Intel Arc A770M vs NVIDIA GeForce RTX 2060 Comparison

Where Each One Wins

The benchmark split between these two GPUs is not a simple generational sweep. The Intel Arc A770M takes 4 of the 10 head-to-head tests, while the NVIDIA GeForce RTX 2060 claims 6. But the wins are clustered by workload type, which tells you exactly which card to pick for which job.

The Arc A770M dominates modern, low-level API scenarios. Its biggest win comes in 3DMark Steel Nomad DX12, where it scores 2278 against the RTX 2060's 1242, a massive 83.4% advantage. That is not a small margin; that is a category difference. The Intel card also wins Geekbench OpenCL with 89494 versus 65014 (37.7% ahead) and Geekbench Vulkan with 74422 versus 65846 (13% ahead). These are compute and modern-API tests that reward raw shading throughput and architecture efficiency. The Arc also takes PassMark DirectX 12 with a 32.1% win (70 vs 53), reinforcing that this is a DirectX 12 Ultimate-era part.

The RTX 2060, by contrast, wins every legacy DirectX test in the suite. PassMark DirectX 10 goes 98 to 56 (42.9% ahead), DirectX 11 goes 110 to 69 (37.3% ahead), and even DirectX 9 edges ahead 190 to 178 (6.3% win). The NVIDIA card also wins the more general PassMark G3D test (14111 vs 11774, a 16.6% margin), PassMark G2D (745 vs 711, 4.6% ahead), and PassMark GPU Compute (5488 vs 4778, 12.9% ahead). So if your workload leans on older API paths or the broader PassMark composite, the RTX 2060 is the stronger pick.

The practical takeaway: the Arc A770M is a forward-looking card that excels where games and compute apps use modern APIs. The RTX 2060 is the safer bet for legacy titles and mixed workloads that still rely on older DirectX paths.

Architecture Differences

These two cards come from different design philosophies, and the data reflects that. The Arc A770M uses Intel's Xe-HPG architecture on the DG2-512 chip, built at TSMC on a 6 nm process. It packs 21,700 million transistors into a 406 mm² die, giving a transistor density of 53.4 million per mm². The RTX 2060 uses NVIDIA's Turing architecture on the TU106 chip, also from TSMC but on a 12 nm process. That chip holds 10,800 million transistors on a 445 mm² die, for a density of 24.3 million per mm². The Intel chip is denser by a wide margin, 53.4M versus 24.3M per mm², which is exactly what you expect from a newer process node.

The execution resources tell the same story. The Arc A770M has 4096 shading units, 256 texture mapping units, and 128 ROPs. The RTX 2060 has 1920 shading units, 120 TMUs, and 48 ROPs. Intel's part has more than double the shaders and TMUs, and nearly triple the ROPs. That hardware advantage shows up directly in the pixel and texture rates: the Arc hits 262.4 GPixel/s and 524.8 GTexel/s, while the RTX 2060 manages 80.64 GPixel/s and 201.6 GTexel/s.

Ray tracing hardware is closer than you might expect. The Arc has 32 ray tracing cores, the RTX 2060 has 30. NVIDIA brings 240 tensor cores, which the Intel part lacks entirely. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, so API coverage is identical on paper.

The process and transistor count differences are the root cause of the performance split. Intel used a smaller node to fit more transistors and more execution units into a similar die area. NVIDIA used a larger, older process but still managed a comparable die size. The result is a mobile GPU that targets modern workloads with brute shading power, versus a desktop GPU that leans on mature driver optimization and legacy API strength.

Head-to-Head Benchmarks

The single most important number in this comparison is the 3DMark Steel Nomad DX12 result. The Arc A770M scores 2278, the RTX 2060 scores 1242, and the delta is 83.4% in Intel's favor. That is the largest gap in the entire test suite. If you are building for a modern DirectX 12 game, this is the clearest signal that the Arc is the stronger card.

Geekbench OpenCL shows a similar but smaller advantage: 89494 versus 65014, a 37.7% win for Intel. This test stresses general compute through the OpenCL API, and the Arc's 16.79 TFLOPS FP32 throughput versus the RTX 2060's 6.451 TFLOPS is the hardware reason. The Vulkan test is closer, 74422 versus 65846, a 13% margin. Vulkan can be more driver-sensitive, and the data shows Intel's lead shrinks there compared to OpenCL.

The RTX 2060's wins are concentrated in PassMark's older API tests. DirectX 10 is its best result, 98 versus 56, a 42.9% lead. DirectX 11 follows at 110 versus 69, a 37.3% margin. These are large, consistent wins. The DirectX 9 test is much closer, 190 versus 178, only 6.3% apart, which suggests the Intel card is not completely lost on legacy paths, just clearly behind.

The composite PassMark G3D test favors NVIDIA, 14111 versus 11774, a 16.6% margin. That test mixes various workloads, and the RTX 2060's broader compatibility with older paths gives it the edge. PassMark GPU Compute also goes NVIDIA's way, 5488 versus 4778, a 12.9% win, which is interesting because the Arc wins OpenCL and Vulkan compute but loses the PassMark compute test. The PassMark G2D test is nearly even, 745 versus 711, a 4.6% NVIDIA lead.

The pattern is consistent: Intel wins modern APIs by large margins, NVIDIA wins legacy APIs by large margins, and the composite scores lean NVIDIA because they include more legacy tests.

