NVIDIA GeForce GTX 480 vs NVIDIA P106-090 Comparison

The NVIDIA P106-090 and the NVIDIA GeForce GTX 480 represent two vastly different eras of GPU design, and the benchmark data reflects a clear generational shift. The P106-090, a Pascal-based mining part, decisively outperforms the older Fermi-based GTX 480 in the available metrics, despite the latter’s status as a former flagship. The data shows a single head-to-head benchmark, and the results are not close.

Head-to-Head Benchmarks

The only directly comparable benchmark between the two cards is Geekbench OpenCL, and the NVIDIA P106-090 dominates this test. The P106-090 scores 21,304 points, while the GeForce GTX 480 manages just 13,300 points. This translates to a delta of 60.2 percent in favor of the P106-090, a massive margin that underscores the architectural leap from Fermi to Pascal. The P106-090’s score is not merely higher; it is over one and a half times the GTX 480’s output, indicating that the newer card handles compute workloads with significantly greater efficiency.

Looking at the broader benchmark context, the P106-090 also has results in 3DMark Steel Nomad (DX12) and Geekbench Vulkan, scoring 509 and 18,596 points respectively. The GTX 480 has no corresponding scores in these tests, which is telling in itself. The absence of a Vulkan result for the GTX 480 aligns with its spec sheet, which lists no Vulkan API support, while the P106-090 supports Vulkan 1.4. In the OpenCL test, the P106-090’s average benchmark score of 13,470 places it in the 54th percentile of all GPUs, slightly above the GTX 480’s 53rd percentile. This percentile difference is small, but the raw score gap in the head-to-head test is anything but.

The P106-090’s performance advantage is further contextualized by its nearest rivals. Its average benchmark score of 13,470 is within 0.5 percent of the AMD Radeon Pro 555 (13,407) and just 0.3 percent behind the NVIDIA GeForce GTX 570 (13,515). This places the P106-090 in a performance tier that is competitive with those older cards. The GTX 480, with an average score of 13,300, sits in a similar neighborhood, trailing the AMD Radeon RX 5500M (13,356) by 0.4 percent. Both cards are clustered around the same average performance level, yet the head-to-head OpenCL result shows the P106-090 pulling far ahead, suggesting its average is dragged down by other factors or that the OpenCL workload specifically favors its architecture.

The Verdict

Based strictly on the data, the NVIDIA P106-090 is the superior card for compute performance. It wins the only shared benchmark by a 60.2 percent margin, which is a decisive victory. The GTX 480, while historically significant, cannot match the P106-090’s raw throughput in OpenCL. The P106-090 also offers a more modern feature set, including support for DirectX 12_1 and Vulkan 1.4, whereas the GTX 480 is limited to DirectX 12 (11_0) and has no Vulkan support listed. For any workload that leverages these APIs, the P106-090 is the only viable option of the two.

The GTX 480 does have one advantage in the data: its launch MSRP was 499 USD, a figure that reflects its original flagship positioning. However, this price point is historical context, not a current recommendation. The P106-090’s launch MSRP is not listed, so no direct pricing comparison can be made. From a pure performance standpoint, the P106-090 is the clear winner. The GTX 480’s 1,345.0 GFLOPS of FP32 performance is dwarfed by the P106-090’s 2.352 TFLOPS, and its pixel rate of 21.03 GPixel/s is less than a third of the P106-090’s 73.49 GPixel/s. The verdict is unambiguous: the P106-090 is the card to choose for anyone prioritizing compute performance.

Where Each One Wins

The NVIDIA P106-090 is the winner in compute-heavy tasks, as evidenced by its 60.2 percent lead in the Geekbench OpenCL benchmark. This makes it the better choice for general-purpose GPU computing, including OpenCL-accelerated applications. Its support for Vulkan 1.4 also gives it an edge in modern graphics APIs, even though it lacks display outputs. The P106-090’s higher texture rate (73.49 GTexel/s vs 42.06 GTexel/s) and pixel rate (73.49 GPixel/s vs 21.03 GPixel/s) suggest it would also handle rasterization workloads more effectively, though no direct gaming benchmarks are available.

