AMD Radeon 660M vs Intel Iris Xe MAX Graphics Comparison

The benchmark data places the Intel Iris Xe MAX Graphics and AMD Radeon 660M in adjacent performance tiers, but their results tell very different stories depending on the API and workload. In the only direct head-to-head test available, Intel’s DG1-based part decisively outpaces AMD’s Rembrandt IGP in OpenCL compute, while the AMD chip counters with a strong Vulkan showing that Intel cannot match—because no Vulkan result exists for the Intel part. The Intel card also holds a slight edge in overall percentile ranking, though the AMD part’s average score across two tests is pulled down by its weaker OpenCL number.

Head-to-Head Benchmarks

The sole direct comparison in the data is Geekbench OpenCL, where the Intel Iris Xe MAX Graphics scores 14,315 against the AMD Radeon 660M’s 12,876. That is an 11.2% victory for Intel, a substantial margin for integrated-class hardware. This result is consistent with Intel’s position in the broader database: its score sits within a razor-thin 0.3% of the NVIDIA GeForce GTX 1070 Ti (14,277) and just 0.3% behind the AMD Radeon Vega 11 (14,352). In other words, Intel’s OpenCL performance places it in genuine desktop-discrete territory, despite being an IGP with a 25 W TDP.

The AMD Radeon 660M, by contrast, trails in OpenCL but compensates in Vulkan. Its Geekbench Vulkan score is 14,748, which is 11.2% higher than its own OpenCL result and 3.0% above Intel’s OpenCL number. This Vulkan figure is the single highest benchmark score recorded for either product in this comparison. It indicates that AMD’s RDNA 2.0 architecture is significantly more efficient under modern graphics APIs, likely due to its hardware ray tracing cores and DirectX 12 Ultimate feature set. However, because Intel’s database entry lacks any Vulkan result, the head-to-head comparison remains incomplete—the two parts never face each other in the same API test.

Looking at the wider rival context, Intel’s OpenCL score places it among a cluster of much larger discrete GPUs. The AMD Radeon RX Vega 11 (14,385) is only 0.5% faster, and the NVIDIA GeForce GTX TITAN (14,373) is 0.4% faster. These are negligible margins—less than the typical run-to-run variance. Intel’s 56th percentile ranking among all GPUs underscores that this is not a low-end part; it outperforms roughly 44% of every GPU ever tested in the database.

The AMD Radeon 660M’s average benchmark score is 13,812, which reflects a split personality: 12,876 in OpenCL and 14,748 in Vulkan. This average places it within 0.1% of the NVIDIA RTX A2000 Mobile (13,821), a professional mobile GPU, and 0.5% ahead of the AMD Radeon RX 7900 XT (13,745) in this specific metric. The 660M’s 55th percentile ranking is just one point below Intel’s, indicating that despite losing the OpenCL face-off, the AMD part is globally competitive. The deltaPct values in its rival list are all under 1.5%, meaning the 660M sits in an extremely tight performance band where a single test can flip the ordering.

The Verdict

The data supports a clear split decision. For OpenCL compute workloads, the Intel Iris Xe MAX Graphics is the unambiguous winner, delivering 11.2% more performance than the AMD Radeon 660M. This is a large enough gap to matter in productivity tasks that leverage OpenCL, such as video encoding, image processing, or scientific simulations. Intel’s 56th percentile ranking, combined with its near-parity to the GTX 1070 Ti (deltaPct 0.3%), suggests it punches well above its IGP classification.

For Vulkan-based gaming and modern graphics APIs, the AMD Radeon 660M is the stronger choice, based on its 14,748 Vulkan score. This figure exceeds Intel’s best result by 3.0% and reflects AMD’s architectural advantages—RDNA 2.0 with 6 ray tracing cores and DirectX 12 Ultimate support. Gamers running Vulkan titles should expect smoother performance on the AMD part, despite its lower raw OpenCL throughput.

The choice ultimately depends on the target software stack. If the workload is OpenCL-centric, Intel wins decisively. If Vulkan is the primary API, AMD’s advantage is clear from the data, even though no direct Vulkan comparison exists. A user who needs both would face a trade-off: Intel for compute, AMD for graphics. The 660M’s average score of 13,812 versus Intel’s 14,315 gives Intel a 3.6% overall average lead, but that gap is almost entirely driven by AMD’s weak OpenCL result.

FAQ

Q: Which GPU has the higher Geekbench OpenCL score?

A: The Intel Iris Xe MAX Graphics scores 14,315 versus 12,876 for the AMD Radeon 660M, an 11.2% advantage for Intel.

Q: Does the AMD Radeon 660M outperform Intel in any benchmark?

A: Yes. The AMD Radeon 660M scores 14,748 in Geekbench Vulkan, which is 3.0% higher than Intel’s OpenCL result. Intel has no Vulkan score in the database.

Q: How does the Intel Iris Xe MAX compare to the NVIDIA GeForce GTX 1070 Ti?

A: Intel’s OpenCL score of 14,315 is 0.3% higher than the GTX 1070 Ti’s 14,277, putting them in a statistical tie.

Q: What is the AMD Radeon 660M’s average benchmark score?