Specification Differences

The memory subsystem is a major differentiator. The Arc A770M has 16 GB of GDDR6 on a 256-bit bus, delivering 512.0 GB/s of bandwidth. The RTX 2060 has 6 GB of GDDR6 on a 192-bit bus, delivering 336.0 GB/s. That is 10 GB more capacity and 176 GB/s more bandwidth for Intel. For texture-heavy or large-dataset workloads, the Arc has a clear hardware advantage. Memory clock differs too: Intel runs at 2000 MHz with 16 Gbps effective, NVIDIA at 1750 MHz with 14 Gbps effective.

Clock speeds show Intel pushing higher. The Arc base clock is 1650 MHz with a 2050 MHz boost. The RTX 2060 runs a 1365 MHz base and 1680 MHz boost. Those higher clocks, combined with more execution units, explain the FP32 throughput gap: 16.79 TFLOPS versus 6.451 TFLOPS. FP16 is similarly lopsided, 33.59 TFLOPS versus 12.90 TFLOPS, both at 2:1 ratio.

Power and physical design diverge sharply. The Arc A770M is rated at 120 W TDP and is an IGP (integrated graphics processor) form factor, meaning it is designed to be soldered onto a board. The RTX 2060 is a 160 W dual-slot card with a 1x 8-pin power connector and a suggested PSU of 450 W. It measures 229 mm long, 113 mm tall, and 35 mm wide. The Arc has no length, height, or width listed, and its display outputs are listed as "Portable Device Dependent," while the RTX 2060 has 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, and 1x USB Type-C. The bus interface differs too: PCIe 4.0 x16 for Intel, PCIe 3.0 x16 for NVIDIA.

The RTX 2060 has a release date of January 6, 2019, and a launch MSRP of 349 USD. The Arc A770M has no release date or launch MSRP in the database. Both are end-of-life products. The RTX 2060 lists a predecessor (GeForce 10) and successor (GeForce 30), while the Arc has neither.

FAQ

Q: Which card is faster in DirectX 12 Ultimate workloads?

A: The Intel Arc A770M wins the 3DMark Steel Nomad DX12 test by 83.4% (2278 vs 1242) and the PassMark DirectX 12 test by 32.1% (70 vs 53). Both cards support DirectX 12 Ultimate (12_2), but the Intel card executes it with significantly more performance.

Q: Does the RTX 2060 win any benchmark by a large margin?

A: Yes. The RTX 2060 wins PassMark DirectX 10 by 42.9% (98 vs 56) and PassMark DirectX 11 by 37.3% (110 vs 69). These are the largest NVIDIA wins in the head-to-head data.

Q: How much more memory bandwidth does the Arc A770M have?

A: The Arc A770M has 512.0 GB/s of bandwidth on a 256-bit bus, while the RTX 2060 has 336.0 GB/s on a 192-bit bus. That is a 176 GB/s difference in favor of Intel.

Q: Which card has more shading units?

A: The Arc A770M has 4096 shading units, 256 TMUs, and 128 ROPs. The RTX 2060 has 1920 shading units, 120 TMUs, and 48 ROPs. Intel's part has more than double the shaders and TMUs, and nearly triple the ROPs.

Q: Are these cards the same physical size?

A: No. The RTX 2060 is a dual-slot card measuring 229 mm long, 113 mm tall, and 35 mm wide, with a 1x 8-pin power connector. The Arc A770M is an IGP form factor with no listed dimensions, designed for portable devices.

Q: Which card wins more benchmarks overall?

A: The RTX 2060 wins 6 of the 10 head-to-head tests, while the Arc A770M wins 4. However, the Arc's wins are in modern API tests with much larger margins, particularly the 83.4% 3DMark Steel Nomad result.

The Verdict

The data points to a clear split based on workload, not a single winner. If your target is modern DirectX 12 or Vulkan gaming, the Intel Arc A770M is the stronger choice. Its 3DMark Steel Nomad score of 2278 versus 1242 is an 83.4% lead, and its Vulkan and OpenCL compute results are 13% and 37.7% ahead respectively. The 16 GB memory capacity and 512.0 GB/s bandwidth also make it the better option for large assets or compute-heavy tasks. The 120 W TDP and IGP form factor mean it fits into portable devices where space is limited.

If your library leans on older games or you need broad compatibility with DirectX 10 and 11 titles, the RTX 2060 is the safer pick. It wins PassMark DirectX 10 by 42.9% and DirectX 11 by 37.3%, and it takes the composite PassMark G3D test by 16.6%. The 160 W TDP, dual-slot design, and 450 W suggested PSU make it a standard desktop card, easier to slot into an existing PC. Its 6 GB of memory is smaller, but its legacy API performance is clearly better.

The average benchmark scores reflect this tradeoff. The Arc A770M sits at 18383 average and 62nd percentile among all GPUs, while the RTX 2060 sits at 15290 average and 58th percentile. The Arc is the higher-ranked card overall, but the RTX 2060's nearest rivals include the RTX 3050 OEM and RX 7600, while the Arc's nearest rivals include the Radeon RX 460 and RTX 3060 Mobile. The Intel card competes in a higher performance tier on average.

Choose the Arc A770M for modern API performance and future-proofing. Choose the RTX 2060 for legacy game compatibility and a conventional desktop form factor. The benchmark data does not support a single universal recommendation, it supports two distinct use cases.