NVIDIA GeForce RTX 5070 Mobile vs NVIDIA P104-100 Comparison

Head-to-Head Benchmarks

The NVIDIA GeForce RTX 5070 Mobile dominates the NVIDIA P104-100 in the recorded head-to-head benchmarks, winning both tests with substantial margins. In Geekbench OpenCL, the RTX 5070 Mobile scores 122,238 against the P104-100's 52,368, a delta of 57.2% in favor of the newer part. The Vulkan results are even more lopsided: the RTX 5070 Mobile posts 116,960 versus 45,165, a 61.4% advantage. These are not close contests; the RTX 5070 Mobile is in a different performance class entirely for compute workloads.

The P104-100's best showing comes in the 3DMark Steel Nomad DX12 test, where it records 1,413 points. However, the RTX 5070 Mobile has no corresponding score in the database for this test, so a direct comparison is unavailable. The RTX 5070 Mobile compensates with a much broader benchmark profile, including Passmark scores across DirectX 9 (214), DirectX 10 (129), DirectX 11 (192), DirectX 12 (93), G2D (896), G3D (20,355), and GPU compute (8,279). These results paint a picture of a versatile GPU capable of handling legacy APIs and modern compute tasks alike.

Looking at average benchmark scores, the P104-100 actually holds a slight edge: 32,982 versus 29,928 for the RTX 5070 Mobile. This is because the P104-100's average is built on fewer, higher-scoring tests, while the RTX 5070 Mobile's average is dragged down by its Passmark legacy API scores. The percentile rankings tell a similar story: the P104-100 sits at the 77th percentile of all GPUs, while the RTX 5070 Mobile sits at the 75th. This means the P104-100 outperforms a slightly larger fraction of the database's GPU population, despite losing both head-to-head matchups badly.

The nearest rivals for the P104-100 reinforce its mid-tier positioning. It sits within 0.7% of the NVIDIA T600 Mobile (32,849), T550 Mobile (33,161), RTX 3050 Mobile (33,170), and AMD Radeon Pro 570 (33,207). The RTX 5070 Mobile, by contrast, trades blows with desktop-class hardware: its nearest rivals include the RTX 3070 Ti (29,945, delta -0.1%), RTX 2080 Ti (29,783, delta 0.5%), RX 6800 (30,095, delta -0.6%), and RX 6700 (30,433, delta -1.7%). The delta percentages show the RTX 5070 Mobile is effectively on par with these high-end desktop cards, a remarkable feat for a mobile part.

FAQ

Q: Which GPU wins in Geekbench OpenCL performance?

A: The NVIDIA GeForce RTX 5070 Mobile wins decisively, scoring 122,238 against the P104-100's 52,368, a 57.2% advantage.

Q: How does the RTX 5070 Mobile compare to the P104-100 in Vulkan compute?

A: The RTX 5070 Mobile leads by 61.4%, posting 116,960 versus the P104-100's 45,165 in Geekbench Vulkan.

Q: Which GPU has a higher average benchmark score?

A: The P104-100 has a higher average at 32,982 versus 29,928 for the RTX 5070 Mobile, despite losing both head-to-head tests.

Q: What does the percentile ranking indicate?

A: The P104-100 ranks at the 77th percentile of all GPUs, while the RTX 5070 Mobile ranks at the 75th, meaning the P104-100 outranks a slightly larger share of the database.

Q: Are there any benchmarks where the P104-100 wins?

A: In the recorded head-to-head data, the P104-100 has zero wins. The RTX 5070 Mobile wins both available comparisons. However, the P104-100 does have a 3DMark Steel Nomad DX12 score of 1,413, for which the RTX 5070 Mobile has no recorded result.

Q: Which GPU has stronger nearest rivals?

A: The RTX 5070 Mobile's nearest rivals include the RTX 3070 Ti, RTX 2080 Ti, RX 6800, and RX 6700, all within 1.7% of its average score. The P104-100's nearest rivals are the T600 Mobile, T550 Mobile, RTX 3050 Mobile, and Radeon Pro 570, all within 0.7%.

