GPU Comparison

Head-to-Head Benchmarks

The recorded head-to-head data shows a clear split between the two cards, with the Intel Arc A750 winning 4 of 10 tests and the NVIDIA GeForce RTX 2060 SUPER winning 6. The margins, however, tell a more nuanced story than the win count alone. The Arc A750 secures its largest victory in the 3DMark Steel Nomad DX12 test, scoring 2612 against 2011, a delta of 29.9%. This is a substantial lead in a modern, DirectX 12-focused workload, indicating that the Intel architecture performs strongly under current-generation API conditions. The Arc A750 also shows a 28.1% advantage in Geekbench OpenCL, posting 98554 versus 76957, and a 10.6% edge in Geekbench Vulkan, with 85631 against 77402. These compute-oriented wins suggest that raw throughput and API utilization favor the Intel card in scenarios that leverage asynchronous compute and modern driver paths.

The NVIDIA GeForce RTX 2060 SUPER counters with decisive victories in the Passmark legacy and rasterization suites. Its most pronounced win is in Passmark DirectX 11, where it scores 130 against the Arc A750's 72, a delta of 44.6%. Similarly, in DirectX 10, the NVIDIA card leads 111 to 65, a 41.4% gap. The Passmark DirectX 9 test also goes to NVIDIA, 218 to 181, a 17% margin. Beyond those API-specific tests, the RTX 2060 SUPER takes the Passmark G3D test with 16462 versus 12534, a 23.9% lead, and the G2D test with 854 versus 732, a 14.3% margin. The Passmark GPU Compute test favors NVIDIA as well, 6721 to 5368, a 20.1% difference. The only other Intel win is in Passmark DirectX 12, where the Arc A750 scores 70 against 61, a 14.8% margin.

What these numbers reveal is a generational and architectural divergence. The Intel card excels in modern, high-level DX12 and compute-heavy benchmarks, while the NVIDIA card dominates older DirectX versions and traditional 2D/3D rasterization tests. The average benchmark score confirms the overall picture: the Arc A750 sits at 20582, while the RTX 2060 SUPER averages 18093. This places the Intel part in the 66th percentile of all GPUs, while the NVIDIA card sits at the 62nd percentile. The nearest rivals for the Arc A750 include the Intel Arc B570 (average 20556, delta 0.1%) and the NVIDIA GeForce RTX 3070 Mobile (average 20534, delta 0.2%), showing that its aggregate performance is tightly clustered with those parts. The RTX 2060 SUPER's nearest rivals include the NVIDIA GeForce RTX 3060 Mobile (average 18159, delta -0.4%) and the AMD Radeon Pro 5700 (average 18189, delta -0.5%), placing it in a similar performance tier.

Architecture Differences

The two cards represent fundamentally different design philosophies. The Intel Arc A750 is built on the Xe-HPG architecture, using the DG2-512 chip fabricated on a 6 nm process at TSMC. It packs 21,700 million transistors into a 406 mm² die, yielding a transistor density of 53.4 million per mm². The NVIDIA GeForce RTX 2060 SUPER uses the Turing architecture with the TU106 chip on a 12 nm process, also from TSMC. It contains 10,800 million transistors on a larger 445 mm² die, giving a density of only 24.3 million per mm². The process node advantage for Intel is clear: the smaller 6 nm node allows nearly double the transistor density, which explains the Arc A750's higher compute throughput despite a smaller physical die.

In terms of core configuration, the Arc A750 fields 3584 shading units, 224 texture mapping units, and 112 render output units. The RTX 2060 SUPER has 2176 shading units, 136 TMUs, and 64 ROPs. This gives the Intel card a 64.7% advantage in shading units, a 64.7% advantage in TMUs, and a 75% advantage in ROPs. The pixel rate reflects this: the Arc A750 achieves 268.8 GPixel/s versus 105.6 GPixel/s for NVIDIA. Texture rate similarly favors Intel at 537.6 GTexel/s against 224.4 GTexel/s. The FP32 throughput is 17.20 TFLOPS for Intel versus 7.181 TFLOPS for NVIDIA, a 2.4x difference. FP16 performance follows the same pattern, with the Arc A750 at 34.41 TFLOPS (2:1) and the RTX 2060 SUPER at 14.36 TFLOPS (2:1). These raw compute metrics place the Intel card in a different performance class, though the benchmark results show that real-world software does not always translate that throughput into wins.

