Intel UHD Graphics 750 vs NVIDIA GeForce GTX 675M Comparison

The Verdict

The benchmark data positions the NVIDIA GeForce GTX 675M as the stronger choice for raw compute workloads, but the margin is slim. In the single available head-to-head test (Geekbench OpenCL), the GTX 675M scores 6946 against the Intel UHD Graphics 750's 6743, a 2.9% advantage. The GTX 675M also holds a higher average benchmark score of 6946 versus 7441 for the Intel part—wait, that is reversed: the Intel part's average is 7441, which is actually higher. This discrepancy matters. The Intel UHD Graphics 750's average score of 7441 places it in the 40th percentile of all GPUs, while the GTX 675M's 6946 average sits at the 39th percentile. The Intel part's average is pulled up by its Vulkan score of 8138, a test the GTX 675M has no data for. For users running OpenCL-only workloads, the GTX 675M is marginally ahead. For those who can leverage Vulkan, the Intel part demonstrates a significant advantage. The data suggests that the GTX 675M is the safer pick for legacy compute compatibility, while the Intel UHD Graphics 750 is better suited for modern API adoption. Neither card is a performance leader; both sit below the 41st percentile, but the Intel part's higher average and Vulkan capability make it the more future-proof option.

Architecture Differences

The two GPUs represent fundamentally different design philosophies separated by nearly a decade of silicon evolution. The Intel UHD Graphics 750 is built on Rocket Lake silicon using Intel's 14 nm+++ process node, a refined version of a mature manufacturing technology. It is an integrated graphics processor (IGP) with a Ring Bus interface, meaning it shares system memory and has no dedicated VRAM. Its architecture is Generation 12.1, featuring 256 shading units, 16 texture mapping units, and 8 raster output units. The GTX 675M, in contrast, is a discrete mobile GPU from NVIDIA's Fermi 2.0 architecture, fabricated on TSMC's 40 nm node. It uses the GF114 chip, which contains 1,950 million transistors on a 332 mm² die, translating to a transistor density of 5.9M per mm². The Fermi part fields 384 shading units, 64 TMUs, and 32 ROPs—significantly more hardware resources than the Intel IGP.

The memory subsystems diverge completely. The Intel part relies on System Shared memory, with bandwidth listed as "System Dependent," making its performance highly variable based on the host platform's RAM configuration. The GTX 675M features 2 GB of dedicated GDDR5 memory on a 256-bit bus, providing a fixed 96.00 GB/s bandwidth. This dedicated memory is a critical architectural advantage for the NVIDIA part, as it eliminates contention with the CPU for memory access. The Intel part's power envelope is drastically lower at 15 W TDP versus the GTX 675M's 100 W TDP, reflecting the IGP's integration into the CPU package against the discrete MXM Module form factor. Clock behavior also differs: the Intel part has a base clock of 300 MHz and a boost of 1300 MHz, while the GTX 675M's core clocks are not listed, only its memory clock of 750 MHz (3 Gbps effective). The API support shows the Intel part is more modern, with DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, while the GTX 675M supports DirectX 12 (11_0) and OpenGL 4.6 but has no Vulkan support listed.

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test, where the NVIDIA GeForce GTX 675M wins with a score of 6946 against the Intel UHD Graphics 750's 6743. The delta is -2.9% from the Intel part's perspective, meaning the GTX 675M is 2.9% faster in this specific workload. This is a narrow margin—well within the range of run-to-run variance that one might expect from benchmark results. The GTX 675M's nearest rivals in the data include the AMD Radeon R5 M240 (6975, -0.4% relative to the GTX) and the NVIDIA GeForce GTX 680M (7023, -1.1%), showing that the GTX 675M sits just below those parts. The Intel UHD Graphics 750's nearest rivals include the AMD Radeon HD 8850M (7447, -0.1%) and the NVIDIA GeForce GTX 1650 (7472, -0.4%), placing the Intel part in a slightly higher performance tier on average.

The Intel part's average benchmark score of 7441 is 7.1% higher than the GTX 675M's 6946 average. This difference is driven entirely by the Intel part's Vulkan score of 8138, which is its stronger result. The GTX 675M has no Vulkan benchmark data, so its average is based solely on the OpenCL score. In the OpenCL test specifically, the GTX 675M's 6946 exceeds the Intel part's 6743 by 203 points. The Intel part's OpenCL score of 6743 is actually lower than its Vulkan score by 1395 points, indicating that its architecture performs better under Vulkan. The GTX 675M's OpenCL score of 6946 places it slightly closer to its nearest rival, the AMD Radeon R5 M240 (6975), than the Intel part's OpenCL score is to its nearest rival, the AMD Radeon HD 8850M (7447). This suggests that in OpenCL, the GTX 675M is more competitive within its peer group than the Intel part is within its own.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The Intel UHD Graphics 750 has a higher average benchmark score of 7441, compared to the NVIDIA GeForce GTX 675M's 6946. This places the Intel part in the 40th percentile of all GPUs versus the 39th percentile for the NVIDIA part.

Q: In the direct head-to-head OpenCL test, who wins?

