NVIDIA GeForce RTX 5090 SE vs NVIDIA RTX 3500 Mobile Ada Generation Comparison

Where Each One Wins

The RTX 5090 SE and the RTX 3500 Mobile Ada Generation occupy entirely different segments of the GPU market, and the data reflects that split clearly. The RTX 5090 SE is a desktop-class part built around the GB202 chip, while the RTX 3500 Mobile is an integrated graphics processor (IGP) designed for portable workstations.

In raw compute throughput, the RTX 5090 SE dominates. Its FP32 rating of 66.94 TFLOPS is roughly 4.2 times the 15.82 TFLOPS of the RTX 3500 Mobile. That gap extends to every measured throughput category: pixel rate, texture rate, and memory bandwidth. The RTX 5090 SE delivers 380.3 GPixel/s against 98.88 GPixel/s for the mobile part, and its texture rate of 1,045.9 GTexel/s is more than four times the 247.2 GTexel/s of the RTX 3500 Mobile.

Memory capacity and bandwidth tell a similar story. The RTX 5090 SE carries 24 GB of GDDR7 across a 384-bit bus, yielding 1.34 TB/s of bandwidth. The RTX 3500 Mobile uses 12 GB of GDDR6 on a 192-bit bus for 432.0 GB/s. That is a 3.1x bandwidth advantage for the desktop part, which matters heavily for large datasets, high-resolution textures, and compute workloads that stream data.

The use-case split is straightforward. The RTX 5090 SE wins every measurable performance category. The RTX 3500 Mobile wins in power envelope and physical integration: its 100 W TDP versus 500 W, its IGP form factor versus dual-slot, and its lack of external power connectors versus a single 16-pin connector. The mobile part also uses PCIe 4.0 x16, while the desktop part uses PCIe 5.0 x16.

For portable workstations, the RTX 3500 Mobile is the only viable option of the two, given its IGP design. For fixed desktop workstations with adequate power delivery, the RTX 5090 SE is the performance leader without qualification.

Architecture Differences

The two GPUs come from different architectural generations. The RTX 5090 SE uses Blackwell 2.0 on the GB202 chip, while the RTX 3500 Mobile uses Ada Lovelace on the AD104 chip. Both are fabricated by TSMC on a 5 nm process, and their transistor densities are similar: 122.9 million transistors per square millimeter for the GB202 versus 121.8 million for the AD104.

The chip scale differs enormously. The GB202 die measures 750 mm² and contains 92,200 million transistors. The AD104 die is 294 mm² with 35,800 million transistors. That is a 2.6x transistor count advantage and a 2.6x die area advantage for the desktop part, which explains the large gap in shading units, TMUs, ROPs, RT cores, and tensor cores.

The RTX 5090 SE has 14,080 shading units, 440 TMUs, 160 ROPs, 110 RT cores, and 440 tensor cores. The RTX 3500 Mobile has 5,120 shading units, 160 TMUs, 64 ROPs, 40 RT cores, and 160 tensor cores. The desktop part thus has 2.75x the shading units, 2.75x the TMUs, 2.5x the ROPs, 2.75x the RT cores, and 2.75x the tensor cores.

Clock speeds also differ. The RTX 5090 SE runs at a 1740 MHz base and 2377 MHz boost. The RTX 3500 Mobile runs at 1110 MHz base and 1545 MHz boost. The desktop part boosts 53.9% higher, which compounds the massive shader-count advantage.

Memory architecture is another clear divider. The RTX 5090 SE uses GDDR7 at 1750 MHz (28 Gbps effective) with a 384-bit bus and 24 GB capacity. The RTX 3500 Mobile uses GDDR6 at 2250 MHz (18 Gbps effective) with a 192-bit bus and 12 GB capacity. The GDDR7 standard and wider bus give the desktop part its 1.34 TB/s bandwidth, versus 432.0 GB/s for the mobile part.

Power delivery differs fundamentally. The RTX 5090 SE has a 500 W TDP, requires a 1x 16-pin power connector, and needs a 900 W suggested PSU. The RTX 3500 Mobile has a 100 W TDP, uses no external power connectors, and lists no PSU requirement. The desktop card is dual-slot, 267 mm long, 111 mm tall, and 40 mm wide. The mobile part is an IGP with no listed dimensions.

Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Both are built by NVIDIA. The RTX 5090 SE belongs to the GeForce 50-series, while the RTX 3500 Mobile is listed as part of the GeForce 30-series generation in the database, despite using Ada Lovelace architecture. The mobile part's display outputs are described as portable device dependent, whereas the desktop card provides 1x HDMI 2.1b and 3x DisplayPort 2.1b.

Head-to-Head Benchmarks

The database records no direct benchmark entries for either GPU in this comparison, and the wins counters show zero for both. The nearest rival lists are empty, and the average benchmark score for each part is zero. This means the head-to-head comparison must rest entirely on the specification-level measurements recorded in the database.

The largest single-category win for the RTX 5090 SE is in FP32 compute. At 66.94 TFLOPS, it is 4.23 times the 15.82 TFLOPS of the RTX 3500 Mobile. That ratio is the defining performance gap between the two parts. FP16 compute is identical to FP32 on both GPUs (1:1 ratio), so the same 4.23x gap applies to half-precision workloads.

