GPU Comparison

# NVIDIA GeForce GTX 675MX vs NVIDIA T600

The NVIDIA GeForce GTX 675MX, a Kepler-based mobile part from 2012, and the NVIDIA T600, a Turing-based workstation card from 2021, represent two very different design eras. In the recorded head-to-head benchmark, the T600 dominates decisively, but the GTX 675MX still holds relevance in specific legacy scenarios. The data shows a single benchmark win for the T600, with the GTX 675MX failing to secure any wins in the direct comparison. This is not a close contest; the T600 is the clear performance leader in modern workloads, while the GTX 675MX remains a capable option for older applications that rely on its particular feature set.

Where Each One Wins

The use-case split is stark. The NVIDIA T600 wins the only direct head-to-head benchmark recorded, the Geekbench OpenCL test, and does so by a massive margin. With a score of 27,875 versus 10,723, the T600 is 61.5 percent ahead of the GTX 675MX in that test. This indicates that for compute-oriented tasks, OpenCL acceleration, and modern GPU-accelerated workflows, the T600 is the definitive choice. Its Turing architecture, higher memory bandwidth, and newer feature set give it a substantial edge in any application that leverages contemporary APIs or compute frameworks.

The GTX 675MX, by contrast, has no head-to-head wins in the database. Its strengths lie in its raw architectural characteristics rather than benchmark victories. It carries 960 shading units, 80 texture mapping units, and 32 ROPs, which are higher shader and TMU counts than the T600's 640 shading units and 40 TMUs. In older DirectX 11 or OpenGL titles that were designed around Kepler-era hardware, the GTX 675MX could be expected to perform competitively, though the recorded data does not directly confirm this. The GTX 675MX also has a wider 256-bit memory bus, which historically benefits bandwidth-sensitive legacy games, even though its total memory bandwidth of 115.2 GB/s is lower than the T600's 160.0 GB/s. For users running very old software that does not support newer APIs, the GTX 675MX's architecture may be more familiar and stable, but the data cannot validate this advantage.

Architecture Differences

The two GPUs come from different architectural generations and are built for different purposes. The GTX 675MX uses the GK104 chip, built on the Kepler architecture, fabricated on a 28 nm process at TSMC. It integrates 3,540 million transistors on a die size of 294 mm², yielding a transistor density of 12.0M per mm². The T600 uses the TU117 chip, based on the Turing architecture, fabricated on a 12 nm process, also at TSMC. It packs 4,700 million transistors into a smaller 200 mm² die, achieving a much higher transistor density of 23.5M per mm². This density difference is a clear indicator of the generational leap in manufacturing and design efficiency.

Memory architecture also diverges significantly. The GTX 675MX uses 2 GB of GDDR5 memory on a 256-bit bus, with a memory clock of 900 MHz (3.6 Gbps effective) and bandwidth of 115.2 GB/s. The T600 uses 4 GB of GDDR6 memory on a 128-bit bus, with a memory clock of 1250 MHz (10 Gbps effective) and bandwidth of 160.0 GB/s. Despite having half the bus width, the T600 delivers 38.9 percent more bandwidth due to its faster memory technology. The T600 also has a higher pixel rate (42.72 GPixel/s versus 13.08 GPixel/s) and a slightly higher texture rate (53.40 GTexel/s versus 52.32 GTexel/s), despite having fewer TMUs. Its FP32 compute is rated at 1.709 TFLOPS, which is 36 percent higher than the GTX 675MX's 1,255.7 GFLOPS, and it also supports FP16 at 3.418 TFLOPS (2:1), which the GTX 675MX does not have.

Power and form factor further separate the two. The GTX 675MX has a TDP of 100 W and uses an MXM module slot, while the T600 has a TDP of 40 W, is a single-slot card, and uses a PCIe 3.0 x16 interface. The T600's power efficiency is a major advantage, delivering higher performance at less than half the power draw. The T600 also supports a newer DirectX version (12_1 versus 11_0) and a much newer Vulkan version (1.4 versus 1.2.175). Both support OpenGL 4.6.

Head-to-Head Benchmarks

The only recorded head-to-head benchmark is Geekbench OpenCL, and it is a blowout. The T600 scores 27,875, while the GTX 675MX scores 10,723. The delta is -61.5 percent from the perspective of the GTX 675MX, meaning the T600 is 61.5 percent faster. This is a massive gap that reflects the architectural and process node differences. The T600's 160.0 GB/s of memory bandwidth, 1.709 TFLOPS of FP32 compute, and Turing-era design all contribute to this result. The GTX 675MX's 115.2 GB/s bandwidth and 1,255.7 GFLOPS of FP32 compute are simply not enough to compete in this modern compute workload.