A: The 660M averages 13,812 across its two tests (OpenCL 12,876 and Vulkan 14,748). Intel’s average is 14,315 from its single OpenCL test.

Q: Which GPU has a higher percentile ranking?

A: The Intel Iris Xe MAX Graphics ranks at the 56th percentile, while the AMD Radeon 660M ranks at the 55th percentile.

Q: Is the AMD Radeon 660M close to any professional mobile GPU?

A: Yes. Its average score of 13,812 is within 0.1% of the NVIDIA RTX A2000 Mobile’s 13,821.

Specification Differences

The two GPUs diverge sharply on core specifications. Intel’s DG1 chip packs 768 shading units, 48 texture mapping units, and 24 raster operation pipelines. AMD’s Rembrandt chip, by contrast, has exactly half the shading units (384), half the TMUs (24), and fewer ROPs (16). Despite this, AMD’s boost clock is higher: 1900 MHz versus Intel’s 1650 MHz. Intel’s base clock is much lower at 300 MHz, while AMD starts at 1500 MHz.

Memory configurations are fundamentally different. Intel uses dedicated 4 GB of LPDDR4X on a 128-bit bus, delivering 68.26 GB/s of bandwidth. AMD’s memory is System Shared, meaning it borrows from system RAM with System Dependent bandwidth and no dedicated pool. This gives Intel a major advantage in memory latency and consistency, but AMD’s higher clocks partially compensate.

The TDP gap is significant: Intel consumes 25 W while AMD draws 40 W. Both are integrated graphics (IGP) with no expansion slot. Intel lists a suggested PSU of 200 W, while AMD provides no such figure. Intel’s pixel rate is 39.60 GPixel/s versus AMD’s 30.40 GPixel/s; Intel’s texture rate is 79.20 GTexel/s versus AMD’s 45.60 GTexel/s. FP32 compute shows Intel at 2.534 TFLOPS and AMD at 1,459.2 GFLOPS (roughly 1.46 TFLOPS). FP16 numbers are 5.069 TFLOPS for Intel and 2.918 TFLOPS for AMD.

Architecture Differences

Intel’s Iris Xe MAX uses the DG1 chip built on Generation 12.1 architecture (Xe Graphics), fabricated on Intel’s 10 nm process with a 95 mm² die size. AMD’s Radeon 660M uses the Rembrandt chip on RDNA 2.0 architecture (Navi II IGP), built on TSMC’s 6 nm process with a 208 mm² die size and 13,100 million transistors. AMD’s process node is smaller and denser (63.0M transistors per mm²), but the die is more than twice the size.

The most consequential architectural difference is ray tracing. AMD includes 6 dedicated ray tracing cores; Intel has none. This aligns with AMD’s DirectX 12 Ultimate (12_2) support versus Intel’s DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4. AMD’s RDNA 2.0 is a modern graphics-first architecture, while Intel’s Generation 12.1 is a compute-oriented design—reflected in the benchmark split.

AMD’s display outputs are Portable Device Dependent, meaning they vary by laptop implementation. Intel lists No outputs, indicating it is purely a compute accelerator in some configurations. Both use PCIe 4.0 x8 interfaces. AMD’s predecessor is Vega II IGP and successor is Navi III IGP; Intel’s predecessor is simply Graphics and successor is Alchemist. Intel’s release date is 2020-10-30; AMD’s is 2022-01-03. Both are end-of-life products.

Where Each One Wins

Intel Iris Xe MAX Graphics wins in OpenCL compute-heavy tasks. Its 11.2% lead over the Radeon 660M in Geekbench OpenCL, combined with higher pixel rate (39.60 GPixel/s vs 30.40 GPixel/s) and texture rate (79.20 GTexel/s vs 45.60 GTexel/s), makes it the better pick for video transcoding, render farms, or any workload that leverages OpenCL. The dedicated 4 GB LPDDR4X memory with 68.26 GB/s bandwidth also gives it a decisive edge over AMD’s system-shared memory, which suffers from System Dependent bandwidth. Intel’s higher FP32 throughput (2.534 TFLOPS vs 1,459.2 GFLOPS) reinforces this compute advantage.

AMD Radeon 660M wins in Vulkan graphics and ray-traced workloads. Its Vulkan score of 14,748 is the highest single benchmark in this comparison, and its 6 ray tracing cores enable hardware-accelerated ray tracing that Intel cannot perform at all. The DirectX 12 Ultimate support means better feature parity with modern consoles and PC games. AMD’s higher boost clock (1900 MHz vs 1650 MHz) helps in latency-sensitive scenarios, and its 40 W TDP—while higher—suggests it can sustain higher clocks under load. For gamers prioritizing Vulkan titles or developers targeting DirectX 12 Ultimate features, the 660M is the data-backed choice.

For mixed workloads, Intel holds the average-score advantage. Intel’s 14,315 average beats AMD’s 13,812 by 3.6%, and Intel’s 56th percentile exceeds AMD’s 55th. But that average is based on a single test for Intel versus two for AMD. The fair interpretation is that Intel is more consistent in compute, while AMD is more specialized in graphics. Users who cannot predict their API mix should lean Intel for raw throughput; users who know they will use Vulkan should accept the OpenCL deficit for the graphics win.