NVIDIA Quadro RTX 5000 Mobile Refresh

The NVIDIA Quadro RTX 5000 Mobile Refresh is a Turing-architecture mobile workstation GPU built on the TU104B chip, manufactured at TSMC on a 12 nm process. The database places it in the GeForce 50-series series field, while the product generation is listed as Quadro Turing-M (Tx000). It integrates 3072 shading units, 192 TMUs, 64 ROPs, 48 RT cores, and 384 tensor cores, with 16 GB of GDDR6 memory and a 110 W TDP. Production status is end-of-life, the benchmarks array is empty, and the GPU sits at the 50th percentile of all GPUs in the database.

Power and Cooling

The TDP is 110 W, which is the only power envelope listed in the FACT PACK. No suggested PSU value is provided, so there is no database-backed power-supply recommendation for this part. The power connectors field is "None," meaning the GPU does not use auxiliary PCIe power connectors from a desktop PSU. Slot width is IGP, indicating an integrated mobile form factor, and display outputs are described as "Portable Device Dependent."

The underlying silicon context is a 12 nm TSMC process with 13,600 million transistors on a 545 mm² die, giving a transistor density of 25.0M per mm². Those numbers frame the 110 W TDP: the device is a high-transistor-count mobile part, but the data does not include a cooler specification or thermal solution. Because the product is IGP slot width and uses no listed power connectors, the host laptop platform is responsible for both power delivery and cooling. There are no dimensions for length, height, or width in the data, and no PSU wattage figure is recorded.

Ray Tracing and Feature Set

The Turing architecture implementation includes 48 RT cores and 384 tensor cores. The RT core count appears in the data as a distinct hardware resource alongside shading units, and the tensor core count is listed separately. API support is broad: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The DirectX 12 Ultimate entry includes the 12_2 feature level, which is the highest DirectX 12 tier in the data.

The compute-oriented feature set is recorded through FP32 and FP16 throughput: FP32 is 9.400 TFLOPS, and FP16 is 18.80 TFLOPS at a 2:1 ratio. Pixel fill rate is 97.92 GPixel/s, and texture fill rate is 293.8 GTexel/s. These are the only performance-related feature numbers in the pack; there are no game-specific feature flags, no DLSS entries, and no additional ray-tracing performance figures. The RT and tensor core counts, together with the API list, define the feature set for this mobile workstation GPU.

How It Compares

The nearestRivals array in the FACT PACK is empty. That means there are no rival names, no rival scores, and no deltaPct values to analyze in this section. The only comparative placement field is percentileVsAllGpus, which is 50. That places the Quadro RTX 5000 Mobile Refresh at the median of the database's GPU population.

Because the nearestRivals data is empty, no per-rival paragraphs with percentage deltas can be written. The avgBenchmarkScore is 0, and the benchmarks array is empty, so there is no measured score to compare against the GPUs around it. The absence of rival entries is a data limitation rather than a performance verdict; the percentile still indicates that half of the GPUs in the database sit above or below this part. Without nearestRivals, all broader comparisons are restricted to the single percentile field.

FAQ

Q: What is the TDP and what power connectors does it require?

A: The TDP is 110 W. The data lists no suggested PSU, and the power connectors field is "None." Slot width is IGP.

Q: How much memory does it have and what is the bus width?

A: It has 16 GB of GDDR6 memory on a 256-bit bus, with 448.0 GB/s of bandwidth. The memory clock is 1750 MHz, listed as 14 Gbps effective.

Q: Does it support ray tracing and tensor operations?

A: Yes, the GPU includes 48 RT cores and 384 tensor cores. It is based on the Turing architecture.

Q: Which graphics APIs are supported?

A: The data lists DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: Is this GPU still in production?

A: No, production status is "End-of-life." The predecessor is listed as Quadro Pascal-M and the successor as Ampere-MW.

Q: What is the host interface?

A: The bus interface is PCIe 3.0 x16. Display outputs are described as "Portable Device Dependent."

Who Should Consider It

The data contains no measured benchmark scores, so this section cannot cite specific frame rates or settings tiers. Instead, the recommendation is grounded in the memory configuration, compute resources, and placement data. The 16 GB GDDR6 frame buffer and 448.0 GB/s of bandwidth point toward workloads that need large resident textures or datasets, which is a common requirement for high-resolution content. The 3072 shading units, 192 TMUs, and 64 ROPs provide the geometry and pixel-processing resources, with pixel fill rate of 97.92 GPixel/s and texture fill rate of 293.8 GTexel/s.

Because the part is at the 50th percentile among all GPUs, it is not positioned as an extreme high-end outlier in the database. It is end-of-life and uses an IGP slot width, so it is best suited to existing mobile workstations rather than new desktop builds. Users who need a mobile workstation GPU with a 16 GB memory buffer and Turing-generation RT/tensor cores are the intended audience. The empty benchmarks array means no database-backed statement can be made about which resolution or quality setting will be playable; the qualification rests on the memory subsystem and API support.

Memory Subsystem

The memory subsystem is 16 GB of GDDR6, connected over a 256-bit bus. Memory clock is 1750 MHz, which the FACT PACK also expresses as 14 Gbps effective. The resulting bandwidth is 448.0 GB/s. That bandwidth is the rate at which data can move between the memory and the 3072 shading units.

For high-resolution workloads, the 16 GB capacity is the most direct resource. Larger frame buffers reduce the pressure to swap textures or geometry out of VRAM, and 448.0 GB/s provides the throughput to feed the TMUs and ROPs. The texture rate is 293.8 GTexel/s and the pixel rate is 97.92 GPixel/s, both of which depend on the memory subsystem delivering data fast enough. This is the complete memory picture in the FACT PACK; no additional memory bandwidth or cache figures are listed.

Benchmark Performance

The benchmarks array is empty, and the avgBenchmarkScore is 0. No application-level test scores are recorded, and no deltaPct values exist because the nearestRivals array is also empty. The only database-level position metric is percentileVsAllGpus = 50, meaning the GPU sits at the midpoint of all GPUs in the database.

The performance-related numbers that do exist are the clock rates and throughput figures. The base clock is 1035 MHz and the boost clock is 1530 MHz. FP32 compute is 9.400 TFLOPS, while FP16 compute is 18.80 TFLOPS at a 2:1 ratio. Pixel fill rate is 97.92 GPixel/s and texture fill rate is 293.8 GTexel/s.

What the data shows is a specification profile rather than a measured benchmark profile. The 50th percentile placement is the only comparative score-like field, and it indicates a median position. Without benchmark scores, no exact percentage lead or deficit over any rival can be computed, and no nearest-rival deltas are available. The verdict from the data is clear: this GPU's position is median in the database, its memory configuration is sizable, and its measured application performance is simply not recorded.