Intel Arc A370M vs NVIDIA GeForce GTX 1060 5 GB Comparison

Where Each One Wins

The benchmark data splits cleanly between these two GPUs. The NVIDIA GeForce GTX 1060 5 GB wins both recorded tests, taking the Geekbench OpenCL and Geekbench Vulkan workloads. The Intel Arc A370M does not win a single head-to-head comparison in the database. That gives the Pascal card a 2-0 sweep, but the margin of victory matters more than the raw count.

For compute-heavy OpenCL tasks, the GTX 1060 5 GB leads by 12.5%. That is a substantial gap, indicating the NVIDIA card handles general-purpose GPU compute with noticeably more headroom. The Vulkan test shows an even larger spread, with the GTX 1060 5 GB ahead by 18.5%. This suggests the older Pascal architecture still has an edge in low-level graphics API workloads, likely due to its wider memory bus and higher raw pixel throughput.

The Arc A370M is not without merit, however. Its strength lies in power efficiency and portability, not raw benchmark scores. With a 35 W TDP versus 120 W for the GTX 1060 5 GB, the Intel part delivers a meaningful fraction of the performance at less than a third of the power draw. For thin-and-light laptops where thermal headroom is scarce, that efficiency profile is the entire story. The data shows the A370M sits at the 74th percentile among all GPUs, while the GTX 1060 5 GB sits at the 78th. Both are above average, but the NVIDIA card occupies a higher tier.

In terms of raw score, the GTX 1060 5 GB averages 33,694 across both tests, while the Arc A370M averages 29,175. That is a 15.5% average deficit for the Intel part. Neither GPU is a flagship, but the Pascal card consistently outruns the Alchemist mobile chip in every measured scenario. The wins are not narrow either; they range from double-digit to nearly one-fifth ahead.

Architecture Differences

The two GPUs come from different eras and design philosophies. The GTX 1060 5 GB uses the GP106 chip built on TSMC's 16 nm process, packing 4,400 million transistors into a 200 mm² die. The Arc A370M uses the DG2-128 chip on TSMC's 6 nm node, with 7,200 million transistors in a smaller 157 mm² package. The newer process gives Intel a transistor density of 45.9M per mm², more than double the 22.0M per mm² of the Pascal chip.

Memory configurations diverge sharply. The NVIDIA card has 5 GB of GDDR5 on a 160-bit bus, delivering 160.2 GB/s of bandwidth. The Intel part has 4 GB of GDDR6 on a 64-bit bus, yielding 112.0 GB/s. The wider bus is a clear advantage for the GTX 1060 5 GB, contributing to its higher fill rates: 68.36 GPixel/s pixel rate and 136.7 GTexel/s texture rate. The A370M posts 65.60 GPixel/s and 131.2 GTexel/s, slightly lower but not far off.

Compute resources tell a similar story. The GTX 1060 5 GB has 1,280 shading units, 80 texture mapping units, and 40 ROPs. The Arc A370M has 1,024 shading units, 64 TMUs, and 32 ROPs. The NVIDIA card also leads in FP32 throughput at 4.375 TFLOPS versus 4.198 TFLOPS for Intel. However, the Arc A370M crushes the Pascal part in FP16: 8.397 TFLOPS versus a meager 68.36 GFLOPS. That 1:64 ratio on the GTX 1060 5 GB means it essentially lacks half-precision compute, while the A370M's 2:1 ratio makes it capable for machine learning inference workloads.

The Arc A370M includes 8 ray tracing cores, a feature the GTX 1060 5 GB lacks entirely. It also supports DirectX 12 Ultimate (12_2), while the NVIDIA card is limited to DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4. The Intel part uses PCIe 4.0 x8, while the GTX 1060 5 GB uses PCIe 3.0 x16. For bandwidth-hungry workloads, the newer interface helps, but the older card's x16 lane count compensates partially.

Power and physical design differ completely. The GTX 1060 5 GB is a dual-slot card requiring a 6-pin power connector and a 300 W PSU, stretching 250 mm in length. The Arc A370M is an integrated graphics processor (IGP) with no power connector and no defined dimensions, built for portable devices. The NVIDIA card is end-of-life, released in late 2017, while the Intel part is also end-of-life but came in early 2022.

Head-to-Head Benchmarks

The Geekbench OpenCL test shows the GTX 1060 5 GB scoring 33,397 against the Arc A370M's 29,676. That is a 12.5% delta in favor of NVIDIA. In practical terms, this means OpenCL compute workloads, such as physics simulations or video encoding filters, will complete noticeably faster on the Pascal card. The gap is not marginal; it is a clear tier difference.

