AMD Radeon RX 6650M vs Intel Arc A770 Comparison

AMD Radeon RX 6650M and Intel Arc A770 are both end-of-life graphics cards, but they target very different segments: one is a mobile IGP for laptops, the other a dual-slot desktop add-in board. The benchmark data shows a clear overall winner in the Intel Arc A770, which takes both head-to-head tests, but the RX 6650M holds its own in the broader performance percentile rankings. The Arc A770 posts an average benchmark score of 68,809 against the RX 6650M’s 71,768, meaning the mobile AMD part actually has a higher aggregate score despite losing the direct comparisons. This creates an interesting dynamic where the desktop card wins the specific tests run, yet the laptop chip sits in the 91st percentile of all GPUs versus the Intel card’s 90th percentile.

Head-to-Head Benchmarks

The direct comparison consists of two Geekbench tests, and Intel wins both decisively. In Geekbench OpenCL, the Arc A770 scores 109,175 against the RX 6650M’s 65,800, a delta of -39.7% for the AMD part. That is a massive gap, with Intel leading by roughly two-thirds of its own score. The Vulkan test narrows the margin, but Intel still takes it: 94,284 for the Arc A770 versus 77,735 for the RX 6650M, a -17.6% delta. The Arc A770’s advantage in OpenCL is nearly double its Vulkan edge, suggesting the Intel architecture scales better in compute-heavy OpenCL workloads than in graphics-oriented Vulkan tasks.

The RX 6650M’s nearest rivals include the NVIDIA TITAN X Pascal at 72,098 (-0.5%), AMD Radeon Pro Vega 64 at 72,379 (-0.8%), and AMD Radeon Vega Frontier Edition at 73,370 (-2.2%). Its average of 71,768 places it just below the TITAN X Pascal, meaning the mobile chip trades blows with a flagship desktop GPU from a previous generation. The Arc A770’s rivals are similarly close: NVIDIA CMP 90HX at 69,000 (-0.3%), AMD Radeon Instinct MI25 at 68,562 (+0.4%), AMD Radeon Pro WX 8200 at 69,870 (-1.5%), and NVIDIA Quadro P6000 at 69,986 (-1.7%). The Intel card sits almost exactly between the CMP 90HX and Instinct MI25, showing it lands in a tight cluster of workstation and mining-oriented cards.

Notably, the RX 6650M’s average benchmark score of 71,768 beats every single one of the Arc A770’s listed rivals, including the CMP 90HX and Quadro P6000. This means that in the aggregate, the mobile AMD part outperforms the desktop Intel card’s closest competitors, even though it loses the head-to-head tests. The wins for Intel are real, but the overall picture is more nuanced than a simple 2-0 sweep.

Architecture Differences

The RX 6650M uses the Navi 23 chip built on RDNA 2.0 architecture, fabricated on a 7 nm TSMC process. It packs 11,060 million transistors onto a 237 mm² die, giving a transistor density of 46.7 million per mm². The Arc A770 uses the DG2-512 chip with Xe-HPG architecture, built on a 6 nm TSMC process. Intel’s die is far larger at 406 mm² and holds 21,700 million transistors, yielding a higher density of 53.4 million per mm². The Intel chip has nearly double the transistor count and a 71% larger die, but the RX 6650M achieves better average benchmark scores per transistor.

Memory configurations diverge sharply. The RX 6650M has 8 GB of GDDR6 on a 128-bit bus, delivering 224.0 GB/s of bandwidth. The Arc A770 doubles capacity to 16 GB GDDR6 on a 256-bit bus, quadrupling bandwidth to 512.0 GB/s. This memory advantage is likely a key factor in Intel’s OpenCL win, as compute workloads often scale with bandwidth. The RX 6650M’s memory runs at 14 Gbps effective, while the Arc A770’s runs at 16 Gbps effective.

Compute resources also favor Intel heavily. The Arc A770 has 4,096 shading units, 256 TMUs, and 128 ROPs, versus the RX 6650M’s 1,792 shading units, 112 TMUs, and 64 ROPs. Intel also leads in ray tracing cores, 32 to 28. The FP32 throughput tells the story: 19.66 TFLOPS for Intel versus 8.659 TFLOPS for AMD, a 2.27x advantage. Pixel rate is 307.2 GPixel/s for Intel versus 154.6 GPixel/s for AMD, and texture rate is 614.4 GTexel/s versus 270.6 GTexel/s. The RX 6650M’s smaller footprint shows in its 120 W TDP and IGP slot width with no power connectors, while the Arc A770 is a dual-slot card requiring 1x 6-pin plus 1x 8-pin power and a 550 W suggested PSU.

Both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, so API compatibility is identical. The RX 6650M connects via PCIe 4.0 x8, while the Arc A770 uses PCIe 4.0 x16. Display outputs differ: the AMD part is portable-device dependent, while Intel offers 1x HDMI 2.1 and 3x DisplayPort 2.0.

