Intel Arc A380 vs NVIDIA GeForce GTX 1070 Comparison

Head-to-Head Benchmarks

The recorded data shows a clear and decisive overall advantage for the NVIDIA GeForce GTX 1070, which wins 9 of the 10 shared benchmark tests. The average benchmark score for the GTX 1070 is 9780, while the Intel Arc A380 sits at 8558, a gap of roughly 14% in favor of the older NVIDIA part. In terms of global standing, the GTX 1070 lands in the 47th percentile of all GPUs, with the Arc A380 just behind at the 44th percentile.

The largest single victory for the GTX 1070 comes in the PassMark DirectX 9 test, where it scores 197 against the Arc A380's 73, a delta of 169.9%. This is followed closely by the PassMark DirectX 11 result, where the GTX 1070 posts 100 versus 38, a 163.2% advantage. The pattern holds across other legacy and modern API tests: in PassMark DirectX 10, the GTX 1070 leads 82 to 37 (121.6% ahead), and in PassMark DirectX 12, it wins 48 to 35 (37.1% ahead). The PassMark G3D test, which is a strong indicator of general 3D rendering performance, shows the GTX 1070 at 13498 compared to the Arc A380's 6252, a 115.9% lead. Compute workloads also favor NVIDIA heavily, with the PassMark GPU Compute score of 6102 versus 2762, a 120.9% difference.

The 3DMark Steel Nomad DX12 test, a more modern and demanding workload, also goes to the GTX 1070 with a score of 1082 against 808, a 33.9% margin. In the GeekBench OpenCL test, the GTX 1070 scores 44700, which is 16.9% higher than the Arc A380's 38224. Even the 2D-focused PassMark G2D test shows a win for NVIDIA at 846 versus 610, a 38.7% difference.

The one test where the Intel Arc A380 emerges victorious is the GeekBench Vulkan benchmark. Here, the Arc A380 scores 36736, which is 39.8% higher than the GTX 1070's 22121. This is a substantial margin and indicates that the Arc A380's Vulkan driver and architecture are particularly well-suited to this specific workload. However, this single win does not offset the broader trend across the other nine tests, where the GTX 1070 frequently doubles or more than doubles the Arc A380's output.

Looking at nearest rivals for context, the GTX 1070's average score of 9780 places it essentially even with the AMD FirePro W5000 (9803, delta -0.2%) and slightly ahead of the NVIDIA Quadro M2000M (9832, delta -0.5%) and Tesla M10 (9724, delta 0.6%). The Arc A380's average of 8558 is similarly clustered around its own set of rivals, including the AMD FirePro W5170M (8595, delta -0.4%) and AMD Radeon HD 8870M (8462, delta 1.1%). This positioning suggests that while the two cards are not directly competing in the same performance tier, the GTX 1070 operates at a substantially higher performance level overall.

Architecture Differences

The fundamental architectural gap between these two GPUs is vast, reflecting their different release periods and design goals. The NVIDIA GeForce GTX 1070 is built on the Pascal architecture using the GP104 chip, manufactured on a 16 nm process at TSMC. The Intel Arc A380 uses the Xe-HPG architecture with the DG2-128 chip, produced on a much more advanced 6 nm process, also at TSMC. Despite the newer process, both chips contain the same transistor count of 7,200 million, but the die sizes differ sharply: the GTX 1070's die is 314 mm², while the Arc A380's is only 157 mm². This results in a transistor density of 22.9M per mm² for the Pascal chip versus 45.9M per mm² for the Intel part, showing the Arc A380 packs nearly twice as many transistors per area.

The memory subsystems are also quite different. The GTX 1070 uses 8 GB of GDDR5 on a 256-bit bus, delivering 256.3 GB/s of bandwidth with memory clocked at 2002 MHz (8 Gbps effective). The Arc A380 has 6 GB of GDDR6 on a much narrower 96-bit bus, yielding 186.0 GB/s of bandwidth, with memory at 1937 MHz (15.5 Gbps effective). Although the GDDR6 memory runs at nearly double the effective speed, the narrow bus limits the overall bandwidth to about 27% less than the GTX 1070.

