AMD Radeon RX 7650 GRE vs NVIDIA GeForce RTX 5090 D Comparison

Head-to-Head Benchmarks

The benchmark data leaves no ambiguity: the NVIDIA GeForce RTX 5090 D dominates the AMD Radeon RX 7650 GRE in every recorded test. In the 3DMark Steel Nomad DX12 test, the RTX 5090 D scores 14,326 points against 2,336 for the RX 7650 GRE, a delta of -83.7% for the AMD card. That is not a marginal gap; it is a difference of 12,000 points on a single workload, placing the two cards in entirely different performance tiers.

The Geekbench OpenCL result tells the same story. The RTX 5090 D posts 310,674 points, while the RX 7650 GRE manages 83,109, a -73.2% delta. The NVIDIA card is roughly 3.7 times faster in raw compute throughput according to this test. The RTX 5090 D also records a Geekbench Vulkan score of 376,915, a test the RX 7650 GRE has no recorded result for in the database.

The average benchmark score reinforces the hierarchy. The RTX 5090 D averages 77,712 across its recorded tests, placing it in the 92nd percentile of all GPUs in the database. The RX 7650 GRE averages 42,723, sitting in the 83rd percentile. Both cards outperform their nearest rivals, but the distance between them is substantial. The RTX 5090 D's nearest rival, the AMD Radeon RX 6650M XT, trails by only 1.1%, while the RX 7650 GRE's closest competitor, the NVIDIA GeForce RTX 4070 SUPER, sits just 1.2% ahead.

The RTX 5090 D also shows its strength across a broader set of legacy DirectX tests, with PassMark scores of 231 in DirectX 10, 371 in DirectX 11, 219 in DirectX 12, and 434 in DirectX 9, alongside a G3D score of 44,065 and a GPU compute score of 28,396. The RX 7650 GRE has no recorded PassMark results, so cross-comparison in those workloads is not possible from the data.

Architecture Differences

The two cards come from different manufacturing nodes and vastly different silicon. The AMD RX 7650 GRE uses the Navi 33 chip on TSMC's 6 nm process, with 13,300 million transistors on a 204 mm² die, yielding a transistor density of 65.2 million per mm². The NVIDIA RTX 5090 D uses the GB202 chip on a 5 nm TSMC process, packing 92,200 million transistors into a 750 mm² die, for a density of 122.9 million per mm². The NVIDIA die is nearly four times larger and holds nearly seven times more transistors.

The architectural generation gap is equally wide. AMD's card belongs to the RDNA 3.0 architecture, part of the Navi III (RX 7000) generation, with the internal codename Hotpink Bonefish. NVIDIA's card uses Blackwell 2.0 in the GeForce 50 series. The predecessor and successor chains differ too: AMD's card follows Navi II and precedes Navi IV, while NVIDIA's card follows GeForce 40 and precedes GeForce 60.

Execution resources show the scale difference. The RX 7650 GRE has 2,048 shading units, 128 texture mapping units, 64 raster output units, and 32 ray tracing cores. The RTX 5090 D has 21,760 shading units, 680 TMUs, 176 ROPs, 170 ray tracing cores, and 680 tensor cores. The NVIDIA card also carries tensor cores, which the AMD card does not list at all. Pixel rate for the NVIDIA card is 423.6 GPixel/s versus 172.5 GPixel/s for AMD. Texture rate is 1,636.8 GTexel/s versus 345.0 GTexel/s. FP32 and FP16 compute are both 104.8 TFLOPS on the NVIDIA card, against 22.08 TFLOPS on the AMD card.

Memory configurations diverge completely. The RX 7650 GRE uses 8 GB of GDDR6 on a 128-bit bus, with 288.0 GB/s of bandwidth and memory clocked at 2250 MHz (18 Gbps effective). The RTX 5090 D uses 32 GB of GDDR7 on a 512-bit bus, with 1.79 TB/s of bandwidth and memory clocked at 1750 MHz (28 Gbps effective). The NVIDIA card has four times the capacity and roughly six times the bandwidth.

The Verdict

The recorded data supports one conclusion: the RTX 5090 D is the stronger card in every measured dimension. It wins both head-to-head benchmarks, holds a higher average score, sits in a higher percentile, and comes from a newer architecture generation with substantially more execution resources. The RX 7650 GRE is not competitive in raw performance against this particular NVIDIA part.

The RX 7650 GRE's place in the database is nonetheless respectable. Its 83rd percentile ranking and average score of 42,723 put it within 1.2% of the RTX 4070 SUPER, a strong mid-range result. The RTX 5090 D, at the 92nd percentile with an average of 77,712, sits near the top of the performance distribution. The gap in average score is roughly 82%, which aligns with the head-to-head deltas.