The Verdict

The data is unambiguous for raw compute throughput: the NVIDIA GeForce RTX 5070 Mobile is the stronger GPU. It wins both head-to-head benchmarks by margins exceeding 57%, and its nearest rivals are desktop-class cards like the RTX 3070 Ti and RTX 2080 Ti. Anyone prioritizing OpenCL or Vulkan compute performance should choose the RTX 5070 Mobile without hesitation.

The P104-100, however, holds its own in the broader database context. Its average benchmark score of 32,982 exceeds the RTX 5070 Mobile's 29,928, and its 77th percentile ranking beats the RTX 5070 Mobile's 75th. This suggests the P104-100 benefits from a more consistent score distribution across the tests it participates in. For users who value a higher average score or percentile standing, the P104-100 is the better choice based on those metrics.

The RTX 5070 Mobile is the obvious pick for modern workloads requiring high FP32 throughput (13.13 TFLOPS versus 6.655 TFLOPS), ray tracing support, or tensor core acceleration. The P104-100, with its older Pascal architecture and mining-oriented design, offers no such features. However, the P104-100's 256-bit memory bus and 320.3 GB/s bandwidth exceed the RTX 5070 Mobile's 128-bit bus and 384.0 GB/s bandwidth in bus width, though not in raw bandwidth. For legacy API compatibility, the RTX 5070 Mobile's Passmark DirectX 9 score of 214 and DirectX 10 score of 129 indicate it handles older workloads adequately.

Specification Differences

The two GPUs differ across nearly every specification. The P104-100 uses a GP104 chip on a 16 nm TSMC process, while the RTX 5070 Mobile uses a GB206 chip on a 5 nm TSMC process. Transistor counts diverge sharply: 7,200 million for the P104-100 versus 21,900 million for the RTX 5070 Mobile, despite the RTX 5070 Mobile having a smaller die (181 mm² versus 314 mm²). Transistor density reflects this: 22.9M per mm² for the P104-100 versus 121.0M per mm² for the RTX 5070 Mobile.

Clock speeds tell a different story. The P104-100 runs at 1607 MHz base and 1733 MHz boost, while the RTX 5070 Mobile runs at 907 MHz base and 1425 MHz boost. The P104-100's higher clocks are offset by the RTX 5070 Mobile's superior architecture. Memory configurations also differ: the P104-100 has 4 GB of GDDR5X on a 256-bit bus, while the RTX 5070 Mobile has 8 GB of GDDR7 on a 128-bit bus. Bandwidth favors the RTX 5070 Mobile at 384.0 GB/s versus 320.3 GB/s.

Compute resources are heavily skewed toward the RTX 5070 Mobile. It has 4,608 shading units, 144 TMUs, 48 ROPs, 36 RT cores, and 144 tensor cores. The P104-100 has 1,920 shading units, 120 TMUs, and 64 ROPs, with no RT or tensor cores. Pixel rate favors the P104-100 (110.9 GPixel/s versus 68.40 GPixel/s), while texture rates are nearly identical (208.0 GTexel/s versus 205.2 GTexel/s). FP32 throughput massively favors the RTX 5070 Mobile (13.13 TFLOPS versus 6.655 TFLOPS), and FP16 is drastically different: 13.13 TFLOPS for the RTX 5070 Mobile versus just 104.0 GFLOPS for the P104-100.

Power and physical characteristics differ fundamentally. The P104-100 is a dual-slot card requiring a single 8-pin connector with a suggested PSU of 200 W, while the RTX 5070 Mobile is an IGP (integrated graphics processor) with a 50 W TDP and no power connectors. The P104-100 has no display outputs, while the RTX 5070 Mobile's outputs are portable device dependent. The bus interface diverges: PCIe 1.0 x4 for the P104-100 versus PCIe 5.0 x16 for the RTX 5070 Mobile. The P104-100 measures 267 mm in length; the RTX 5070 Mobile has no recorded dimensions.

Architecture Differences

The architectural gap is generational. The P104-100 is built on Pascal, released December 2017, and belongs to NVIDIA's Mining GPUs generation. The RTX 5070 Mobile is Blackwell 2.0, released April 2025, and belongs to the GeForce 50 Mobile series. The process node improvement from 16 nm to 5 nm enables the RTX 5070 Mobile to pack three times as many transistors into a smaller die.