Compute resources differ significantly. The GTX 1070 has 1920 shading units, 120 texture mapping units, and 64 ROPs. The Arc A380 is configured with 1024 shading units, 64 TMUs, and 32 ROPs. The GTX 1070's pixel rate is 107.7 GPixel/s and its texture rate is 202.0 GTexel/s, both far higher than the Arc A380's 65.60 GPixel/s and 131.2 GTexel/s. In raw FP32 throughput, the GTX 1070 reaches 6.463 TFLOPS compared to the Arc A380's 4.198 TFLOPS. However, the FP16 capabilities tell a different story: the GTX 1070 only manages 101.0 GFLOPS (at a 1:64 ratio), while the Arc A380 delivers 8.397 TFLOPS (at a 2:1 ratio). This means the Arc A380 has vastly superior half-precision compute, a feature that matters for certain emerging workloads.

The Arc A380 also includes 8 dedicated ray tracing cores, a feature entirely absent from the Pascal-based GTX 1070. This is a generational addition that enables hardware-accelerated ray tracing, which the older NVIDIA card cannot perform. The API support also differs: the GTX 1070 supports DirectX 12 (12_1), while the Arc A380 supports DirectX 12 Ultimate (12_2), the higher feature level. Both cards support OpenGL 4.6 and Vulkan 1.4.

Power and interface specifications are also notably different. The GTX 1070 has a TDP of 150 W with a suggested PSU of 450 W, while the Arc A380 has a TDP of just 75 W and a suggested PSU of 250 W. The GTX 1070 uses a PCIe 3.0 x16 interface, whereas the Arc A380 uses PCIe 4.0 x8. Display outputs differ as well: the GTX 1070 offers 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a, while the Arc A380 provides 1x HDMI 2.1 and 3x DisplayPort 2.0.

Where Each One Wins

The benchmark data paints a clear picture of where each GPU excels, and the use-case split is heavily tilted toward the NVIDIA card. The GTX 1070 is the dominant choice for nearly all traditional gaming and rendering workloads. Its wins span DirectX 9, 10, 11, and 12 tests, with margins ranging from 37.1% to 169.9%. This indicates broad strength across both legacy and modern API-based games. The PassMark G3D score of 13498 versus 6252 further reinforces that the GTX 1070 is the superior card for general 3D rendering, including most current game titles that rely on conventional rasterization.

The GTX 1070 also wins in compute-heavy tasks, as shown by the PassMark GPU Compute result (6102 versus 2762, a 120.9% lead) and the GeekBench OpenCL score (44700 versus 38224, a 16.9% lead). This makes it the better option for GPU-accelerated applications such as video encoding, scientific simulation, and machine learning inference that rely on FP32 or integer operations. The 3DMark Steel Nomad DX12 result (1082 versus 808, a 33.9% lead) confirms that even in modern, demanding graphics workloads, the GTX 1070 maintains a significant performance edge.

The single area where the Intel Arc A380 wins is the GeekBench Vulkan benchmark, where it scores 36736 versus 22121, a 39.8% advantage. This suggests that the Arc A380's Vulkan driver implementation is exceptionally well-optimized, and for applications that use the Vulkan API, the Arc A380 may deliver better performance than its other results would suggest. Additionally, the Arc A380 has hardware ray tracing cores and superior FP16 performance (8.397 TFLOPS versus 101.0 GFLOPS), making it a more future-proof choice for workloads that leverage these features, such as ray-traced lighting effects or FP16-accelerated AI inferencing. The Arc A380 also has a significantly lower TDP of 75 W versus 150 W, making it a far more power-efficient option for compact or low-power systems.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 1070 has an average benchmark score of 9780, while the Intel Arc A380 has an average of 8558, a difference of approximately 14% in favor of the GTX 1070.