Ray tracing hardware differs in configuration. The Arc A750 has 28 ray tracing cores, while the RTX 2060 SUPER has 34. NVIDIA also includes 272 tensor cores, a feature absent from the Intel card's specification sheet. Clock speeds diverge as well: the Arc A750 runs at a base of 2050 MHz and a boost of 2400 MHz, while the RTX 2060 SUPER runs at 1470 MHz base and 1650 MHz boost. The Intel card's higher clocks, combined with its wider execution resources, explain its compute advantage. Memory subsystems are similar in capacity and type: both have 8 GB of GDDR6 on a 256-bit bus. The Arc A750, however, achieves a memory bandwidth of 512.0 GB/s using 2000 MHz memory (16 Gbps effective), while the RTX 2060 SUPER tops out at 448.0 GB/s with 1750 MHz memory (14 Gbps effective). This 14.3% bandwidth advantage for Intel likely contributes to its strong showing in memory-intensive compute benchmarks.

Power and interface characteristics differ as well. The Arc A750 has a TDP of 225 W and requires a 550 W suggested PSU, with 1x 6-pin and 1x 8-pin power connectors. The RTX 2060 SUPER has a lower TDP of 175 W, a 450 W suggested PSU, and a single 8-pin connector. The Intel card uses PCIe 4.0 x16, while NVIDIA uses PCIe 3.0 x16. Display outputs also differ: the Arc A750 offers 1x HDMI 2.1 and 3x DisplayPort 2.0, while the RTX 2060 SUPER provides 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, and 1x USB Type-C. Both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The NVIDIA card is physically smaller at 229 mm length, 113 mm height, and 35 mm width, while the Intel card's dimensions are not recorded in the database. Both are dual-slot designs.

Where Each One Wins

The data points to distinct usage profiles. The Intel Arc A750 is the clear choice for modern DirectX 12 workloads and compute-centric applications. Its 29.9% lead in 3DMark Steel Nomad DX12, a demanding next-generation test, indicates that games built around DX12 features will see a substantial performance uplift on the Intel card. The 28.1% advantage in Geekbench OpenCL and the 10.6% edge in Geekbench Vulkan further cement its position for GPU compute tasks, including rendering, machine learning inference (where OpenCL is common), and Vulkan-based titles. The 14.8% win in Passmark DirectX 12 reinforces this pattern. Users running modern engines, Vulkan ports, or general-purpose GPU compute should favor the Arc A750 based on the recorded scores.

The NVIDIA GeForce RTX 2060 SUPER, conversely, dominates legacy API performance and traditional rasterization. Its 44.6% lead in Passmark DirectX 11 and 41.4% lead in DirectX 10 make it the safer choice for older games, many of which still rely on DX11-era code paths. The 17% margin in DirectX 9 is also notable, though that API is largely obsolete. The 23.9% advantage in Passmark G3D, which aggregates gaming performance, suggests that the NVIDIA card retains an edge in the broader 3D rendering landscape. The 20.1% win in GPU compute, a Passmark-specific test, adds another feather to NVIDIA's cap, indicating that its compute efficiency per watt may be superior despite lower raw throughput. The 14.3% lead in G2D (2D graphics) is a niche but real advantage for desktop and windowed workloads.

The aggregate average scores put the Arc A750 ahead overall, 20582 versus 18093, a 13.8% difference. This suggests that for a balanced mix of modern and legacy tests, the Intel card edges out the NVIDIA card. However, the percentile rankings are close: 66th for Intel versus 62nd for NVIDIA. The nearest-rival data shows that the Arc A750 competes with the Intel Arc B570 and the RTX 3070 Mobile, while the RTX 2060 SUPER sits alongside the RTX 3060 Mobile and Radeon Pro 5700. This implies that the Intel card targets a slightly higher performance tier in the database's overall rankings, even if its legacy API performance lags.

