NVIDIA GeForce GTX 670MX vs NVIDIA Quadro 4000M Comparison

FAQ

Q: How do the two GPUs compare in overall benchmark performance?

A: The NVIDIA GeForce GTX 670MX has an average benchmark score of 5721, while the NVIDIA Quadro 4000M scores 5211. The GTX 670MX sits at the 33rd percentile among all GPUs, while the Quadro 4000M sits at the 30th percentile.

Q: Which GPU wins in the head-to-head OpenCL test?

A: The GTX 670MX wins the Geekbench OpenCL test with a score of 6125 versus the Quadro 4000M's 5211, a margin of 17.5%. This is the only head-to-head benchmark recorded in the database.

Q: What are the memory specifications of each GPU?

A: The GTX 670MX has 3 GB of GDDR5 memory on a 192-bit bus with 67.20 GB/s bandwidth. The Quadro 4000M has 2 GB of GDDR5 memory on a 256-bit bus with 80.00 GB/s bandwidth.

Q: What process nodes are used for each chip?

A: The GTX 670MX uses a 28 nm process with 3,540 million transistors on a 294 mm² die. The Quadro 4000M uses a 40 nm process with 1,950 million transistors on a 332 mm² die.

Q: Which GPU has higher thermal design power?

A: The Quadro 4000M has a higher thermal design power at 100 W, whereas the GTX 670MX is rated at 75 W.

Q: What are the supported APIs for each GPU?

A: Both support DirectX 12 (11_0) and OpenGL 4.6. The GTX 670MX also supports Vulkan 1.2.175, while the Quadro 4000M has no Vulkan entry in the database.

Architecture Differences

The two GPUs come from different NVIDIA architectures and generations. The GeForce GTX 670MX is built on the Kepler architecture with the GK104 chip, part of the GeForce 600M generation. The Quadro 4000M is built on the Fermi architecture with the GF104 chip, part of the Quadro Fermi-M (x000M) generation.

The manufacturing process differs significantly. The GTX 670MX uses a 28 nm process from TSMC, while the Quadro 4000M uses a 40 nm process, also from TSMC. This process difference explains part of the performance gap: the 28 nm node allows for substantially higher transistor density. The GTX 670MX packs 3,540 million transistors into a 294 mm² die, yielding a density of 12.0 million transistors per mm². The Quadro 4000M contains 1,950 million transistors on a larger 332 mm² die, for a density of 5.9 million transistors per mm².

Shader resources are very different. The GTX 670MX has 960 shading units, 80 texture mapping units, and 24 raster operation pipelines. The Quadro 4000M has 336 shading units, 56 texture mapping units, and 32 raster operation pipelines. The GTX 670MX therefore has nearly three times the shading units, more than 40% more TMUs, but fewer ROPs than the Quadro.

Compute rates reflect these differences. The GTX 670MX delivers 1,153.9 GFLOPS of FP32 compute, 12.02 GPixel/s pixel fill rate, and 48.08 GTexel/s texture fill rate. The Quadro 4000M delivers 638.4 GFLOPS, 6.650 GPixel/s, and 26.60 GTexel/s. In every rate category, the GTX 670MX leads by a substantial margin, though the Quadro's higher ROP count gives it a relatively smaller deficit in pixel throughput than in shading.

Memory architecture also differs. The GTX 670MX uses a 192-bit memory bus with 3 GB of GDDR5, clocked at 700 MHz for 2.8 Gbps effective, giving 67.20 GB/s bandwidth. The Quadro 4000M uses a wider 256-bit bus with 2 GB of GDDR5, clocked at 625 MHz for 2.5 Gbps effective, giving 80.00 GB/s bandwidth. The Quadro actually has higher memory bandwidth due to the wider bus, despite slower memory clocks and less capacity.

The bus interface differs as well. The GTX 670MX uses PCIe 3.0 x16, while the Quadro 4000M uses MXM-B (3.0), reflecting its intended mobile workstation form factor. Both are listed as "Portable Device Dependent" for display outputs, meaning outputs vary by laptop implementation. The Quadro 4000M is listed as an MXM Module for slot width, while the GTX 670MX has no slot width entry.

Neither GPU has RT cores or tensor cores. Both are end-of-life products. The GTX 670MX was released on 2012-09-30, and the Quadro 4000M on 2011-02-21. The GTX 670MX has a predecessor in GeForce 500M and successor in GeForce 700M; the Quadro 4000M has a predecessor in Quadro FX Mobile and successor in Quadro Kepler-M.

Head-to-Head Benchmarks

The database contains exactly one head-to-head benchmark between these two GPUs: the Geekbench OpenCL test. In this test, the GeForce GTX 670MX scores 6125, while the Quadro 4000M scores 5211. The GTX 670MX wins by 17.5%.

A 17.5% lead in OpenCL is meaningful for compute-oriented workloads. The GTX 670MX's higher shading unit count (960 versus 336) and higher FP32 throughput (1,153.9 GFLOPS versus 638.4 GFLOPS) directly support this result. The difference is larger than the raw shading unit ratio might suggest, because the GTX 670MX also benefits from the smaller 28 nm process and the associated clock behavior.

The average benchmark scores in the database tell a similar story. The GTX 670MX averages 5721 across its two recorded benchmarks (Geekbench OpenCL at 6125 and Geekbench Vulkan at 5316). The Quadro 4000M has only one recorded benchmark, the Geekbench OpenCL at 5211, which is also its average score.

Context from nearest rivals reinforces the positioning. The GTX 670MX sits between the NVIDIA GeForce GTX 550 Ti (average 5731, 0.2% above) and the Intel Iris Pro Graphics P6300 (average 5712, 0.2% below). It is also 0.5% above the AMD Radeon HD 8790M and 2.1% above the NVIDIA Quadro M500M. The Quadro 4000M sits between the NVIDIA GeForce 940M (average 5284, 1.4% above) and the AMD Radeon R7 M260X (average 5161, 1% below). It is also 1.1% above the NVIDIA Quadro K3100M and 0.5% below the NVIDIA GeForce GTX 760M.