Q: In which benchmark does the Intel Arc A380 outperform the GTX 1070?

A: The Arc A380 wins the GeekBench Vulkan test, scoring 36736 compared to the GTX 1070's 22121, which is a 39.8% higher result.

Q: How do the memory bandwidth figures compare between the two cards?

A: The GTX 1070 has a memory bandwidth of 256.3 GB/s using GDDR5 on a 256-bit bus, whereas the Arc A380 has 186.0 GB/s using GDDR6 on a 96-bit bus.

Q: Does the Intel Arc A380 support hardware ray tracing?

A: Yes, the Arc A380 includes 8 dedicated ray tracing cores, while the GTX 1070 has no ray tracing cores at all.

Q: What is the transistor density difference between the two chips?

A: The GTX 1070's GP104 chip has a transistor density of 22.9M per mm² on a 314 mm² die, while the Arc A380's DG2-128 chip has a density of 45.9M per mm² on a 157 mm² die, despite both having 7,200 million transistors.

Q: Which card has a higher FP32 compute throughput?

A: The GTX 1070 delivers 6.463 TFLOPS of FP32 performance, which is higher than the Arc A380's 4.198 TFLOPS.

Specification Differences

The two GPUs differ across nearly every major specification category, reflecting their distinct generations and design philosophies.

Process Node and Die: The GTX 1070 uses a 16 nm process with a 314 mm² die, while the Arc A380 uses a 6 nm process with a 157 mm² die. Both have 7,200 million transistors, but the transistor density is 22.9M per mm² for the GTX 1070 and 45.9M per mm² for the Arc A380.

Memory: The GTX 1070 has 8 GB of GDDR5 on a 256-bit bus with 256.3 GB/s bandwidth. The Arc A380 has 6 GB of GDDR6 on a 96-bit bus with 186.0 GB/s bandwidth. Memory clocks are 2002 MHz (8 Gbps effective) for the GTX 1070 and 1937 MHz (15.5 Gbps effective) for the Arc A380.

Compute Units: The GTX 1070 has 1920 shading units, 120 TMUs, and 64 ROPs. The Arc A380 has 1024 shading units, 64 TMUs, and 32 ROPs. The GTX 1070 also has higher pixel and texture rates: 107.7 GPixel/s and 202.0 GTexel/s versus 65.60 GPixel/s and 131.2 GTexel/s.

FP32 and FP16 Performance: The GTX 1070 achieves 6.463 TFLOPS FP32 and 101.0 GFLOPS FP16 (1:64 ratio). The Arc A380 achieves 4.198 TFLOPS FP32 and 8.397 TFLOPS FP16 (2:1 ratio).

Ray Tracing: The GTX 1070 has no ray tracing cores, while the Arc A380 includes 8 RT cores.

Power and Cooling: The GTX 1070 has a TDP of 150 W with a suggested PSU of 450 W. The Arc A380 has a TDP of 75 W with a suggested PSU of 250 W. Both are dual-slot cards with a single 8-pin power connector.

Interfaces and Outputs: The GTX 1070 uses PCIe 3.0 x16, while the Arc A380 uses PCIe 4.0 x8. Display outputs are 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a for the GTX 1070, versus 1x HDMI 2.1 and 3x DisplayPort 2.0 for the Arc A380.

API Support: The GTX 1070 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The Arc A380 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Physical Dimensions: The GTX 1070 measures 267 mm in length, 112 mm in height, and 40 mm in width. The Arc A380 measures 222 mm in length, 114 mm in height, and 42 mm in width.

Release and Status: The GTX 1070 was released on 2016-06-09 with a launch MSRP of 379 USD. The Arc A380 was released on 2022-06-13 with a launch MSRP of 149 USD. Both are end-of-life products, with the GTX 1070 succeeding the GeForce 900 series and preceding the GeForce 20, while the Arc A380 succeeds Xe Graphics and precedes Battlemage.