The Verdict

The benchmark results indicate that the Intel Arc A750 is the stronger card for users running modern DirectX 12 and Vulkan applications, as well as compute-heavy workloads. Its 29.9% lead in 3DMark Steel Nomad DX12 and 28.1% lead in OpenCL are decisive, and its average score of 20582 places it 13.8% above the RTX 2060 SUPER's 18093. The Arc A750's architectural advantages, including a 6 nm process, higher core counts, and faster memory bandwidth, directly support these wins. For a user prioritizing current-generation games and GPU compute, the data clearly favors the Intel card.

The NVIDIA GeForce RTX 2060 SUPER remains the better option for legacy software compatibility and traditional DirectX 11-era gaming. Its 44.6% and 41.4% leads in DX11 and DX10, respectively, are substantial, and its Passmark G3D score of 16462 against 12534 shows a 23.9% advantage in aggregate 3D performance. The NVIDIA card also wins the Passmark GPU compute test by 20.1%, which may surprise given Intel's raw FP32 advantage, but the recorded data is unambiguous. For users with a library of older titles or software that has not adopted modern APIs, the RTX 2060 SUPER is the safer pick despite its lower overall average score. Its lower TDP of 175 W versus 225 W also makes it a more power-efficient option, though the database does not record measured power draw.

The launch MSRP for the Intel Arc A750 is 289 USD, while the NVIDIA GeForce RTX 2060 SUPER has a launch MSRP of 399 USD. Both cards are end-of-life in production status. The release dates show the Intel card launched on 2022-10-11, while the NVIDIA card launched on 2019-07-08, meaning the Intel part is a much newer design. The successor to the Arc A750 is Battlemage, while the RTX 2060 SUPER's successor is GeForce 30. The predecessor of the Arc A750 is Xe Graphics, and the RTX 2060 SUPER's predecessor is GeForce 10. For a new build focused on future-proofing with DX12 and Vulkan, the Arc A750 is the data-backed choice. For a system that must run a broad array of older software without issue, the RTX 2060 SUPER's legacy API wins make it the prudent selection.

FAQ

Q: Which card wins in DirectX 12 performance?

A: The Intel Arc A750 wins both the 3DMark Steel Nomad DX12 test (2612 versus 2011, a 29.9% margin) and the Passmark DirectX 12 test (70 versus 61, a 14.8% margin).

Q: How do the cards compare in compute workloads?

A: The Arc A750 leads Geekbench OpenCL by 28.1% (98554 versus 76957) and Geekbench Vulkan by 10.6% (85631 versus 77402), but the RTX 2060 SUPER wins Passmark GPU Compute by 20.1% (6721 versus 5368).

Q: What is the average benchmark score difference?

A: The Intel Arc A750 averages 20582, while the NVIDIA GeForce RTX 2060 SUPER averages 18093, a 13.8% advantage for the Intel card.

Q: Which card is better for older DirectX 9 and DirectX 11 games?

A: The RTX 2060 SUPER wins Passmark DirectX 9 by 17% (218 versus 181) and Passmark DirectX 11 by 44.6% (130 versus 72), making it the stronger choice for legacy titles.

Q: What are the memory bandwidth figures for each card?

A: The Arc A750 has a memory bandwidth of 512.0 GB/s (2000 MHz, 16 Gbps effective), while the RTX 2060 SUPER has 448.0 GB/s (1750 MHz, 14 Gbps effective). Both use 8 GB of GDDR6 on a 256-bit bus.

Q: How do the transistor counts and process nodes compare?

A: The Arc A750 uses a 6 nm process with 21,700 million transistors on a 406 mm² die, while the RTX 2060 SUPER uses a 12 nm process with 10,800 million transistors on a 445 mm² die. The Intel card has a transistor density of 53.4M per mm² versus 24.3M per mm² for NVIDIA.