AMD Radeon AI PRO R9700 vs NVIDIA GeForce GTX 1630 Comparison

Head-to-Head Benchmarks

The benchmark data presents a starkly one-sided comparison, with the AMD Radeon AI PRO R9700 winning both head-to-head tests by decisive margins. In Geekbench OpenCL, the AMD card scores 118,202 against the GTX 1630’s 24,858, a delta of -79% from the AMD card’s perspective. This means the AMD card delivers approximately 3.75 times the raw compute throughput of the NVIDIA part in this workload. The Vulkan result tells a similar story, though with a slightly narrower gap: the R9700 posts 63,550 versus 23,695 for the GTX 1630, a -62.7% delta that still leaves the NVIDIA card trailing by a factor of roughly 2.7.

Looking at the broader benchmark averages, the two cards occupy surprisingly similar percentile positions despite their divergent performance profiles. The GTX 1630 sits at the 70th percentile among all GPUs with an average benchmark score of 24,277, while the R9700 ranks at the 69th percentile with an average of 23,315. This seeming contradiction—where the newer, far more powerful card scores lower in average—can be attributed to the composition of the benchmark suite. The R9700’s average includes several PassMark legacy tests (DirectX 9, 10, 11, and 12 scores of 363, 155, 272, and 83, respectively) alongside its 3DMark Steel Nomad score of 7,006, which drags its aggregate down. The GTX 1630’s average draws from only two strong Geekbench results, giving it a higher mean despite losing both head-to-head matchups.

The nearest-rival data adds context. The GTX 1630’s closest competitors by average score include the GeForce GTX 780 Ti (24,236, +0.2% delta), RTX 2080 SUPER (24,170, +0.4%), RX 6600 XT (24,442, -0.7%), and RX 6800S (24,063, +0.9%). The R9700’s neighbors are the Radeon R9 M290X (23,276, +0.2%), RX 6600M (23,273, +0.2%), Pro Vega 16 (23,250, +0.3%), and NVIDIA P106-100 (23,249, +0.3%). These deltas are all within one percentage point, indicating that the average benchmark score is not a reliable differentiator here—the two cards simply do not compete in the same performance class despite their percentile proximity.

Architecture Differences

The architectural gap between these two cards is generational and profound. The GTX 1630 uses the TU117 chip on NVIDIA’s Turing architecture, built on TSMC’s 12 nm process with 4,700 million transistors on a 200 mm² die, yielding a transistor density of 23.5 million per square millimeter. The R9700 employs the Navi 48 chip on AMD’s RDNA 4.0 architecture, fabricated on TSMC’s 4 nm node with 53,900 million transistors across a 357 mm² die—a density of 151.0 million per square millimeter. That is over six times the transistor density of the NVIDIA part, and more than eleven times the total transistor count.

The compute resources differ by an order of magnitude. The GTX 1630 packs 512 shading units, 32 texture mapping units, and 16 raster output pipelines. The R9700 fields 4,096 shading units, 256 TMUs, and 128 ROPs—eight times the shaders and TMUs, and eight times the ROPs. The R9700 also includes 64 dedicated ray tracing cores, a feature entirely absent from the GTX 1630. NVIDIA’s card has no RT cores and no tensor cores, reflecting its entry-level positioning within the GeForce 16 family.

Clock behavior also diverges sharply. The GTX 1630 runs a base clock of 1740 MHz and a boost of 1785 MHz—a narrow operating window. The R9700 has a base of 1660 MHz, a game clock of 2350 MHz, and an aggressive boost of 2920 MHz, representing a 76% boost over its base frequency. Memory clocks tell a similar story: the NVIDIA card uses 1500 MHz (12 Gbps effective) GDDR6, while the AMD card’s memory runs at 2518 MHz (20.1 Gbps effective).

The resulting throughput figures are staggering in their disparity. The GTX 1630 delivers 28.56 GPixel/s pixel fill rate, 57.12 GTexel/s texture rate, and 1.828 TFLOPS of FP32 compute. The R9700 achieves 373.8 GPixel/s, 747.5 GTexel/s, and 47.84 TFLOPS FP32—roughly 13 times the pixel throughput, 13 times the texture throughput, and 26 times the FP32 compute. The AMD card also supports FP16 at a 1:1 ratio (47.84 TFLOPS), whereas the GTX 1630’s FP16 throughput of 3.656 TFLOPS comes at a 2:1 penalty relative to FP32.

Where Each One Wins

The data indicates that the R9700 is the unequivocal performance leader in every benchmark where both cards were tested. Its Geekbench OpenCL win by 79% reflects its massive compute advantage, which would benefit workloads such as GPU-accelerated rendering, scientific simulation, and machine learning inference. The Vulkan win by 62.7% suggests strong modern gaming and compute API performance, supported by its DirectX 12 Ultimate (12_2) feature level and 64 RT cores.