Where Each One Wins

The Arc A770 wins in raw compute and memory-bound scenarios. Its 512.0 GB/s bandwidth and 19.66 TFLOPS FP32 make it the clear choice for OpenCL-heavy workloads, as evidenced by the 39.7% lead in that benchmark. The 16 GB memory capacity is double the RX 6650M’s 8 GB, which matters for large datasets or high-resolution textures. The 225 W TDP and dual-slot design indicate it can sustain higher clocks under load, with a 2400 MHz boost clock versus 2416 MHz for the AMD part.

The RX 6650M wins in efficiency and portability. Its 120 W TDP is nearly half the Intel card’s 225 W, and it requires no power connectors, making it suitable for thin laptops. The 7 nm process with lower transistor density (46.7M/mm² versus 53.4M/mm²) suggests a more power-efficient design. Despite losing head-to-head tests, the RX 6650M’s average benchmark score of 71,768 is 4.3% higher than the Arc A770’s 68,809, indicating it performs better across a wider range of workloads not captured in the two direct tests. Its Vulkan score of 77,735 is only 17.6% behind Intel, a smaller gap than the OpenCL deficit.

The RX 6650M also holds a percentile advantage, sitting in the 91st percentile versus the Arc A770’s 90th. This means the AMD part ranks higher against all GPUs in the database, despite the Intel card’s direct wins. The Arc A770’s nearest rival, the NVIDIA CMP 90HX, is virtually tied at 69,000, while the RX 6650M’s closest competitor, the TITAN X Pascal, is also nearly tied at 72,098. Both cards land in competitive clusters, but the AMD part’s cluster sits slightly higher overall.

The Verdict

From the data alone, the Intel Arc A770 is the stronger choice for desktop users who prioritize raw compute performance. It wins both head-to-head tests, with a 39.7% lead in OpenCL and 17.6% in Vulkan. The 16 GB memory and 512.0 GB/s bandwidth give it a clear edge for memory-intensive applications. If you need maximum FP32 throughput or large frame buffers, the Arc A770 is the pick. Its launch MSRP was 329 USD.

The AMD Radeon RX 6650M is the better pick for mobile systems or users who care about overall benchmark standing. It achieves a higher average score (71,768) and better percentile (91st) despite losing both direct tests. The 120 W TDP and IGP form factor make it the only viable option for laptops. For a desktop builder, the RX 6650M is not a practical choice — it is an integrated mobile part with no power connectors or display outputs of its own.

Strictly from the data, the Arc A770 wins the head-to-head, but the RX 6650M wins the aggregate. The Intel card is for compute-heavy desktop workloads; the AMD card is for portable efficiency with competitive Vulkan performance. Neither card is a clear all-rounder, but the Arc A770’s dual benchmark victories make it the default recommendation for anyone who can accommodate a dual-slot, 225 W card with a 550 W PSU.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon RX 6650M scores 71,768 on average, which is higher than the Intel Arc A770’s 68,809.

Q: How much faster is the Arc A770 in Geekbench OpenCL?

A: The Arc A770 scores 109,175 versus the RX 6650M’s 65,800, a delta of -39.7% for the AMD part.

Q: Does the RX 6650M win any head-to-head benchmark?

A: No. The data shows the Arc A770 wins both tested benchmarks (OpenCL and Vulkan), giving Intel a 2-0 record in wins.

Q: What is the memory capacity difference?

A: The Arc A770 has 16 GB of GDDR6, while the RX 6650M has 8 GB. The Intel card also has a 256-bit bus versus 128-bit, yielding 512.0 GB/s versus 224.0 GB/s bandwidth.

Q: Which card has a higher transistor count?

A: The Arc A770 has 21,700 million transistors on its DG2-512 chip, versus 11,060 million on the RX 6650M’s Navi 23.

Q: Do both cards support the same APIs?

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

Specification Differences

The two cards differ in nearly every physical and compute specification. The RX 6650M uses a 7 nm process with 11,060 million transistors on a 237 mm² die, while the Arc A770 uses 6 nm with 21,700 million transistors on a 406 mm² die. Transistor density is 46.7M/mm² for AMD and 53.4M/mm² for Intel. Clock speeds are close: the RX 6650M has a base of 2068 MHz and boost of 2416 MHz, versus 2100 MHz base and 2400 MHz boost for Intel. Memory differs by capacity (8 GB versus 16 GB), bus width (128-bit versus 256-bit), and bandwidth (224.0 GB/s versus 512.0 GB/s). The RX 6650M has 1,792 shading units, 112 TMUs, and 64 ROPs, while the Arc A770 has 4,096 shading units, 256 TMUs, and 128 ROPs. Ray tracing cores are 28 versus 32. FP32 is 8.659 TFLOPS for AMD and 19.66 TFLOPS for Intel. Pixel rate is 154.6 GPixel/s versus 307.2 GPixel/s, and texture rate is 270.6 GTexel/s versus 614.4 GTexel/s. TDP is 120 W for the RX 6650M versus 225 W for the Arc A770. The AMD part is an IGP with no power connectors, while Intel is dual-slot with 1x 6-pin and 1x 8-pin and a 550 W suggested PSU. Bus interface is PCIe 4.0 x8 versus x16. Display outputs are portable-device dependent for AMD, versus 1x HDMI 2.1 and 3x DisplayPort 2.0 for Intel. Release dates differ: the RX 6650M launched on 2022-01-03, and the Arc A770 on 2022-10-11.