Intel Iris Pro Graphics 6200 vs NVIDIA GeForce GTX 670MX Comparison

The NVIDIA GeForce GTX 670MX and Intel Iris Pro Graphics 6200 represent two fundamentally different approaches to mobile graphics, and the benchmark data shows they are far closer in overall performance than their architectural differences suggest. While the GTX 670MX is a discrete part from NVIDIA’s Kepler generation aimed at gaming laptops, the Iris Pro 6200 is an integrated GPU from Intel’s Broadwell generation. The data shows that each card wins one of the two head-to-head tests, with the GTX 670MX taking a decisive victory in OpenCL and the Iris Pro 6200 countering in Vulkan. This split makes the choice between them entirely dependent on the workload.

Head-to-Head Benchmarks

The most striking result in the data is the Geekbench OpenCL score, where the NVIDIA GeForce GTX 670MX scores 6,125 against the Intel Iris Pro Graphics 6200’s 4,556. That is a 34.4% advantage for the NVIDIA part, a massive gap that highlights the discrete GPU’s raw compute advantage. The GTX 670MX’s average benchmark score of 5,721 is also telling, as it sits within 0.2% of the GeForce GTX 550 Ti and 0.5% of the AMD Radeon HD 8790M, indicating it trades blows with those older desktop and mobile parts. In OpenCL, the GTX 670MX’s 960 shading units and 1,153.9 GFLOPS of FP32 performance are clearly leveraged, while the Iris Pro’s 384 shading units and 844.8 GFLOPS cannot keep up.

The Vulkan test flips the script. Here, the Intel Iris Pro Graphics 6200 scores 6,032, beating the GTX 670MX’s 5,316 by 11.9%. This is a significant upset, especially considering that the GTX 670MX is a discrete GPU with dedicated GDDR5 memory and a 192-bit bus. The Iris Pro’s average benchmark score of 6,117 is also higher than the GTX 670MX’s 5,721, and its nearest rival list includes the NVIDIA GeForce MX230 at a 0.7% delta, showing it competes with much newer entry-level discrete parts. The Vulkan result suggests that Intel’s driver implementation for the Iris Pro 6200 is particularly strong, or that the workload favors its architecture’s specific strengths.

When looking at the broader picture, the GTX 670MX’s average score of 5,721 places it in the 33rd percentile of all GPUs, while the Iris Pro 6200’s 6,117 places it in the 35th percentile. This means the Intel part is actually ranked slightly higher globally, despite losing the OpenCL test. The head-to-head data shows a 1-1 split in wins, but the delta percentages are lopsided: the GTX 670MX’s OpenCL win is 34.4%, while the Iris Pro’s Vulkan win is only 11.9%. This suggests that while the Intel part is more consistent overall, the NVIDIA part has a higher ceiling in compute-heavy tasks.

Where Each One Wins

The NVIDIA GeForce GTX 670MX is the clear winner in OpenCL-based workloads, which typically include general-purpose compute tasks, physics simulations, and some professional applications. Its 34.4% lead in that test is backed by its hardware: 960 shading units, 80 TMUs, and a texture rate of 48.08 GTexel/s. The card’s pixel rate of 12.02 GPixel/s is also nearly double the Iris Pro’s 6.60 GPixel/s, which matters for fill-rate-bound scenarios. If the workload is compute-oriented and uses OpenCL, the GTX 670MX is the only rational choice between these two.

The Intel Iris Pro Graphics 6200 wins in Vulkan, which is the modern low-overhead graphics API. Its 11.9% lead in that test is notable, but the more interesting data point is its overall consistency. The Iris Pro scores 6,032 in Vulkan and 7,764 in Geekbench Metal, a test the GTX 670MX does not even have a result for. This suggests the Intel part has better driver support for modern APIs and is more future-proof in that regard. The Iris Pro’s texture rate of 52.80 GTexel/s is actually higher than the GTX 670MX’s 48.08 GTexel/s, which is a surprise given the NVIDIA part’s larger memory bandwidth of 67.20 GB/s versus the Intel’s system-dependent shared memory.

For gaming specifically, the Vulkan win is significant because many modern titles use Vulkan or DirectX 12. The Iris Pro supports DirectX 12 (11_1) and Vulkan 1.0, while the GTX 670MX supports DirectX 12 (11_0) and Vulkan 1.2.175. The NVIDIA part has the newer Vulkan version, but the Intel part still wins the benchmark. The Iris Pro’s higher average score and percentile ranking suggest it is the more balanced performer, while the GTX 670MX is a specialist that dominates one API but falls behind in another.

Architecture Differences

The architectural gap between these two is generational and fundamental. The NVIDIA GeForce GTX 670MX uses the GK104 chip built on a 28 nm process at TSMC, with 3,540 million transistors on a 294 mm² die. This gives it a transistor density of 12.0M per mm². It is a Kepler architecture part from the GeForce 600M generation, released in late 2012. The Intel Iris Pro Graphics 6200, by contrast, uses the Broadwell GT3e chip built on Intel’s 14 nm process, with no transistor count or die size listed in the data. It is a Generation 8.0 architecture part from the HD Graphics (Broadwell) generation, released in late 2014.

The process node difference is stark: 28 nm versus 14 nm. This gives Intel a significant efficiency advantage, which is reflected in the TDP. The GTX 670MX has a TDP of 75 W, while the Iris Pro 6200 has a TDP of just 15 W. That is a 5x difference in power draw, and the Iris Pro achieves a higher average benchmark score while consuming a fraction of the power. The GTX 670MX has no power connectors listed, suggesting it is a low-power discrete part, but it still dwarfs the Intel IGP in power consumption.