The NVIDIA GeForce GTX 480, despite losing the compute benchmark, retains a niche advantage in legacy compatibility. It has display outputs (2x DVI and 1x mini-HDMI 1.3a), making it functional for actual display use, whereas the P106-090 has no outputs at all. The GTX 480 also has a wider 384-bit memory bus compared to the P106-090’s 192-bit bus, though its memory bandwidth is slightly lower (177.4 GB/s vs 192.2 GB/s). For users with older software that relies on DirectX 11_0, the GTX 480’s API support is sufficient, but its lack of Vulkan support limits its future-proofing. The GTX 480’s 1536 MB of memory is half of the P106-090’s 3 GB, which could be a limiting factor in memory-intensive applications.

FAQ

Q: Which card performs better in Geekbench OpenCL?

A: The NVIDIA P106-090 scores 21,304 points, which is 60.2 percent higher than the NVIDIA GeForce GTX 480’s score of 13,300 points.

Q: Does the GTX 480 support Vulkan?

A: No, the GTX 480 has no Vulkan support listed in its specifications, while the P106-090 supports Vulkan 1.4.

Q: What is the memory size difference between the two cards?

A: The P106-090 has 3 GB of GDDR5 memory, while the GTX 480 has 1536 MB (1.5 GB) of GDDR5 memory.

Q: Which card has a higher average benchmark score?

A: The P106-090 has an average benchmark score of 13,470, which is slightly higher than the GTX 480’s average of 13,300.

Q: What is the TDP of each card?

A: The P106-090 has a TDP of 75 W, whereas the GTX 480 has a TDP of 250 W.

Q: Does either card have display outputs?

A: Only the GTX 480 has display outputs, featuring 2x DVI and 1x mini-HDMI 1.3a. The P106-090 has no display outputs.

Architecture Differences

The two cards are built on fundamentally different architectures and process nodes. The P106-090 uses the GP106 chip based on the Pascal architecture, manufactured on a 16 nm process at TSMC. It contains 4,400 million transistors on a 200 mm² die, resulting in a transistor density of 22.0M per mm². In contrast, the GTX 480 uses the GF100 chip based on the older Fermi architecture, built on a 40 nm process, also at TSMC. It packs 3,100 million transistors onto a much larger 529 mm² die, yielding a transistor density of just 5.9M per mm². This difference in density is a direct result of the process node advancement, allowing the P106-090 to fit more transistors in a smaller area.

The memory subsystems also differ significantly. The P106-090 has 3 GB of GDDR5 memory on a 192-bit bus, with a bandwidth of 192.2 GB/s and an effective memory speed of 8 Gbps. The GTX 480 has 1536 MB of GDDR5 memory on a wider 384-bit bus, but its bandwidth is lower at 177.4 GB/s, with an effective speed of 3.7 Gbps. The P106-090’s higher memory clock compensates for its narrower bus, resulting in superior overall bandwidth.

Compute resources are also distributed differently. The P106-090 has 768 shading units, 48 TMUs, and 48 ROPs, while the GTX 480 has 480 shading units, 60 TMUs, and 48 ROPs. Despite having fewer shading units, the GTX 480 has more TMUs. However, the P106-090’s higher clocks (1354 MHz base, 1531 MHz boost vs no base or boost clocks listed for the GTX 480) and architectural efficiency give it a massive lead in FP32 performance: 2.352 TFLOPS versus 1,345.0 GFLOPS. The P106-090 also supports FP16 at 36.74 GFLOPS, a feature absent from the GTX 480’s spec sheet. Power consumption reflects these differences, with the P106-090 drawing 75 W and the GTX 480 drawing 250 W. The P106-090 requires a single 6-pin power connector and a 250 W PSU, while the GTX 480 needs a 6-pin and an 8-pin connector and a 600 W PSU. The cards also differ in their bus interfaces, with the P106-090 using PCIe 1.0 x1 and the GTX 480 using PCIe 2.0 x16.