NVIDIA P104-100 vs NVIDIA RTX A500 Mobile Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA RTX A500 Mobile has the higher average benchmark score at 39,568, compared to 32,982 for the NVIDIA P104-100. The A500 Mobile also places in the 82nd percentile of all GPUs, while the P104-100 sits in the 77th percentile.

Q: Why does the P104-100 win the head-to-head benchmarks if its average score is lower?

A: The head-to-head tests only include two specific workloads: Geekbench OpenCL and Geekbench Vulkan. The P104-100 wins both of those tests. Its average score is dragged down by the inclusion of a third benchmark, 3DMark Steel Nomad DX12, where it records 1,413 points. That test is not part of the shared head-to-head comparison.

Q: What memory configurations do these two cards use?

A: Both cards have 4 GB of memory, but they use different types and buses. The RTX A500 Mobile uses GDDR6 on a 64-bit bus with 96.00 GB/s of bandwidth. The P104-100 uses GDDR5X on a 256-bit bus with 320.3 GB/s of bandwidth.

Q: Which card has more shading units?

A: The RTX A500 Mobile has 2,048 shading units, while the P104-100 has 1,920. However, the P104-100 has more texture mapping units (120 versus 64) and more raster output units (64 versus 32).

Q: Do both cards support ray tracing?

A: No. The RTX A500 Mobile includes 16 ray tracing cores and 64 tensor cores. The P104-100 has no ray tracing cores and no tensor cores listed in the database.

Q: What are the production statuses of these GPUs?

A: Both are end-of-life products. The RTX A500 Mobile was released on 2022-03-21, while the P104-100 was released on 2017-12-11.

Where Each One Wins

The data splits cleanly along workload lines. The NVIDIA P104-100 wins every shared benchmark. In Geekbench OpenCL, it scores 52,368 against 41,263 for the RTX A500 Mobile, a delta of 21.2%. In Geekbench Vulkan, it scores 45,165 against 37,873, a delta of 16.1%. The P104-100 also has a recorded result in 3DMark Steel Nomad DX12 (1,413) that the A500 Mobile does not share, so that test cannot be compared directly.

However, the RTX A500 Mobile is not without its own territory. Its average benchmark score of 39,568 is substantially higher than the P104-100's 32,982, and it ranks in the 82nd percentile versus 77th. The A500 Mobile also brings features the P104-100 cannot match: 16 ray tracing cores, 64 tensor cores, and DirectX 12 Ultimate (12_2) support. The P104-100 only reaches DirectX 12 (12_1). For workloads that use ray tracing, tensor operations, or modern DirectX 12 Ultimate features, the A500 Mobile is the only option of the two.

The P104-100 is a mining-oriented card with no display outputs. The RTX A500 Mobile is an integrated graphics processor (IGP) with display outputs described as "Portable Device Dependent," meaning it is designed for laptops where the display connection is routed through the system. For compute benchmarks, the P104-100 has the edge. For a mobile workstation with display capability and modern API support, the A500 Mobile is the practical choice.

Architecture Differences

The RTX A500 Mobile is built on the Ampere architecture using the GA107S chip, fabricated on Samsung's 8 nm process. The P104-100 uses the older Pascal architecture with the GP104 chip, fabricated on TSMC's 16 nm process. This is a two-generation jump in process technology, and the transistor data reflects it: the A500 Mobile packs 8,700 million transistors into a 200 mm² die, for a density of 43.5 million transistors per mm². The P104-100 has 7,200 million transistors on a much larger 314 mm² die, for a density of 22.9 million per mm². The A500 Mobile achieves nearly double the transistor density.

The architectural differences go beyond the node. The A500 Mobile supports DirectX 12 Ultimate (12_2), which includes hardware ray tracing via its 16 RT cores and AI acceleration via its 64 tensor cores. The P104-100, from the Pascal generation, has neither RT cores nor tensor cores, and its DirectX support stops at 12_1. Both GPUs support OpenGL 4.6 and Vulkan 1.4, so those API levels are identical.

Memory architecture also diverges sharply. The A500 Mobile uses a 64-bit memory bus with GDDR6, while the P104-100 uses a 256-bit bus with GDDR5X. The P104-100's wider bus gives it 320.3 GB/s of bandwidth, more than three times the A500 Mobile's 96.00 GB/s. This is a critical difference for memory-bound workloads.

The compute capabilities reflect the generational split. The A500 Mobile delivers 6.296 TFLOPS of FP32 and the same 6.296 TFLOPS of FP16 (a 1:1 ratio), which is characteristic of Ampere's design. The P104-100 delivers 6.655 TFLOPS of FP32, slightly higher, but only 104.0 GFLOPS of FP16 (a 1:64 ratio). For FP16 workloads, the A500 Mobile is dramatically faster. The pixel rate also favors the P104-100: 110.9 GPixel/s versus 49.18 GPixel/s, and the texture rate favors it too: 208.0 GTexel/s versus 98.37 GTexel/s.

Specification Differences

The two GPUs differ in nearly every major specification category. The process node is 8 nm for the A500 Mobile versus 16 nm for the P104-100. The chip is GA107S for the A500 Mobile and GP104 for the P104-100. Transistor counts are 8,700 million versus 7,200 million, and die sizes are 200 mm² versus 314 mm².

