AMD Radeon RX 7900 GRE vs NVIDIA P104-100 Comparison

The Verdict

The benchmark data tells an unambiguous story: the AMD Radeon RX 7900 GRE is the dominant performer across every shared test, with the NVIDIA P104-100 trailing by margins ranging from roughly 55% to over 70%. The RX 7900 GRE wins all three head-to-head benchmarks, and the deltas are not close—they are generational gaps. The P104-100, a mining-focused GPU from 2017, simply cannot compete with the modern RDNA 3.0 architecture in any measured workload.

For users seeking maximum compute performance or modern API support, the RX 7900 GRE is the only rational choice. It delivers 3.4x the OpenCL score, 2.2x the Vulkan score, and 3.4x the 3DMark Steel Nomad score of the P104-100. The P104-100, meanwhile, offers no display outputs, making it unsuitable for any conventional desktop use case. The data suggests the P104-100 is a niche artifact—useful only in specific mining or compute scenarios where its lower power draw (suggested PSU of 200 W versus 600 W) and older architecture might be acceptable trade-offs. However, even there, the performance gap is so vast that the RX 7900 GRE's advantages in raw throughput and memory capacity likely outweigh its higher power requirements.

Architecture Differences

The two GPUs come from entirely different eras and design philosophies. The NVIDIA P104-100 is built on the Pascal architecture, fabricated on a 16 nm TSMC process, with 7,200 million transistors packed into a 314 mm² die. The AMD Radeon RX 7900 GRE uses the RDNA 3.0 architecture on a 5 nm TSMC process, with 57,700 million transistors on a 529 mm² die. This represents a massive leap in transistor density: the RX 7900 GRE achieves 109.1M transistors per mm², nearly five times the P104-100's 22.9M per mm².

The compute resources differ dramatically. The P104-100 has 1,920 shading units, 120 texture mapping units, and 64 ROPs. The RX 7900 GRE more than doubles those numbers with 5,120 shading units, 320 TMUs, and 160 ROPs. The RX 7900 GRE also includes 80 ray tracing cores, a feature entirely absent from the P104-100. Memory configurations are equally divergent: the P104-100 uses 4 GB of GDDR5X on a 256-bit bus with 320.3 GB/s bandwidth, while the RX 7900 GRE uses 16 GB of GDDR6 on the same 256-bit bus but with 576.0 GB/s bandwidth—an 80% increase.

Clock speeds tell a nuanced story. The P104-100 has a higher base clock at 1607 MHz versus 1287 MHz, but the RX 7900 GRE boosts to 2245 MHz versus 1733 MHz. The RX 7900 GRE also has a game clock of 1880 MHz, a specification the P104-100 lacks entirely. The memory clocks show a similar pattern: the P104-100 runs at 1251 MHz (10 Gbps effective), while the RX 7900 GRE runs at 2250 MHz (18 Gbps effective). These architectural differences compound into the massive performance gaps seen in benchmarks.

Where Each One Wins

The RX 7900 GRE wins every single benchmark in the shared set, so the "where each wins" section is one-sided. However, the nature of the wins varies by workload type. In 3DMark Steel Nomad, a DirectX 12 test, the RX 7900 GRE scores 4814 versus 1413 for the P104-100—a 240% advantage. This suggests the RX 7900 GRE is dramatically better suited for modern gaming workloads, especially those leveraging DirectX 12 Ultimate features like ray tracing.

In Geekbench OpenCL, the RX 7900 GRE scores 175758 versus 52368, a 235% advantage. This indicates a massive lead in general-purpose compute tasks that rely on OpenCL acceleration. The Vulkan test shows the smallest relative gap but still a commanding win: 99850 versus 45165, a 121% advantage. This suggests the RX 7900 GRE is also superior in Vulkan-based applications, though the relative gap is narrower than in DirectX 12 or OpenCL.

The P104-100 has no wins, but its profile suggests it could theoretically be preferable in scenarios requiring minimal power consumption or physical space. Its suggested PSU is 200 W versus 600 W for the RX 7900 GRE, and it uses a single 8-pin connector versus two 8-pin connectors. However, with no display outputs, the P104-100 cannot function as a traditional graphics card for gaming or content creation. Its niche is purely computational, and even there, the RX 7900 GRE's benchmark supremacy makes it the superior choice unless power constraints are absolute.

FAQ

Q: Which GPU has better DirectX 12 performance?

A: The AMD Radeon RX 7900 GRE wins decisively in 3DMark Steel Nomad (DirectX 12), scoring 4814 versus 1413 for the NVIDIA P104-100, a 70.6% lead.

Q: How do the two compare in OpenCL compute workloads?

A: The RX 7900 GRE scores 175758 in Geekbench OpenCL, while the P104-100 scores 52368—a 70.2% gap in favor of AMD.

Q: Does the NVIDIA P104-100 support display output?

A: No. The P104-100 has no display outputs, making it unsuitable for any conventional monitor or VR headset connection.

Q: What is the memory capacity difference?

A: The RX 7900 GRE has 16 GB of GDDR6 memory, while the P104-100 has 4 GB of GDDR5X. Both use a 256-bit bus, but the RX 7900 GRE's bandwidth is 576.0 GB/s versus 320.3 GB/s.

Q: Which GPU has a higher boost clock?

