTX 970 is vastly faster. In the Geekbench OpenCL test, it scores 29,982 versus the GTX 675M’s 6,946, a 331.6% difference.

Q: Can the GTX 675M be used in a desktop PC?

A: No. The GTX 675M is an MXM Module with an MXM-B (3.0) bus interface and no power connectors. Its display outputs are "Portable Device Dependent," meaning it is designed for laptops.

Q: Does the GTX 675M support Vulkan?

A: No. The data lists Vulkan as null for the GTX 675M, while the GTX 970 supports Vulkan 1.4.

Q: How does the GTX 970 compare to its nearest rivals?

A: The GTX 970 has an average benchmark score of 7,157. Its nearest rival, the Intel Iris Pro Graphics P580, scores 7,170, which is 0.2% higher. The NVIDIA GeForce GTX 750 scores 7,222, which is 0.9% higher.

Q: What is the memory bandwidth difference?

A: The GTX 970 has a bandwidth of 224.4 GB/s, while the GTX 675M has 96.00 GB/s. Both use a 256-bit bus, but the GTX 970 runs its GDDR5 memory at 7 Gbps effective, versus 3 Gbps effective for the GTX 675M.

Q: Which card has a higher pixel fill rate?

A: The GTX 970 has a pixel rate of 65.97 GPixel/s, which is significantly higher than the GTX 675M’s 9.92 GPixel/s.

Architecture Differences

The GTX 970 is built on the Maxwell 2.0 architecture using the GM204 chip, fabricated on a 28 nm process at TSMC. It contains 5,200 million transistors on a 398 mm² die, giving a transistor density of 13.1M per mm². The GTX 675M uses the Fermi 2.0 architecture with the GF114 chip, on a 40 nm process, also at TSMC. It has 1,950 million transistors on a 332 mm² die, with a density of 5.9M per mm².

The architectural leap is substantial. Maxwell 2.0 introduces features like DirectX 12 (12_1) and Vulkan 1.4 support, whereas Fermi 2.0 only reaches DirectX 12 (11_0) and has no Vulkan support. The GTX 970 has 1,664 shading units, 104 TMUs, and 56 ROPs. The GTX 675M has 384 shading units, 64 TMUs, and 32 ROPs. This is a 4.3x difference in shading units, which directly correlates with the compute performance gap seen in the benchmarks.

The GTX 970 also has a much higher clock speed: a base of 1050 MHz and a boost of 1178 MHz. The GTX 675M has no listed base or boost clock, only a memory clock of 750 MHz (3 Gbps effective). The GTX 970’s memory clock is 1753 MHz (7 Gbps effective). The process node difference (28 nm vs 40 nm) explains part of the efficiency and clock speed advantage, but the core design differences are the primary driver of the performance gap.

Specification Differences

The two cards differ in nearly every specification except for the memory bus width, which is 256-bit for both. The most critical differences are:

• Memory size: GTX 970 has 4 GB GDDR5; GTX 675M has 2 GB GDDR5.
• Memory bandwidth: GTX 970 has 224.4 GB/s; GTX 675M has 96.00 GB/s.
• Process node: GTX 970 is 28 nm; GTX 675M is 40 nm.
• Transistors: GTX 970 has 5,200 million; GTX 675M has 1,950 million.
• Die size: GTX 970 is 398 mm²; GTX 675M is 332 mm².
• Shading units: GTX 970 has 1,664; GTX 675M has 384.
• TMUs: GTX 970 has 104; GTX 675M has 64.
• ROPs: GTX 970 has 56; GTX 675M has 32.
• Pixel rate: GTX 970 is 65.97 GPixel/s; GTX 675M is 9.92 GPixel/s.
• Texture rate: GTX 970 is 122.5 GTexel/s; GTX 675M is 39.68 GTexel/s.
• FP32 compute: GTX 970 is 3.920 TFLOPS; GTX 675M is 952.3 GFLOPS.
• TDP: GTX 970 is 148 W; GTX 675M is 100 W.
• Slot width: GTX 970 is dual-slot; GTX 675M is MXM Module.
• Power connectors: GTX 970 has 2x 6-pin; GTX 675M has none.
• Bus interface: GTX 970 is PCIe 3.0 x16; GTX 675M is MXM-B (3.0).
• Display outputs: GTX 970 has 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.2; GTX 675M is portable device dependent.
• DirectX support: GTX 970 is 12 (12_1); GTX 675M is 12 (11_0).
• Vulkan support: GTX 970 is 1.4; GTX 675M is null.
• Dimensions: GTX 970 is 267 mm long, 111 mm high, 40 mm wide; GTX 675M has no listed dimensions.
• Release date: GTX 970 is 2014-09-18; GTX 675M is 2012-03-21.
• Launch MSRP: GTX 970 is 329 USD; GTX 675M has none listed.

These differences reinforce the conclusion that the GTX 970 is a fundamentally more capable piece of hardware, designed for desktop performance, while the GTX 675M is a constrained mobile solution from an older generation.