Clock speeds favor the P104-100. Its base clock is 1,607 MHz and boost is 1,733 MHz, while the A500 Mobile runs at 832 MHz base and 1,537 MHz boost. Memory clocks are also different: the A500 Mobile runs at 1,500 MHz with 12 Gbps effective, while the P104-100 runs at 1,251 MHz with 10 Gbps effective.

Memory specifications are a major differentiator. Both have 4 GB, but the A500 Mobile uses GDDR6 on a 64-bit bus with 96.00 GB/s bandwidth. The P104-100 uses GDDR5X on a 256-bit bus with 320.3 GB/s bandwidth. The P104-100's bandwidth advantage is over 3.3 times.

Compute units differ as well. The A500 Mobile has 2,048 shading units, 64 TMUs, 32 ROPs, 16 RT cores, and 64 tensor cores. The P104-100 has 1,920 shading units, 120 TMUs, and 64 ROPs, with no RT or tensor cores. The P104-100 has higher pixel rate (110.9 GPixel/s versus 49.18 GPixel/s) and texture rate (208.0 GTexel/s versus 98.37 GTexel/s). The A500 Mobile has higher FP16 throughput (6.296 TFLOPS versus 104.0 GFLOPS), while the P104-100 has slightly higher FP32 throughput (6.655 TFLOPS versus 6.296 TFLOPS).

Power and physical specifications differ too. The A500 Mobile has a TDP of 30 W, is an IGP slot width, and has no power connectors. The P104-100 has no listed TDP, is dual-slot, requires a single 8-pin power connector, and has a suggested PSU of 200 W. The P104-100 is 267 mm (10.5 inches) long. The bus interface is PCIe 4.0 x8 for the A500 Mobile and PCIe 1.0 x4 for the P104-100. The A500 Mobile has display outputs that are "Portable Device Dependent," while the P104-100 has no display outputs at all.

Head-to-Head Benchmarks

The shared benchmark suite consists of two tests, and the P104-100 wins both. In Geekbench OpenCL, the P104-100 scores 52,368 against 41,263 for the A500 Mobile, a 21.2% advantage. In Geekbench Vulkan, the P104-100 scores 45,165 against 37,873, a 16.1% advantage. These are not marginal wins; the P104-100 is meaningfully faster in both compute APIs.

The A500 Mobile does not win any of the shared head-to-head benchmarks. Its advantage appears in the aggregate: its average benchmark score of 39,568 is 20.0% higher than the P104-100's 32,982. That average includes the P104-100's third benchmark, 3DMark Steel Nomad DX12, where it scores 1,413. That low result pulls the P104-100's average down significantly. The A500 Mobile's nearest rivals in the database are AMD Radeon Pro 575 (39,555, 0% delta), AMD Radeon Pro 575X (39,116, 1.2% delta), AMD Radeon Pro WX 7100 (40,063, -1.2% delta), and AMD Radeon Pro 580 (40,318, -1.9% delta). The P104-100's nearest rivals are NVIDIA T600 Mobile (32,849, 0.4% delta), NVIDIA T550 Mobile (33,161, -0.5% delta), NVIDIA GeForce RTX 3050 Mobile (33,170, -0.6% delta), and AMD Radeon Pro 570 (33,207, -0.7% delta). The A500 Mobile sits in a higher performance tier than the P104-100 based on average scores.

The FP16 comparison is stark. The A500 Mobile delivers 6.296 TFLOPS of FP16, identical to its FP32 figure. The P104-100 delivers only 104.0 GFLOPS of FP16, a 1:64 ratio. In FP16 compute, the A500 Mobile is over 60 times faster, which matters for AI and machine learning workloads that rely on reduced precision.

The Verdict

The choice between these two GPUs depends entirely on the workload. For raw compute throughput in OpenCL and Vulkan, the data favors the NVIDIA P104-100. It wins both head-to-head benchmarks by margins of 21.2% and 16.1%, and it offers over three times the memory bandwidth (320.3 GB/s versus 96.00 GB/s) thanks to its 256-bit bus. It also has higher FP32 throughput (6.655 TFLOPS versus 6.296 TFLOPS) and far higher pixel and texture rates. For compute-heavy tasks that use these APIs and do not require display output, the P104-100 is the stronger card.

For a mobile workstation context, the RTX A500 Mobile is the more complete product. It has display outputs, a 30 W TDP, and no power connectors, making it suitable for integration into a laptop. Its 82nd percentile ranking and average score of 39,568 place it well above the P104-100's 77th percentile and 32,982 average. It also supports DirectX 12 Ultimate (12_2), includes 16 ray tracing cores and 64 tensor cores, and delivers 6.296 TFLOPS of FP16 performance. The P104-100 has no display outputs, no ray tracing, no tensor cores, and minimal FP16 capability.

The P104-100 is a mining GPU, released 2017-12-11, with no display outputs and a dual-slot, 267 mm design that requires an 8-pin power connector. The A500 Mobile is a mobile workstation GPU, released 2022-03-21, built on a newer 8 nm process with nearly double the transistor density. Users who need a display, ray tracing, tensor cores, or FP16 compute should choose the RTX A500 Mobile. Users who need maximum OpenCL and Vulkan throughput and have no display requirements should choose the P104-100. The benchmark data is unambiguous: the P104-100 wins the shared tests, but the A500 Mobile wins on features, efficiency, and overall database standing.