Intel Arc A750 vs NVIDIA GeForce RTX 2080 SUPER Comparison

The NVIDIA GeForce RTX 2080 SUPER and Intel Arc A750 represent two very different approaches to GPU design, separated by three years of architectural evolution. The recorded data shows a clear overall winner in the RTX 2080 SUPER, which claims 9 out of 10 head-to-head benchmark victories, but the single loss is significant and reveals a distinct use case where the Intel card excels. The RTX 2080 SUPER sits at the 69th percentile among all GPUs, while the Arc A750 sits at the 66th, and their average benchmark scores of 24,170 and 20,582 respectively show a substantial gap in overall performance.

Where Each One Wins

The RTX 2080 SUPER dominates the majority of the test suite, winning in every category except one. Its strongest victories come in legacy DirectX workloads, where it more than doubles the Intel card's scores. In PassMark DirectX 11, the NVIDIA card scores 165 versus 72 for Intel, a 129.2% advantage. Similarly, in DirectX 10, the RTX 2080 SUPER scores 140 against 65, a 115.4% lead. These results indicate that for older game titles or applications relying on legacy graphics APIs, the Turing-based card is overwhelmingly superior.

The RTX 2080 SUPER also wins decisively in modern compute and general 3D workloads. In PassMark G3D, it scores 19,490 versus 12,534, a 55.5% margin. Its PassMark GPU Compute score of 8,290 beats Intel's 5,368 by 54.4%. These are not narrow wins; they represent a fundamental performance advantage in rasterization and general-purpose GPU tasks.

The Intel Arc A750's single victory comes in 3DMark Steel Nomad DX12, where it scores 2,612 against the RTX 2080 SUPER's 1,882, a 27.9% lead. This is the most modern benchmark in the suite, targeting DirectX 12 Ultimate features. This result suggests that in high-end, DX12-native workloads, the Arc A750's newer architecture can pull ahead significantly. However, this win does not carry over to other DX12 tests, as the RTX 2080 SUPER still wins PassMark DirectX 12 by 7.1% (75 versus 70).

The data also shows a split in API-specific performance. The RTX 2080 SUPER wins GeekBench Vulkan by 30% (111,284 versus 85,631), but the two cards are nearly tied in GeekBench OpenCL, with NVIDIA winning by just 0.7% (99,226 versus 98,554). This suggests that the NVIDIA card has a stronger Vulkan implementation, while OpenCL performance is effectively a draw.

Architecture Differences

The architectural divide between these two cards is stark. The RTX 2080 SUPER uses the TU104 chip built on Turing architecture, fabricated on a 12 nm process at TSMC. It contains 13,600 million transistors on a 545 mm² die, resulting in a transistor density of 25.0M per mm². In contrast, the Intel Arc A750 uses the DG2-512 chip based on Xe-HPG architecture (Alchemist generation), built on a much newer 6 nm process, also at TSMC. It packs 21,700 million transistors onto a smaller 406 mm² die, achieving a density of 53.4M per mm². This means Intel's design is more than twice as dense as NVIDIA's, reflecting the newer manufacturing node.

The core configurations differ substantially. The RTX 2080 SUPER has 3,072 shading units, 192 texture mapping units, and 64 ROPs. The Arc A750 has more of each: 3,584 shading units, 224 TMUs, and 112 ROPs. The Intel card also has 28 ray tracing cores, while NVIDIA has 48 RT cores. Notably, the RTX 2080 SUPER includes 384 tensor cores, a feature the Arc A750 completely lacks, which is important for AI-accelerated workloads.

Clock speeds also favor Intel. The Arc A750 has a base clock of 2050 MHz and a boost clock of 2400 MHz, while the RTX 2080 SUPER operates at 1650 MHz base and 1815 MHz boost. This clock advantage, combined with the higher core count, helps Intel achieve a theoretical FP32 throughput of 17.20 TFLOPS, compared to NVIDIA's 11.15 TFLOPS. The pixel rate is also much higher on Intel at 268.8 GPixel/s versus 116.2 GPixel/s, and texture rate is 537.6 GTexel/s versus 348.5 GTexel/s.

Memory configurations are similar but not identical. Both cards have 8 GB of GDDR6 memory on a 256-bit bus. The RTX 2080 SUPER runs its memory at 15.5 Gbps effective, yielding 495.9 GB/s bandwidth. The Arc A750 runs at 16 Gbps effective, yielding 512.0 GB/s. Intel's memory is slightly faster, but the difference is small.

The cards also differ in power requirements. The RTX 2080 SUPER has a TDP of 250 W and a suggested PSU of 600 W. The Arc A750 has a TDP of 225 W and a suggested PSU of 550 W. Both use dual-slot coolers and require 1x 6-pin plus 1x 8-pin power connectors. The bus interface differs, with NVIDIA using PCIe 3.0 x16 and Intel using PCIe 4.0 x16, which provides double the bandwidth for data transfers.

Head-to-Head Benchmarks

The largest win for the RTX 2080 SUPER is in PassMark DirectX 11, where it leads by 129.2%. This is followed closely by DirectX 10 at 115.4%. These legacy API tests show NVIDIA's mature driver stack and older architecture handling older workloads with ease. In DirectX 9, the NVIDIA card wins by 25.4% (227 versus 181), still a comfortable margin.