The launch timeline shows the RTX 5090 D released on 2025-01-29, and the RX 7650 GRE followed on 2025-02-06. Both cards remain in active production. The NVIDIA card carries a launch MSRP of 2,299 USD, while the AMD card carries a launch MSRP of 279 USD. The price difference is enormous, but the performance difference is equally enormous, and the data shows no workload where the AMD card closes the gap.

Specification Differences

The two cards differ in nearly every specification field. The RX 7650 GRE uses a 6 nm process; the RTX 5090 D uses 5 nm. Transistor counts are 13,300 million versus 92,200 million. Die size is 204 mm² versus 750 mm². Transistor density is 65.2M per mm² versus 122.9M per mm². Base clocks are 1720 MHz versus 2017 MHz, and boost clocks are 2695 MHz versus 2407 MHz. The AMD card has a game clock of 2350 MHz; the NVIDIA card has no recorded game clock.

Memory differs in capacity, type, bus width, bandwidth, and clock. The AMD card uses 8 GB GDDR6 on a 128-bit bus with 288.0 GB/s bandwidth. The NVIDIA card uses 32 GB GDDR7 on a 512-bit bus with 1.79 TB/s bandwidth. Shading units are 2,048 versus 21,760. TMUs are 128 versus 680. ROPs are 64 versus 176. Ray tracing cores are 32 versus 170. Tensor cores are absent on the AMD card and number 680 on the NVIDIA card.

Power and connectivity also differ. TDP is 170 W for AMD and 575 W for NVIDIA. The AMD card uses a single 8-pin power connector and suggests a 450 W PSU; the NVIDIA card uses a single 16-pin connector and suggests a 950 W PSU. The AMD card runs on PCIe 4.0 x8; the NVIDIA card uses PCIe 5.0 x16. Display outputs are HDMI 2.1a plus three DisplayPort 2.1 on AMD, versus HDMI 2.1b plus three DisplayPort 2.1b on NVIDIA. Physical dimensions differ: the AMD card is 204 mm long and 115 mm tall, while the NVIDIA card is 304 mm long, 137 mm tall, and 48 mm wide. Both are dual-slot cards.

FAQ

Q: Which card has the higher average benchmark score?

A: The NVIDIA GeForce RTX 5090 D has an average benchmark score of 77,712, compared to 42,723 for the AMD Radeon RX 7650 GRE.

Q: How large is the performance gap in the 3DMark Steel Nomad DX12 test?

A: The RTX 5090 D scores 14,326 points versus 2,336 for the RX 7650 GRE, a delta of -83.7% for the AMD card.

Q: What memory configurations do the two cards use?

A: The RX 7650 GRE uses 8 GB of GDDR6 on a 128-bit bus with 288.0 GB/s bandwidth. The RTX 5090 D uses 32 GB of GDDR7 on a 512-bit bus with 1.79 TB/s bandwidth.

Q: Do both cards support the same DirectX version?

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

Q: Which card has more ray tracing cores?

A: The RTX 5090 D has 170 ray tracing cores. The RX 7650 GRE has 32.

Q: What is the transistor density of each card?

A: The RX 7650 GRE has a transistor density of 65.2M per mm². The RTX 5090 D has a density of 122.9M per mm².

Where Each One Wins

The RTX 5090 D wins every recorded benchmark. In 3DMark Steel Nomad DX12, it leads by 83.7%. In Geekbench OpenCL, it leads by 73.2%. It also has additional recorded results in Geekbench Vulkan and multiple PassMark tests, none of which the RX 7650 GRE has in the database. For any workload represented by these tests, the NVIDIA card is the clear choice.

The RX 7650 GRE has no benchmark wins in the recorded data. Its nearest rivals are the RTX 4070 SUPER and Quadro M6000 24 GB, both within 1.2%, which places it in a competitive mid-range bracket. The RTX 5090 D's nearest rivals are the RX 6650M XT, RX 6850M XT, and Tesla P100 variants, all within 2.4%, showing it leads its own tier by a smaller margin than its lead over the AMD card.

From the data, the RTX 5090 D suits workloads that demand maximum compute throughput, high memory bandwidth, and large memory capacity, given its 32 GB GDDR7 pool and 1.79 TB/s bandwidth. The RX 7650 GRE, with its 8 GB GDDR6 memory and 288.0 GB/s bandwidth, fits scenarios where the smaller power envelope of 170 W and single 8-pin connector are priorities. The recorded benchmarks, however, consistently favor the NVIDIA card across both DirectX and OpenCL workloads.