These rival comparisons place the GTX 670MX in a slightly higher performance tier overall. The GTX 670MX is competitive with desktop-class GeForce GTX 550 Ti, while the Quadro 4000M sits closer to entry-level mobile parts like the GeForce 940M and GTX 760M. The 17.5% head-to-head lead is consistent with the percentile gap: 33rd versus 30th percentile among all GPUs.

Specification Differences

The following fields differ between the two GPUs, based on the recorded data:

• Chip: GK104 (GTX 670MX) versus GF104 (Quadro 4000M)
• Architecture: Kepler versus Fermi
• Generation: GeForce 600M versus Quadro Fermi-M (x000M)
• Process node: 28 nm versus 40 nm
• Transistors: 3,540 million versus 1,950 million
• Die size: 294 mm² versus 332 mm²
• Transistor density: 12.0M / mm² versus 5.9M / mm²
• Base clock: 601 MHz versus no base clock listed
• Boost clock: 601 MHz versus no boost clock listed
• Memory clock: 700 MHz (2.8 Gbps effective) versus 625 MHz (2.5 Gbps effective)
• Memory size: 3 GB versus 2 GB
• Memory bus width: 192 bit versus 256 bit
• Memory bandwidth: 67.20 GB/s versus 80.00 GB/s
• Shading units: 960 versus 336
• TMUs: 80 versus 56
• ROPs: 24 versus 32
• Pixel rate: 12.02 GPixel/s versus 6.650 GPixel/s
• Texture rate: 48.08 GTexel/s versus 26.60 GTexel/s
• FP32: 1,153.9 GFLOPS versus 638.4 GFLOPS
• TDP: 75 W versus 100 W
• Slot width: not listed versus MXM Module
• Bus interface: PCIe 3.0 x16 versus MXM-B (3.0)
• Vulkan support: 1.2.175 versus none listed
• Release date: 2012-09-30 versus 2011-02-21
• Predecessor: GeForce 500M versus Quadro FX Mobile
• Successor: GeForce 700M versus Quadro Kepler-M

Fields that are the same include manufacturer (both NVIDIA), foundry (both TSMC), memory type (both GDDR5), power connectors (both "None"), display outputs (both "Portable Device Dependent"), DirectX support (both 12 (11_0)), OpenGL support (both 4.6), production status (both end-of-life), and the absence of RT cores and tensor cores.

Notably, the Quadro 4000M has no base or boost clock entries, and no game clock. The GTX 670MX has a base clock of 601 MHz and a boost clock of 601 MHz, meaning it does not dynamically boost beyond its base in the recorded data.

Where Each One Wins

GeForce GTX 670MX wins in compute-heavy workloads. The 17.5% OpenCL lead is the clearest signal. With 960 shading units, 80 TMUs, and 1,153.9 GFLOPS of FP32 throughput, the GTX 670MX is designed for parallel processing tasks. Its 28 nm process allows higher transistor density and lower power draw per transistor, contributing to its 75 W TDP despite significantly more compute resources than the Quadro's 100 W TDP.

GeForce GTX 670MX wins in texture-heavy rendering. Its 48.08 GTexel/s texture rate is 80% higher than the Quadro's 26.60 GTexel/s. The 80 TMUs versus 56 TMUs, combined with a higher texture rate, favors applications that sample many textures per frame.

GeForce GTX 670MX wins in pixel throughput despite fewer ROPs. The GTX 670MX delivers 12.02 GPixel/s versus the Quadro's 6.650 GPixel/s. Although the Quadro has 32 ROPs versus 24, the GTX 670MX's higher clock and shading resources more than compensate. This matters for fill-rate-bound scenarios at high resolutions.

Quadro 4000M wins in memory bandwidth. Its 80.00 GB/s exceeds the GTX 670MX's 67.20 GB/s by 19%. The 256-bit bus, even with slower memory clocks, provides more raw bandwidth. This could benefit workloads that stream large datasets with poor locality, though the effect is limited by the Quadro's lower compute rates.

Quadro 4000M wins in ROP count. With 32 ROPs versus 24, the Quadro has more raster operation units. This may provide an advantage in specific depth-buffer or blend-heavy scenarios, but the recorded pixel rate shows the GTX 670MX still wins overall in pixel throughput.

GeForce GTX 670MX wins in API compatibility. It supports Vulkan 1.2.175, while the Quadro 4000M has no Vulkan entry. Both support DirectX 12 (11_0) and OpenGL 4.6. For modern applications that can use Vulkan, the GTX 670MX has a clear software advantage.

GeForce GTX 670MX wins in capacity and compute density. The 3 GB frame buffer doubles the Quadro's 2 GB capacity for large textures or datasets. The transistor density difference (12.0M / mm² versus 5.9M / mm²) reflects the architectural efficiency leap across one process generation.

Quadro 4000M wins in power envelope efficiency for its class. At 100 W TDP, it consumes more power than the GTX 670MX's 75 W. The GTX 670MX therefore delivers higher performance with 25% lower TDP, making it the more efficient option in the recorded data.

In summary, the GTX 670MX is the clear performance leader in every recorded benchmark and most architectural metrics. The Quadro 4000M retains advantages only in memory bandwidth, ROP count, and its MXM module form factor, which may matter for specific mobile workstation integrations. For general-purpose compute, gaming, or rendering workloads, the GTX 670MX is the stronger part. For memory-bandwidth-sensitive applications on a mobile workstation platform, the Quadro 4000M offers a narrower but real edge.