The GTX 1630’s sole advantage lies in its physical footprint and power envelope. It is a single-slot card measuring 145 mm in length, 69 mm in height, and 18 mm in width, requiring no auxiliary power connectors and a 250 W suggested PSU. The R9700 is a dual-slot card at 267 mm by 111 mm by 40 mm, needing a 16-pin connector and a 700 W PSU. The NVIDIA card’s 75 W TDP versus the AMD card’s 300 W TDP means the GTX 1630 can be deployed in systems where power delivery and space are constrained, albeit with dramatically lower performance.

For legacy API compatibility, the GTX 1630 supports DirectX 12 (12_1), while the R9700 supports DirectX 12 Ultimate (12_2). Both cards support OpenGL 4.6 and Vulkan 1.4. The R9700’s PassMark results reveal a mixed legacy performance picture: it scores 363 in DirectX 9, 155 in DirectX 10, 272 in DirectX 11, and 83 in DirectX 12, alongside 1,236 in G2D and 26,979 in G3D, and 15,300 in GPU compute. These numbers suggest the R9700 is optimized for modern workloads rather than older API paths.

FAQ

Q: Which card has higher raw compute performance?

A: The AMD Radeon AI PRO R9700 dominates, with 47.84 TFLOPS FP32 versus the GTX 1630’s 1.828 TFLOPS—a 26-fold difference. In Geekbench OpenCL, the R9700 scores 118,202 against 24,858 for the GTX 1630, a -79% delta.

Q: How do the memory subsystems compare?

A: The R9700 offers 32 GB of GDDR6 on a 256-bit bus with 644.6 GB/s bandwidth. The GTX 1630 has 4 GB of GDDR6 on a 64-bit bus with 96.00 GB/s bandwidth. The AMD card provides eight times the memory capacity and roughly 6.7 times the bandwidth.

Q: Does the GTX 1630 have any redeeming advantages?

A: Yes, in physical size and power requirements. The GTX 1630 is a single-slot card at 145 mm length with no power connectors and a 75 W TDP, versus the R9700’s dual-slot 267 mm length, 16-pin connector, and 300 W TDP. The GTX 1630 also needs a 250 W PSU versus 700 W for the R9700.

Q: What are the production statuses of these cards?

A: The GTX 1630 is end-of-life, released on 2022-06-27, while the R9700 is active, released on 2025-07-22. The R9700’s launch MSRP is 1,299 USD.

Q: How do their benchmark averages compare despite the performance gap?

A: The GTX 1630 has an average benchmark score of 24,277 at the 70th percentile, while the R9700 averages 23,315 at the 69th percentile. This is because the R9700’s average includes legacy PassMark tests (DirectX 9-12 scores ranging from 83 to 363) that drag its aggregate down, despite its massive wins in Geekbench tests.

Q: Which card supports ray tracing?

A: Only the R9700, which includes 64 ray tracing cores and supports DirectX 12 Ultimate (12_2). The GTX 1630 has no RT cores and only supports DirectX 12 (12_1).

Specification Differences

| Specification | GTX 1630 | R9700 |

|---|---|---|

| Architecture | Turing | RDNA 4.0 |

| Process Node | 12 nm | 4 nm |

| Transistors | 4,700 million | 53,900 million |

| Die Size | 200 mm² | 357 mm² |

| Transistor Density | 23.5M / mm² | 151.0M / mm² |

| Base Clock | 1740 MHz | 1660 MHz |

| Boost Clock | 1785 MHz | 2920 MHz |

| Game Clock | — | 2350 MHz |

| Memory Clock | 1500 MHz (12 Gbps effective) | 2518 MHz (20.1 Gbps effective) |

| Memory Size | 4 GB | 32 GB |

| Memory Bus Width | 64 bit | 256 bit |

| Memory Bandwidth | 96.00 GB/s | 644.6 GB/s |

| Shading Units | 512 | 4096 |

| TMUs | 32 | 256 |

| ROPs | 16 | 128 |

| RT Cores | — | 64 |

| Pixel Rate | 28.56 GPixel/s | 373.8 GPixel/s |

| Texture Rate | 57.12 GTexel/s | 747.5 GTexel/s |

| FP32 Performance | 1.828 TFLOPS | 47.84 TFLOPS |

| FP16 Performance | 3.656 TFLOPS (2:1) | 47.84 TFLOPS (1:1) |

| TDP | 75 W | 300 W |

| Slot Width | Single-slot | Dual-slot |

| Power Connectors | None | 1x 16-pin |

| Suggested PSU | 250 W | 700 W |

| Bus Interface | PCIe 3.0 x16 | PCIe 5.0 x16 |

| Display Outputs | 1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a | 1x HDMI 2.1b, 3x DisplayPort 2.1a |

| DirectX Support | 12 (12_1) | 12 Ultimate (12_2) |

| Dimensions | 145 mm x 69 mm x 18 mm | 267 mm x 111 mm x 40 mm |

| Production Status | End-of-life | Active |

| Release Date | 2022-06-27 | 2025-07-22 |

| Predecessor | GeForce 10 | Radeon Pro Vega |

| Successor | GeForce 20 | — |