A: The RX 7900 GRE boosts to 2245 MHz, compared to the P104-100's 1733 MHz. The P104-100 has a higher base clock (1607 MHz versus 1287 MHz), but the AMD card's boost advantage is substantial.

Q: Are there any benchmarks where the P104-100 comes close to the RX 7900 GRE?

A: No. The closest result is Geekbench Vulkan, where the RX 7900 GRE leads by 54.8%, still a dominant margin.

Head-to-Head Benchmarks

The head-to-head data reveals a consistent and overwhelming pattern of AMD superiority. Starting with 3DMark Steel Nomad (DirectX 12), the RX 7900 GRE posts a score of 4814, while the P104-100 manages only 1413. The delta of -70.6% means the P104-100 achieves less than a third of the AMD card's performance. This is a massive gap that reflects not just architectural age but also the fundamental differences in compute resources—the RX 7900 GRE has 5,120 shading units versus 1,920, and 80 RT cores versus none.

In Geekbench OpenCL, the gap is nearly identical in percentage terms: the RX 7900 GRE scores 175758 versus 52368, a 70.2% deficit for the P104-100. This test measures raw compute throughput, and the numbers align with the theoretical FP32 performance difference. The RX 7900 GRE delivers 45.98 TFLOPS of FP32 compute, while the P104-100 delivers 6.655 TFLOPS—a 6.9x difference that manifests in the benchmark scores, albeit not linearly.

The Vulkan test is the closest of the three, but still lopsided. The RX 7900 GRE scores 99850, while the P104-100 scores 45165, a 54.8% gap. This narrower margin might be attributable to Vulkan's lower-level API design, which can sometimes benefit older architectures by reducing driver overhead. Yet even here, the RX 7900 GRE more than doubles the P104-100's score.

The P104-100's best relative performance comes in Vulkan, which is notable given its lack of modern features. This could suggest that Vulkan workloads are less dependent on the architectural advantages that the RDNA 3.0 brings to DirectX 12, such as mesh shaders or advanced geometry processing. However, the absolute numbers still favor the RX 7900 GRE overwhelmingly.

It's also worth remembering the RX 7900 GRE has additional benchmark data that the P104-100 lacks entirely. PassMark scores for the AMD card include DirectX 10 (139), DirectX 11 (300), DirectX 12 (107), DirectX 9 (310), G2D (1180), G3D (27089), and GPU Compute (15016). These provide a more complete picture of its capabilities across legacy and modern APIs, while the P104-100 has no comparable data in the pack.

Specification Differences

The two GPUs differ across nearly every measurable specification. The manufacturing processes are generations apart: the P104-100 uses 16 nm TSMC technology, while the RX 7900 GRE uses 5 nm TSMC. Transistor counts reflect this leap: 7,200 million for NVIDIA versus 57,700 million for AMD, with die sizes of 314 mm² and 529 mm² respectively. Transistor density tells the real story—22.9M per mm² for the P104-100 versus 109.1M per mm² for the RX 7900 GRE.

Clock specifications show a mixed picture. The P104-100 has a base clock of 1607 MHz and a boost clock of 1733 MHz, while the RX 7900 GRE has a base clock of 1287 MHz, a game clock of 1880 MHz, and a boost clock of 2245 MHz. Memory clocks are 1251 MHz (10 Gbps effective) for the P104-100 and 2250 MHz (18 Gbps effective) for the RX 7900 GRE.

Memory capacity and type differ fundamentally: the P104-100 offers 4 GB of GDDR5X, while the RX 7900 GRE offers 16 GB of GDDR6. Both use a 256-bit bus, but bandwidth favors AMD at 576.0 GB/s versus 320.3 GB/s. The compute unit counts are starkly different: 1,920 shading units, 120 TMUs, and 64 ROPs for NVIDIA; 5,120 shading units, 320 TMUs, and 160 ROPs for AMD. The RX 7900 GRE also includes 80 RT cores, a feature the P104-100 lacks entirely.

Pixel and texture rates scale accordingly. The P104-100 delivers 110.9 GPixel/s and 208.0 GTexel/s, while the RX 7900 GRE delivers 359.2 GPixel/s and 718.4 GTexel/s. FP32 performance is 6.655 TFLOPS for NVIDIA versus 45.98 TFLOPS for AMD. FP16 performance shows a different ratio: the P104-100 manages 104.0 GFLOPS at a 1:64 ratio, while the RX 7900 GRE achieves 91.96 TFLOPS at a 2:1 ratio—a staggering difference in half-precision throughput.

Power requirements and physical specifications also diverge. The P104-100 has no TDP listed, with a suggested PSU of 200 W and a single 8-pin connector. The RX 7900 GRE has a TDP of 260 W, a suggested PSU of 600 W, and two 8-pin connectors. Both are dual-slot cards, but the P104-100 is 267 mm long, while the RX 7900 GRE is 276 mm long, with dimensions of 110 mm in height and 51 mm in width. The bus interface differs: the P104-100 uses PCIe 1.0 x4, while the RX 7900 GRE uses PCIe 4.0 x16—a massive bandwidth difference. Display outputs are another major divergence: the P104-100 has none, while the RX 7900 GRE offers 1x HDMI 2.1a, 2x DisplayPort 2.1, and 1x USB Type-C. Both support DirectX 12 (though the AMD card supports 12 Ultimate with version 12_2 versus 12_1 for NVIDIA), OpenGL 4.6, and Vulkan 1.4.