The shading unit count is also wildly different. The GTX 670MX has 960 shading units, 80 TMUs, and 24 ROPs. The Iris Pro 6200 has 384 shading units, 48 TMUs, and only 6 ROPs. The NVIDIA part’s ROP count is four times higher, which explains its pixel rate advantage. However, the Intel part’s texture rate is higher, which is unusual given the lower TMU count. This suggests the Broadwell GT3e chip has a more efficient texture pipeline per TMU.

Memory architecture is another major divider. The GTX 670MX has 3 GB of dedicated GDDR5 memory on a 192-bit bus, with 67.20 GB/s of bandwidth. The Iris Pro 6200 uses system shared memory, with a system-dependent bandwidth. This means the NVIDIA part has predictable and dedicated memory performance, while the Intel part’s memory performance depends entirely on the system’s RAM configuration. The GTX 670MX’s memory clock is listed as 700 MHz with 2.8 Gbps effective, while the Intel part’s memory clock is simply "System Shared."

Specification Differences

The most consequential specification difference is TDP: 75 W for the GTX 670MX versus 15 W for the Iris Pro 6200. This alone determines which systems these parts belong in. The GTX 670MX is a discrete add-in part for larger laptops, while the Iris Pro 6200 is an IGP soldered into the CPU package.

Clock speeds differ significantly. The GTX 670MX has a base clock of 601 MHz and a boost clock of 601 MHz, meaning it does not boost at all. The Iris Pro 6200 has a base clock of 300 MHz and a boost clock of 1,100 MHz. The Intel part’s boost is nearly double its base, which is a common pattern for IGPs that need to conserve power when idle. The NVIDIA part runs at a constant 601 MHz, which is a simpler but less efficient approach.

Memory specifications are also split. The GTX 670MX has 3 GB of GDDR5 with a 192-bit bus and 67.20 GB/s bandwidth. The Iris Pro 6200 has system shared memory, which means the size, type, bus width, and bandwidth are all system dependent. This gives the NVIDIA part a fixed performance profile, while the Intel part’s memory performance is variable and depends on the system’s RAM speed and dual-channel configuration.

The bus interface differs: the GTX 670MX uses PCIe 3.0 x16, while the Iris Pro uses a Ring Bus. The NVIDIA part is a true PCIe card, while the Intel part is integrated into the CPU die and communicates via the ring bus. Display outputs are also different: the GTX 670MX is listed as "Portable Device Dependent," while the Iris Pro is "Motherboard Dependent," reflecting their integration levels.

API support shows a split. The GTX 670MX supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The Iris Pro supports DirectX 12 (11_1), OpenGL 4.4, and Vulkan 1.0. The NVIDIA part has a newer OpenGL and Vulkan version, but the Intel part has a slightly newer DirectX feature level. Both are end-of-life products, with the GTX 670MX having a successor in the GeForce 700M series.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The Intel Iris Pro Graphics 6200 has an average benchmark score of 6,117, which is higher than the NVIDIA GeForce GTX 670MX’s 5,721. The Intel part also has a higher percentile ranking at 35 versus 33.

Q: How large is the OpenCL performance gap?

A: The GTX 670MX scores 6,125 in Geekbench OpenCL, while the Iris Pro scores 4,556. That is a 34.4% advantage for the NVIDIA part, making it the clear winner for OpenCL workloads.

Q: Which GPU wins in Vulkan?

A: The Intel Iris Pro Graphics 6200 wins in Vulkan, scoring 6,032 against the GTX 670MX’s 5,316, a delta of 11.9%. This makes the Intel part the better choice for Vulkan-based applications and games.

Q: What is the TDP difference between these two?

A: The NVIDIA GeForce GTX 670MX has a TDP of 75 W, while the Intel Iris Pro Graphics 6200 has a TDP of just 15 W. The Intel part consumes 60 W less power while achieving a higher average benchmark score.

Q: Which GPU has dedicated video memory?

A: The NVIDIA GeForce GTX 670MX has 3 GB of dedicated GDDR5 memory on a 192-bit bus with 67.20 GB/s bandwidth. The Intel Iris Pro uses system shared memory, so its bandwidth is system dependent.

Q: What are the shading unit counts?

A: The GTX 670MX has 960 shading units, while the Iris Pro has 384. The NVIDIA part also has 80 TMUs and 24 ROPs, compared to the Intel part’s 48 TMUs and 6 ROPs.

The Verdict

The data paints a clear picture for specific use cases. If the workload is OpenCL-based, the NVIDIA GeForce GTX 670MX is the only choice. Its 34.4% lead in that test is decisive, and its higher pixel rate of 12.02 GPixel/s versus 6.60 GPixel/s gives it a strong advantage in fill-rate-bound tasks. The 3 GB of dedicated GDDR5 memory also makes it more reliable for applications that need predictable memory performance. This is the part for compute-heavy applications and older games that rely on OpenCL.

If the workload is Vulkan-based, the Intel Iris Pro Graphics 6200 is the better performer, with an 11.9% lead. Its higher average benchmark score of 6,117 and 35th percentile ranking also indicate it is the more balanced overall GPU. The 15 W TDP makes it dramatically more power-efficient, and the fact that it is an IGP means it costs nothing extra in a system. For modern games that use Vulkan or DirectX 12, the Iris Pro is the safer bet.

The real verdict is about context. The GTX 670MX is a discrete part for gaming laptops from 2012, and its 75 W TDP means it requires a larger chassis and better cooling. The Iris Pro 6200 is an integrated GPU from 2014 that sips power and fits in ultrabooks. The benchmark scores show they are in the same performance tier overall, with the Intel part actually ranking slightly higher. Given the 5x TDP difference and the Iris Pro’s higher average score, the Intel part is the more sensible choice for most users. The GTX 670MX only makes sense if OpenCL compute is the primary workload, as its 34.4% lead in that specific test is too large to ignore.