Beyond the head-to-head, the database provides additional context through average benchmark scores and percentile rankings. The GTX 675MX has an average benchmark score of 8,427, placing it in the 43rd percentile of all GPUs. The T600 has an average score of 7,035, placing it in the 39th percentile. This is counterintuitive given the head-to-head result, but it reflects the different benchmark suites each card was subjected to. The T600's average includes Passmark tests (DirectX 9, 10, 11, 12, G2D, G3D, and GPU compute) where it scores 114, 32, 49, 25, 756, 6,479, and 2,402 respectively. The GTX 675MX's average is based on Geekbench Metal (6,131) and Geekbench OpenCL (10,723). The T600's average is dragged down by low Passmark DirectX scores, while its Geekbench OpenCL score is far above the GTX 675MX's. The T600 also has a Geekbench Vulkan score of 25,580, which the GTX 675MX cannot match due to its older architecture.

Specification Differences

The key specification differences, based solely on the recorded data, are as follows:

• Process Node: 28 nm (GTX 675MX) versus 12 nm (T600)
• Transistors: 3,540 million versus 4,700 million
• Die Size: 294 mm² versus 200 mm²
• Transistor Density: 12.0M / mm² versus 23.5M / mm²
• Memory Size: 2 GB GDDR5 versus 4 GB GDDR6
• Memory Bus: 256 bit versus 128 bit
• Memory Clock: 900 MHz (3.6 Gbps effective) versus 1250 MHz (10 Gbps effective)
• Memory Bandwidth: 115.2 GB/s versus 160.0 GB/s
• Shading Units: 960 versus 640
• TMUs: 80 versus 40
• Pixel Rate: 13.08 GPixel/s versus 42.72 GPixel/s
• Texture Rate: 52.32 GTexel/s versus 53.40 GTexel/s
• FP32: 1,255.7 GFLOPS versus 1.709 TFLOPS
• FP16: Not available versus 3.418 TFLOPS (2:1)
• TDP: 100 W versus 40 W
• Slot Width: MXM Module versus Single-slot
• Bus Interface: MXM-B (3.0) versus PCIe 3.0 x16
• Display Outputs: Portable Device Dependent versus 4x mini-DisplayPort 1.4a
• DirectX Support: 12 (11_0) versus 12 (12_1)
• Vulkan Support: 1.2.175 versus 1.4
• Release Date: 2012-09-30 versus 2021-04-11

FAQ

Q: Which GPU has higher memory bandwidth?

A: The NVIDIA T600 has higher memory bandwidth at 160.0 GB/s, compared to the GTX 675MX's 115.2 GB/s, despite the GTX 675MX having a wider 256-bit bus.

Q: How many shading units does each GPU have?

A: The GTX 675MX has 960 shading units, while the T600 has 640 shading units.

Q: Which card has a lower power draw?

A: The T600 has a TDP of 40 W, significantly lower than the GTX 675MX's 100 W.

Q: What is the difference in FP32 compute performance?

A: The T600 delivers 1.709 TFLOPS of FP32 compute, which is higher than the GTX 675MX's 1,255.7 GFLOPS.

Q: Which GPU supports DirectX 12_1?

A: The T600 supports DirectX 12 (12_1), while the GTX 675MX supports DirectX 12 (11_0).

Q: What was the result of the Geekbench OpenCL head-to-head test?

A: The T600 scored 27,875, which is 61.5 percent higher than the GTX 675MX's score of 10,723.

The Verdict

The data is unambiguous: the NVIDIA T600 is the superior GPU in computed performance, memory bandwidth, and power efficiency. Its 61.5 percent lead in Geekbench OpenCL, combined with 36 percent higher FP32 throughput and 38.9 percent more memory bandwidth, makes it the clear choice for modern compute workloads, professional applications, and any task that benefits from DirectX 12_1 or Vulkan 1.4 support. Its 40 W TDP is a major advantage for compact or low-power systems, and its 4 GB of GDDR6 memory doubles the capacity of the GTX 675MX.

The GTX 675MX, while lacking benchmark wins, still has a place in the database for users with legacy systems. Its higher shader unit count (960 versus 640) and wider memory bus (256-bit) may offer advantages in very old software that was optimized for Kepler hardware. Its 43rd percentile ranking versus the T600's 39th percentile also suggests that in the full suite of recorded tests, the GTX 675MX holds its own in some legacy scenarios. However, the T600's 40th percentile (rounded from 39) is based on a broader test suite, and its head-to-head victory is decisive.

For anyone choosing between these two, the T600 is the obvious pick for all modern applications. The GTX 675MX should only be considered if compatibility with an MXM module system is required and the workload is exclusively legacy. Based on the recorded data, the T600 outperforms the GTX 675MX in every measurable modern benchmark, and its architectural advantages ensure it will remain viable for years to come.