The P104-100's Pascal architecture supports DirectX 12 (12_1), while the RTX 5070 Mobile supports DirectX 12 Ultimate (12_2). Both support OpenGL 4.6 and Vulkan 1.4, so API coverage is otherwise equal. The RTX 5070 Mobile adds dedicated ray tracing cores (36) and tensor cores (144), which the P104-100 lacks entirely. This is the single largest architectural differentiator: the RTX 5070 Mobile can accelerate ray-traced workloads and AI inferencing, while the P104-100 cannot.

Memory architecture also reflects the generational leap. The P104-100 uses GDDR5X at 10 Gbps effective, while the RTX 5070 Mobile uses GDDR7 at 24 Gbps effective. This explains how the RTX 5070 Mobile achieves higher bandwidth (384.0 GB/s) despite a narrower 128-bit bus. The P104-100's 256-bit bus is wider, but older memory technology limits its throughput.

Shading architecture favors the RTX 5070 Mobile in raw count: 4,608 shading units versus 1,920. The RTX 5070 Mobile also has more TMUs (144 versus 120), but fewer ROPs (48 versus 64). FP16 throughput is a major divergence: the RTX 5070 Mobile runs FP16 at a 1:1 ratio with FP32 (13.13 TFLOPS each), while the P104-100 runs FP16 at a 1:64 ratio (104.0 GFLOPS). This makes the RTX 5070 Mobile vastly more capable for mixed-precision compute.

The production status differs as well: the P104-100 is end-of-life, while the RTX 5070 Mobile is active. The RTX 5070 Mobile has a predecessor (GeForce 40 Mobile), while the P104-100 has no recorded predecessor or successor.

Where Each One Wins

The RTX 5070 Mobile wins in every direct benchmark comparison, making it the clear choice for compute-intensive applications. Its 57.2% lead in OpenCL and 61.4% lead in Vulkan mean it excels at general-purpose GPU computing, machine learning inference (thanks to tensor cores), and ray-traced rendering. The 1:1 FP16 ratio makes it ideal for workloads that leverage half-precision arithmetic, such as AI training or scientific simulations. Its 8 GB of GDDR7 memory and 384.0 GB/s bandwidth provide ample capacity and speed for modern datasets.

The P104-100 wins in specific niche scenarios. Its higher pixel rate (110.9 GPixel/s versus 68.40 GPixel/s) suggests better performance in fill-rate-bound scenarios, such as certain rasterization tasks. Its 256-bit memory bus, while paired with older GDDR5X, offers a wider memory path that can benefit workloads with high spatial locality. The P104-100's higher average benchmark score (32,982 versus 29,928) and higher percentile ranking (77th versus 75th) indicate it is more consistent relative to the broader GPU population.

For gaming, the RTX 5070 Mobile's Passmark G3D score of 20,355 and DirectX 11 score of 192 demonstrate strong DirectX performance, while its DirectX 12 score of 93 lags behind its own DirectX 11 result. The P104-100 has no Passmark scores in the database, so direct gaming comparisons are impossible. However, the RTX 5070 Mobile's robust compute scores and modern feature set make it the logical pick for contemporary games that leverage ray tracing, DLSS, or DirectX 12 Ultimate features.

The P104-100's lack of display outputs makes it unsuitable for any end-user graphics workload; it was designed for mining, and its data reflects that. The RTX 5070 Mobile, as an IGP with portable device dependent outputs, is designed for laptops and mobile workstations. The RTX 5070 Mobile's 50 W TDP and lack of power connectors make it far more power-efficient, while the P104-100's 200 W suggested PSU indicates a much higher system power draw.

In summary, the RTX 5070 Mobile is the superior GPU for virtually every modern use case, with overwhelming leads in compute benchmarks and a feature set that includes ray tracing, tensor cores, and modern memory technology. The P104-100's only advantages are its higher average score, higher percentile rank, and wider memory bus, which are outweighed by its age, lack of modern features, and mining-specific design.