NVIDIA GeForce 945M vs NVIDIA GeForce GTX 675M Comparison

# NVIDIA GeForce 945M vs NVIDIA GeForce GTX 675M: Mobile GPU Comparison

The GeForce 945M and the GeForce GTX 675M represent two distinct generations of NVIDIA's mobile graphics solutions. The 945M is a Maxwell-based part from the GeForce 900M series, while the GTX 675M is a Fermi 2.0 part from the GeForce 600M series. The database's head-to-head benchmark suite records exactly one competitive benchmark test between the two: Geekbench OpenCL. The results show a significant edge for the Maxwell part.

Head-to-Head Benchmarks

The database records a single head-to-head benchmark, Geekbench OpenCL, with clear results. The NVIDIA GeForce 945M scores 8099, while the NVIDIA GeForce GTX 675M scores 6946. The 945M holds a 16.6% performance advantage in this workload. The margin is substantial, representing a meaningful generational uplift in compute-oriented tasks.

This result aligns with the broader average benchmark scores recorded for each card. The 945M has an average benchmark score of 8099 across all recorded workloads, while the GTX 675M averages 6946. The 945M's percentile rank among all GPUs in the database is 42%, while the GTX 675M sits at the 39th percentile. This suggests that while both parts are firmly in the mid-range of the mobile GPU landscape, the 945M sits slightly higher in the distribution.

Looking at the nearest rivals in the database, the 945M sits close to the NVIDIA GRID K2, which averages 8080, a score just 0.2% lower. The AMD Radeon R9 M360 scores 8129, which is 0.4% higher, indicating the 945M is surrounded by parts within a tight performance band. On the GTX 675M side, the closest rivals are the NVIDIA GeForce GTX 650 Ti Boost at 8067, which is 0.4% lower, and the GTX 950M at 8135, which is 0.4% higher. The data paints a picture of two parts sitting in a dense cluster of similarly performing mobile GPUs.

Where Each One Wins

The recorded data shows a single benchmark test, which makes a comprehensive use-case split difficult. However, the raw specifications and the characteristics of each part suggest clear scenarios where one might be preferred over the other.

The 945M is built on the Maxwell architecture using the GM107 chip, fabricated on a 28 nm process at TSMC. It packs 1,870 million transistors onto a die size of 148 mm², resulting in a transistor density of 12.6 million per mm². The GTX 675M, by contrast, uses the GF114 chip based on the Fermi 2.0 architecture, built on a 40 nm process at TSMC. It has 1,950 million transistors on a 332 mm² die, giving a density of 5.9 million per mm².

The architectural differences are significant. Maxwell is a more modern, power-efficient design. The 945M's base clock runs at 928 MHz with a boost up to 1020 MHz, while the GTX 675M's memory clock is listed at 750 MHz with 3 Gbps effective. These figures suggest the 945M is tuned for lower power consumption and higher efficiency per clock, while the 675M is a larger, more power-hungry chip. The 945M has a TDP of 75 W, whereas the GTX 675M draws 100 W, a 25 W difference that highlights the efficiency gains of the newer architecture.

The 945M features a 128-bit memory bus with 2 GB of DDR3 memory, while the GTX 675M uses a 256-bit bus with 2 GB of GDDR5. The GTX 675M's memory bandwidth is 96.00 GB/s, which is substantially higher than the 28.80 GB/s of the 945M. This is a major difference in memory bandwidth, with the Fermi card having a significantly wider bus and faster memory type.

The shading and pixel processing capabilities also differ notably. The 945M includes 640 shading units and 16 ROPs, while the GTX 675M houses 384 shading units and 32 ROPs. The 945M's pixel rate is 16.32 GPixel/s and texture rate is 40.80 GTexel/s, versus the 675M's 9.920 GPixel/s and 39.68 GTexel/s. The Maxwell chip thus has fewer ROPs but a much higher pixel throughput, while the Fermi chip has a higher texture rate only by a small margin. The 945M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4, whereas the GTX 675M supports DirectX 12 (11_0) and OpenGL 4.6.

The 945M is a portable device dependent (MXM-B 3.0) module with no power connectors, while the GTX 675M is an MXM Module (MXM-B 3.0). Both have portable device dependent display outputs. The 945M's production status is "End-of-life", released in late 2015, succeeding the GeForce 800M, while the GTX 675M is also "End-of-life", released in early 2012, and follows the GeForce 500M.

The Verdict

Based strictly on the database metrics, the NVIDIA GeForce 945M is the faster part in the recorded workload. The head-to-head result shows a 16.6% OpenCL performance lead. This aligns with the 945M's more efficient Maxwell architecture, newer 28 nm process, and its superior transistor density (12.6 million per mm² versus 5.9M on the Fermi part). The 945M executes fewer shader operations (640 versus 384) but achieves a higher pixel rate, which indicates a better instruction scheduling and clock efficiency. The GTX 675M's advantage is found in memory subsystem: its 256-bit bus and GDDR5 memory provide 96.00 GB/s of bandwidth, over 3.3x the bandwidth of the 945M's 28.80 GB/s. For workloads that are memory bound, the GTX 675M is likely to be the stronger part. However, for the general compute and gaming workload captured in the database, the 945M is the winner.

The 945M is best for users who prioritize compute efficiency and modern feature support, including Vulkan 1.4 API. The GTX 675M is a better pick for workloads that require massive memory bandwidth, such as high-resolution texture streaming or large frame buffers, where its 256-bit bus gives it a major edge. It is also a more capable geometry processor with its 32 ROPs and 32 TMUs. But the overall data points to the 945M.

FAQ

Q: How much faster is the NVIDIA GeForce 945M than the GTX 675M?

A: In the recorded Geekbench OpenCL benchmark, the 945M scores 8099 versus 6946, a 16.6% lead.

Q: Which card has higher memory bandwidth?

A: The GTX 675M. It has a 256-bit memory bus with GDDR5 memory providing 96.00 GB/s, while the 945M has a 128-bit bus with DDR3 giving 28.80 GB/s.

Q: Does the 945M support Vulkan?

A: Yes, the 945M supports Vulkan 1.4 as well as DirectX 12 (11_0) and OpenGL 4.6. The GTX 675M supports DirectX 12 (11_0) and OpenGL 4.6 but no Vulkan is recorded.

Q: What is the power draw difference?

A: The 945M has a 75 W TDP, while the GTX 675M has a 100 W TDP.

Q: What is the memory configuration?

A: The 945M has 2 GB DDR3 on a 128-bit bus. The GTX 675M has 2 GB GDDR5 on a 256-bit bus.

Architecture Differences

The two cards are built on different architectures. The 945M uses NVIDIA's Maxwell architecture with a GM107 chip. The GTX 675M uses the Fermi 2.0 architecture with a GF114 chip. The fabrication process differs: the 945 is on a 28 nm process at TSMC, while the GTX uses a 40 nm node, also at TSMC. The 945M's die is smaller at 148 mm², yet it contains a similar number of transistors (1,870 million versus 1,950 million). This makes the Maxwell part the more efficient and denser design. The 945M is a newer release and is classified as End-of-life, as is the GTX 675. The 945M has a base clock of 928 MHz and a boost of 1020 MHz, while the GTX 675M's memory clock is 750 MHz (with a 3 Gbps effective data rate). The 945M has 640 shader units, 40 TMUs, and 16 ROPs, while the GTX 675M has 384 shaders, 64 TMUs, and 32 ROPs.