In the 3DMark Steel Nomad DX12 test, the tables turn. The Arc A750 wins by 27.9%, the only test where it leads. This is a modern, demanding workload that appears to favor Intel's Xe-HPG architecture. The margin is significant, indicating that in pure DX12 rasterization with no legacy compatibility constraints, Intel can outperform NVIDIA.

The GeekBench Vulkan test shows a 30% win for the RTX 2080 SUPER, with scores of 111,284 versus 85,631. This is a large gap in a modern API, suggesting NVIDIA's Vulkan driver is substantially more efficient than Intel's. However, in GeekBench OpenCL, the gap nearly disappears, with NVIDIA winning by just 0.7% (99,226 versus 98,554). This near-tie indicates that compute performance through OpenCL is very similar between the two cards.

The PassMark G3D and GPU Compute tests show NVIDIA winning by 55.5% and 54.4%, respectively. These are broad tests that exercise a variety of workloads, and the consistent margin of around 55% is notable. The PassMark DirectX 12 test is much closer, with NVIDIA winning by only 7.1% (75 versus 70), suggesting that in some DX12 scenarios the two cards are competitive.

Finally, the PassMark G2D test, which measures 2D graphics performance, shows NVIDIA winning by 25.7% (920 versus 732). This is not a gaming workload, but it indicates general desktop and UI rendering performance.

FAQ

Q: Which card is faster in overall average benchmark score?

A: The NVIDIA GeForce RTX 2080 SUPER has an average benchmark score of 24,170, while the Intel Arc A750 scores 20,582. The RTX 2080 SUPER is faster overall.

Q: Does the Intel Arc A750 win any benchmark?

A: Yes, it wins the 3DMark Steel Nomad DX12 test by 27.9%, scoring 2,612 versus 1,882 for the RTX 2080 SUPER.

Q: How do the two cards compare in legacy DirectX performance?

A: The RTX 2080 SUPER wins by 129.2% in DirectX 11 and 115.4% in DirectX 10. It also wins DirectX 9 by 25.4%.

Q: Which card has better Vulkan performance?

A: The RTX 2080 SUPER is significantly better, winning GeekBench Vulkan by 30% with a score of 111,284 versus 85,631.

Q: What is the difference in FP32 compute power?

A: The Intel Arc A750 has 17.20 TFLOPS FP32, while the RTX 2080 SUPER has 11.15 TFLOPS. Intel has a theoretical advantage.

Q: Do both cards support DirectX 12 Ultimate?

A: Yes, both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Specification Differences

The two cards differ in nearly every core specification. The process node is 12 nm for NVIDIA versus 6 nm for Intel. Transistor count is 13,600 million versus 21,700 million, and die size is 545 mm² versus 406 mm². Transistor density is 25.0M per mm² for NVIDIA and 53.4M per mm² for Intel.

The RTX 2080 SUPER has 3,072 shading units, 192 TMUs, and 64 ROPs. The Arc A750 has 3,584 shading units, 224 TMUs, and 112 ROPs. The RTX 2080 SUPER has 48 RT cores and 384 tensor cores, while the Arc A750 has 28 RT cores and no tensor cores.

Clock speeds are 1650/1815 MHz for NVIDIA and 2050/2400 MHz for Intel. Memory bandwidth is 495.9 GB/s for NVIDIA and 512.0 GB/s for Intel. Pixel rate is 116.2 GPixel/s versus 268.8 GPixel/s, and texture rate is 348.5 GTexel/s versus 537.6 GTexel/s. FP32 is 11.15 TFLOPS versus 17.20 TFLOPS.

The TDP is 250 W for NVIDIA and 225 W for Intel. The suggested PSU is 600 W versus 550 W. The bus interface is PCIe 3.0 x16 for NVIDIA and PCIe 4.0 x16 for Intel. Display outputs are 1x HDMI 2.0, 3x DisplayPort 1.4a, and 1x USB Type-C for NVIDIA, versus 1x HDMI 2.1 and 3x DisplayPort 2.0 for Intel.

The Verdict

The data indicates that the NVIDIA GeForce RTX 2080 SUPER is the stronger card for most workloads. It wins 9 of 10 benchmarks, with particularly large margins in legacy DirectX and general 3D compute. For users running older games, Vulkan titles, or general GPU compute tasks, the RTX 2080 SUPER is the clear choice. Its tensor cores also provide hardware acceleration for AI workloads that the Intel card cannot match.

The Intel Arc A750 is the better option for those specifically targeting modern DirectX 12 Ultimate workloads, as demonstrated by its 27.9% win in 3DMark Steel Nomad DX12. It also offers higher theoretical compute throughput (17.20 TFLOPS versus 11.15 TFLOPS), a faster bus interface (PCIe 4.0), and lower power consumption (225 W versus 250 W). However, these advantages do not translate into wins in the broader benchmark suite.

The RTX 2080 SUPER's average score of 24,170 places it near rivals like the GTX 780 Ti (24,236) and RX 6600 XT (24,442), while the Arc A750's 20,582 places it near the Arc B570 (20,556) and RTX 3070 Mobile (20,534). The RTX 2080 SUPER is the higher-performing card overall, and the recorded data supports choosing it for maximum compatibility and performance across varied workloads. The Arc A750 is a niche pick for users focused solely on the latest DX12 features, where it demonstrates a clear, measurable advantage.