A: The NVIDIA GeForce GTX 675M wins the Geekbench OpenCL test with a score of 6946 against the Intel UHD Graphics 750's 6743, representing a 2.9% advantage for the NVIDIA part.

Q: Does the Intel UHD Graphics 750 support Vulkan?

A: Yes, the Intel UHD Graphics 750 supports Vulkan 1.4. Its Geekbench Vulkan score is 8138, which is its higher benchmark result. The NVIDIA GeForce GTX 675M has no Vulkan support listed in the data.

Q: What is the memory configuration difference?

A: The Intel UHD Graphics 750 uses System Shared memory with a System Dependent bandwidth, while the NVIDIA GeForce GTX 675M has 2 GB of dedicated GDDR5 memory on a 256-bit bus with a fixed bandwidth of 96.00 GB/s.

Q: How do their transistor counts compare?

A: The NVIDIA GeForce GTX 675M's GF114 chip contains 1,950 million transistors on a 332 mm² die. The Intel UHD Graphics 750's transistor count is not listed in the data.

Q: What are the power consumption figures?

A: The Intel UHD Graphics 750 has a TDP of 15 W, while the NVIDIA GeForce GTX 675M has a TDP of 100 W. The Intel part is an integrated GPU, whereas the NVIDIA part is a discrete MXM Module.

Where Each One Wins

The NVIDIA GeForce GTX 675M wins in the direct OpenCL compute benchmark, scoring 6946 versus the Intel UHD Graphics 750's 6743. This indicates that for applications relying on OpenCL—a common compute API for older software and certain professional tools—the GTX 675M holds a measurable, though small, advantage. Its dedicated 2 GB GDDR5 memory with 96.00 GB/s bandwidth also gives it a structural edge in scenarios where memory bandwidth is the limiting factor, since the Intel part's bandwidth is "System Dependent" and subject to platform variability. The GTX 675M's higher shading unit count (384 vs 256), TMU count (64 vs 16), and ROP count (32 vs 8) suggest it should handle texture-heavy and fill-rate-limited workloads more effectively, despite its lower pixel rate (9.920 GPixel/s vs 10.40 GPixel/s) being slightly below the Intel part.

The Intel UHD Graphics 750 wins in overall average benchmark performance, with 7441 versus 6946, driven by its Vulkan score of 8138. For modern applications that support Vulkan, the Intel part demonstrates a significant 20.7% advantage over its own OpenCL score, indicating that its architecture is better optimized for this newer API. The Intel part also has a higher pixel rate (10.40 GPixel/s vs 9.920 GPixel/s) and a higher boost clock (1300 MHz), which could translate to better performance in lightweight, resolution-sensitive tasks. Its 15 W TDP makes it suitable for systems with limited thermal and power budgets, while the GTX 675M's 100 W TDP requires more substantial cooling. The Intel part's DirectX 12 (12_1) support is also more feature-complete than the GTX 675M's DirectX 12 (11_0), which may matter for newer games that leverage DirectX 12 Ultimate features.

Specification Differences

The two GPUs differ across nearly every measured specification. The Intel UHD Graphics 750 is manufactured on a 14 nm+++ process by Intel, while the NVIDIA GeForce GTX 675M uses a 40 nm process from TSMC. The Intel part has no listed transistor count or die size, whereas the GTX 675M contains 1,950 million transistors on a 332 mm² die with a density of 5.9M per mm². In terms of core configuration, the Intel part has 256 shading units, 16 TMUs, and 8 ROPs; the GTX 675M has 384 shading units, 64 TMUs, and 32 ROPs. The clock speeds differ, with the Intel part listing a 300 MHz base and 1300 MHz boost, while the GTX 675M only lists a memory clock of 750 MHz (3 Gbps effective) with no core clocks provided.

Memory specifications are entirely different: the Intel part uses System Shared memory with no fixed size, type, bus width, or bandwidth (listed as "System Dependent"), while the GTX 675M has 2 GB of GDDR5 on a 256-bit bus with 96.00 GB/s bandwidth. The compute rates show a mixed picture: the Intel part has a pixel rate of 10.40 GPixel/s and texture rate of 20.80 GTexel/s, while the GTX 675M has a pixel rate of 9.920 GPixel/s and texture rate of 39.68 GTexel/s. The FP32 performance is 665.6 GFLOPS for the Intel part versus 952.3 GFLOPS for the GTX 675M. The Intel part supports FP16 at 1,331.2 GFLOPS (2:1 ratio), while the GTX 675M has no FP16 data. Power consumption is 15 W for the Intel part versus 100 W for the GTX 675M. The form factor is IGP for Intel versus MXM Module for NVIDIA, with the bus interface being Ring Bus for Intel and MXM-B (3.0) for NVIDIA. Display outputs are Motherboard Dependent for the Intel part and Portable Device Dependent for the GTX 675M. API support differs: the Intel part supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, while the GTX 675M supports DirectX 12 (11_0) and OpenGL 4.6 but has no Vulkan support. The GTX 675M has a predecessor (GeForce 500M) and successor (GeForce 700M), while the Intel part has neither listed. Both are end-of-life products, with the Intel part released on 2021-03-29 and the GTX 675M on 2012-03-21.