The Geekbench Vulkan test widens the margin further. The GTX 1060 5 GB posts 33,991, while the Arc A370M manages 28,673. The 18.5% delta is the largest single-test difference in the comparison. Vulkan is a low-overhead API used in modern games and compute applications, and the NVIDIA card's advantage here suggests its driver maturity and hardware scheduler handle the API more efficiently. The Arc A370M's ray tracing cores do not help in this particular test, as Vulkan compute does not leverage RT hardware.

Looking at the nearest rivals for context, the GTX 1060 5 GB's average score of 33,694 places it just 0.2% above the NVIDIA RTX A5000 (33,622) and 0.5% below the AMD Radeon RX 7700S (33,849). It also trails the AMD Radeon HD 7950 by 0.8% (33,951) and the AMD Radeon RX 480 by 0.9% (33,997). These are all extremely close margins, indicating the GTX 1060 5 GB sits in a tightly packed performance cluster.

The Arc A370M's average of 29,175 is nearly identical to the AMD Radeon RX Vega M GH (29,197, a 0.1% deficit) and the AMD FirePro W8000 (29,211, also 0.1% behind). It leads the AMD Radeon RX 470 by 0.6% (28,996) and the AMD Radeon RX 6800M by 1% (28,874). The A370M's position among its rivals is similarly tight, but at a lower absolute performance level.

The pattern is consistent: the GTX 1060 5 GB outperforms the Arc A370M in both API tests, with the Vulkan gap nearly 50% larger than the OpenCL gap. The NVIDIA card's higher memory bandwidth and greater shading unit count likely drive these results. The Intel part's lower power envelope does not translate into a compute advantage in these benchmarks.

FAQ

Q: Which GPU has higher raw compute performance in FP32?

A: The NVIDIA GeForce GTX 1060 5 GB leads with 4.375 TFLOPS of FP32 throughput, while the Intel Arc A370M delivers 4.198 TFLOPS. The difference is small, about 4.2% in favor of NVIDIA.

Q: Does the Intel Arc A370M support ray tracing?

A: Yes, the Arc A370M includes 8 ray tracing cores. The NVIDIA GeForce GTX 1060 5 GB has no ray tracing cores, as it predates that hardware feature.

Q: How do the memory bandwidth figures compare?

A: The GTX 1060 5 GB offers 160.2 GB/s over a 160-bit GDDR5 bus. The Arc A370M provides 112.0 GB/s over a 64-bit GDDR6 bus. NVIDIA's card has a 43% bandwidth advantage.

Q: What is the power draw difference?

A: The GTX 1060 5 GB is rated at 120 W TDP and requires a 6-pin power connector. The Arc A370M is rated at 35 W TDP and is an integrated GPU with no power connector requirement.

Q: Which GPU has better Vulkan performance?

A: The NVIDIA GeForce GTX 1060 5 GB scores 33,991 in Geekbench Vulkan, which is 18.5% higher than the Arc A370M's 28,673. NVIDIA leads decisively in this API.

Q: Are these GPUs still in production?

A: No, both are end-of-life products. The GTX 1060 5 GB was released in late 2017, while the Arc A370M launched in early 2022.

The Verdict

The data points to a clear winner for raw performance. The NVIDIA GeForce GTX 1060 5 GB beats the Intel Arc A370M in every recorded benchmark, with margins ranging from 12.5% to 18.5%. Its average benchmark score of 33,694 sits at the 78th percentile among all GPUs, while the Arc A370M's 29,175 lands at the 74th percentile. If the priority is maximum compute throughput in OpenCL or Vulkan workloads, the GTX 1060 5 GB is the superior choice.

However, the Arc A370M is not without a role. Its 35 W TDP makes it suitable for portable devices where the 120 W GTX 1060 5 GB would be impractical. The Intel part also brings ray tracing support and DirectX 12 Ultimate, features absent from the older Pascal card. For FP16 workloads, the A370M's 8.397 TFLOPS dwarfs the GTX 1060 5 GB's 68.36 GFLOPS, making it the better option for half-precision compute tasks.

The choice depends entirely on the use case. For a desktop system with available power and PCIe slots, the GTX 1060 5 GB delivers more performance per benchmark point and a higher overall percentile ranking. For a thin laptop or a low-power embedded system, the Arc A370M offers a meaningful fraction of that performance at a fraction of the power draw, plus modern feature support like ray tracing.

The GTX 1060 5 GB also benefits from a wider memory bus and higher bandwidth, which likely explains its consistent lead in both API tests. The Arc A370M's narrower 64-bit bus is a bottleneck that its higher clock speeds cannot fully overcome. In the end, the database shows a 2-0 win for NVIDIA, but the Intel part wins on efficiency and feature set. Benchmark scores are not the only metric; the right GPU depends on the system it powers.