Pixel throughput shows a 3.85x advantage for the desktop part: 380.3 GPixel/s versus 98.88 GPixel/s. Texture throughput shows a 4.23x advantage: 1,045.9 GTexel/s versus 247.2 GTexel/s. These ratios are consistent with the shading unit and clock differences, confirming that the desktop part scales proportionally across all fixed-function units.

Memory bandwidth shows a 3.10x advantage: 1.34 TB/s versus 432.0 GB/s. This is the smallest of the major ratios, because the mobile part's GDDR6 runs at a higher memory clock (2250 MHz versus 1750 MHz) even though the effective data rate is lower (18 Gbps versus 28 Gbps). The wider 384-bit bus on the desktop part is what secures the bandwidth lead.

Clock speed differences are notable but secondary. The RTX 5090 SE boosts at 2377 MHz, which is 53.9% higher than the 1545 MHz boost of the RTX 3500 Mobile. The base clock gap is even larger: 1740 MHz versus 1110 MHz, a 56.8% difference. Higher clocks do not compensate for the 2.75x shader-count gap; they compound it.

Transistor counts reinforce the performance picture. The GB202 packs 92,200 million transistors versus 35,800 million in the AD104, a 2.58x difference. Die size scales similarly at 750 mm² versus 294 mm². The density figures are nearly identical (122.9M/mm² versus 121.8M/mm²), which means the performance gap comes from raw silicon area, not design efficiency.

The TDP ratio is the one category where the mobile part wins decisively. At 100 W versus 500 W, the RTX 3500 Mobile uses one-fifth the power of the RTX 5090 SE. That is a 5x efficiency difference in favor of the mobile part, though it comes with proportionally reduced performance.

The Verdict

The data shows two products with no overlapping use cases. The RTX 5090 SE is a desktop GPU with 500 W TDP, dual-slot cooling, a 16-pin power connector, and a 900 W suggested PSU. It requires a PCIe 5.0 x16 slot, a large chassis, and substantial power delivery. Its 24 GB of GDDR7 and 1.34 TB/s bandwidth serve high-end rendering, large model inference, and multi-display setups with three DisplayPort 2.1b outputs and one HDMI 2.1b output.

The RTX 3500 Mobile is an IGP with no external power connectors, no listed dimensions, and portable-device-dependent display outputs. Its 100 W TDP fits within mobile workstation power budgets. Its 12 GB of GDDR6 and 432.0 GB/s bandwidth are sufficient for professional workloads that fit within that memory capacity, but the 4.23x FP32 deficit and 3.10x bandwidth deficit place it in a lower performance tier.

For any performance-bound workload, the RTX 5090 SE is the clear choice, provided the system can supply 500 W and accommodate a dual-slot card. For any power-constrained or portable system, the RTX 3500 Mobile is the only choice, since the RTX 5090 SE cannot physically fit in such a chassis.

The release dates confirm the generational separation: the RTX 3500 Mobile launched in 2023, while the RTX 5090 SE is dated 2025. The RTX 5090 SE lists a launch MSRP of 1,499 USD, while the mobile part has no listed launch MSRP. The predecessor and successor relationships differ as well: the desktop card follows the GeForce 40 series and precedes the GeForce 60 series, while the mobile part follows Ampere-MW and precedes Blackwell-MW.

Neither GPU has recorded benchmark scores in the database, and both sit at the 50th percentile versus all GPUs. This means the comparison is entirely specification-driven. The performance ratios are consistent across every measured category, which gives confidence that the RTX 5090 SE is the superior part for compute and graphics workloads, while the RTX 3500 Mobile is the superior part for power efficiency and portability.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The RTX 5090 SE delivers 66.94 TFLOPS, which is 4.23 times the 15.82 TFLOPS of the RTX 3500 Mobile Ada Generation.

Q: What are the memory specifications of each GPU?

A: The RTX 5090 SE uses 24 GB of GDDR7 on a 384-bit bus with 1.34 TB/s bandwidth. The RTX 3500 Mobile uses 12 GB of GDDR6 on a 192-bit bus with 432.0 GB/s bandwidth.

Q: How does power consumption compare between the two?

A: The RTX 5090 SE has a 500 W TDP and requires a 1x 16-pin power connector with a 900 W suggested PSU. The RTX 3500 Mobile has a 100 W TDP, uses no external power connectors, and lists no PSU requirement.

Q: What architectures do these GPUs use?

A: The RTX 5090 SE uses Blackwell 2.0 on the GB202 chip, while the RTX 3500 Mobile uses Ada Lovelace on the AD104 chip. Both are fabricated by TSMC on a 5 nm process.

Q: Which GPU has more shading units and RT cores?

A: The RTX 5090 SE has 14,080 shading units and 110 RT cores. The RTX 3500 Mobile has 5,120 shading units and 40 RT cores. The desktop part has 2.75x more of both.

Q: Do both GPUs support the same graphics APIs?

A: Yes, both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Their display outputs differ: the RTX 5090 SE provides 1x HDMI 2.1b and 3x DisplayPort 2.1b, while the RTX 3500 Mobile's outputs are portable device dependent.