AMD Radeon RX 7900 XT vs NVIDIA P106-090 Comparison

Where Each One Wins

The data splits these two cards into cleanly separated use cases. The AMD Radeon RX 7900 XT dominates modern 3D rendering workloads, while the NVIDIA P106-090 shows a narrow but notable lead in a single compute-oriented OpenCL test.

In the 3DMark Steel Nomad DX12 benchmark, the RX 7900 XT scores 5,623 against the P106-090’s 509, a delta of 1,004.7%. That is not a marginal gap; it is an order-of-magnitude difference in raw DirectX 12 rendering throughput. The RX 7900 XT also wins the Geekbench Vulkan test decisively, scoring 71,068 versus 18,596, a 282.2% advantage. For anyone running Vulkan-based games or compute, the AMD card is the only viable option between the two.

The P106-090’s single win comes in Geekbench OpenCL, where it scores 21,304 versus the RX 7900 XT’s 12,756. That is a 40.1% lead for the NVIDIA card. This is an unusual result given the overall performance disparity, and it suggests the P106-090’s driver stack or compute scheduler is particularly well-tuned for the OpenCL workload in that specific test. However, this is one isolated data point against two decisive losses.

Looking at the average benchmark scores, the two cards are much closer than the head-to-head results imply. The RX 7900 XT averages 13,745 across all its benchmark entries, while the P106-090 averages 13,470. The RX 7900 XT’s nearest rival, the AMD Radeon 660M, sits at 13,812 (-0.5% delta), and the P106-090’s nearest rival, the NVIDIA GeForce GTX 570, sits at 13,515 (-0.3% delta). Both cards occupy the same percentile band against all GPUs — 55th for the RX 7900 XT, 54th for the P106-090. The average scores are skewed by the fact that the RX 7900 XT has many more benchmark entries, including low DirectX 9 and 10 results, which drag its mean down. The P106-090 only has three benchmark entries, all of which are relatively strong for its class.

The practical takeaway: the RX 7900 XT wins every modern gaming and graphics API test by a wide margin, while the P106-090 has a specific compute niche where it punches above its weight. If the workload is Vulkan or DX12, the choice is obvious. If the workload is OpenCL-heavy and does not require display output, the P106-090 deserves a second look.

Architecture Differences

The architectural gap between these two is generational. The RX 7900 XT uses the Navi 31 chip on the RDNA 3.0 architecture, built on a 5 nm TSMC process. It packs 57,700 million transistors into a 529 mm² die, yielding a transistor density of 109.1M per mm². The P106-090 uses the GP106 chip on the older Pascal architecture, built on a 16 nm TSMC process. It contains 4,400 million transistors on a 200 mm² die, for a density of 22.0M per mm². The RX 7900 XT has over 13 times the transistor count and a 5 times higher density, which explains the massive performance delta in compute-heavy tasks.

The core configurations differ even more starkly. The RX 7900 XT has 5,376 shading units, 336 texture mapping units, and 192 ROPs. It also includes 84 dedicated ray tracing cores, a feature entirely absent from the P106-090, which has no RT cores and no tensor cores. The P106-090 has 768 shading units, 48 TMUs, and 48 ROPs. This is not a refinement of the same design; it is a different class of hardware entirely.

Clock behavior also diverges. The RX 7900 XT has a base clock of 1,387 MHz, a boost clock of 2,394 MHz, and a game clock of 2,025 MHz. The P106-090 has a base clock of 1,354 MHz and a boost clock of 1,531 MHz, with no game clock specification. The RX 7900 XT’s boost clock is over 56% higher than the P106-090’s, and its memory runs at 2,500 MHz (20 Gbps effective) versus the P106-090’s 2,002 MHz (8 Gbps effective).

The memory subsystems are in different leagues. The RX 7900 XT has 20 GB of GDDR6 on a 320-bit bus, delivering 800.0 GB/s of bandwidth. The P106-090 has 3 GB of GDDR5 on a 192-bit bus, delivering 192.2 GB/s. That is a 4.16x bandwidth advantage for the AMD card, which matters enormously for texture-heavy scenes and large data sets.

Power and physical design reflect the performance gap. The RX 7900 XT has a 300 W TDP, requires two 8-pin power connectors, and suggests a 700 W PSU. The P106-090 has a 75 W TDP, a single 6-pin connector, and a 250 W PSU recommendation. The RX 7900 XT is also longer at 276 mm versus the P106-090’s 250 mm, and thicker at 51 mm versus an unspecified width for the NVIDIA card. Both are dual-slot.

The most consequential architectural difference for real-world use is display output. The RX 7900 XT has 1x HDMI 2.1a, 2x DisplayPort 2.1, and 1x USB Type-C. The P106-090 has no display outputs at all. This is a mining-specific card, and the data confirms it: no video output means it cannot function as a primary GPU in a standard desktop. The PCIe interface also differs, with the RX 7900 XT using PCIe 4.0 x16 and the P106-090 using PCIe 1.0 x1, a severely bottlenecked connection for data transfer.

Head-to-Head Benchmarks

The three shared benchmarks tell a consistent story with one outlier. The 3DMark Steel Nomad DX12 test is the clearest indicator of raw graphics power. The RX 7900 XT scores 5,623, and the P106-090 scores 509. The delta is 1,004.7%, meaning the AMD card performs at roughly 11 times the level of the NVIDIA card in this workload. That is not a win; it is a sweep. The RX 7900 XT’s 84 ray tracing cores and 800 GB/s memory bandwidth are the likely drivers, though the data does not break down per-feature contributions.

The Geekbench Vulkan test shows a similar but less extreme gap. The RX 7900 XT scores 71,068, while the P106-090 scores 18,596. The delta is 282.2%, meaning the AMD card is nearly 3.8 times faster. Vulkan is a low-level API that tends to scale with shading unit count and memory bandwidth, both of which heavily favor the RX 7900 XT. The P106-090’s 768 shading units and 192.2 GB/s bandwidth are simply insufficient to compete.

The Geekbench OpenCL result is the outlier. The P106-090 scores 21,304, and the RX 7900 XT scores 12,756. The delta is -40.1% for the AMD card, meaning the NVIDIA card is 40.1% faster. This is counterintuitive given the architectural disparity, but the data is unambiguous. OpenCL performance can be heavily influenced by driver optimization, and Pascal has a mature compute stack. Still, this is a single test, and the RX 7900 XT’s overall compute capability — 51.48 TFLOPS FP32 versus 2.352 TFLOPS FP32 — suggests the OpenCL result is workload-specific rather than representative of general compute performance.

The wins tally reflects this: the RX 7900 XT wins 2 out of 3 head-to-head tests, with the P106-090 taking the third. The magnitude of the RX 7900 XT’s wins (1,004.7% and 282.2%) dwarfs the P106-090’s single win (40.1%). For any weighted average, the AMD card is the superior performer.

Specification Differences

The key specification differences, where both cards have data, are as follows:

• Process node: 5 nm (RX 7900 XT) vs 16 nm (P106-090)
• Transistors: 57,700 million vs 4,400 million
• Die size: 529 mm² vs 200 mm²
• Base clock: 1,387 MHz vs 1,354 MHz
• Boost clock: 2,394 MHz vs 1,531 MHz
• Memory size: 20 GB vs 3 GB
• Memory type: GDDR6 vs GDDR5
• Memory bus: 320 bit vs 192 bit
• Memory bandwidth: 800.0 GB/s vs 192.2 GB/s
• Shading units: 5,376 vs 768
• TMUs: 336 vs 48
• ROPs: 192 vs 48
• RT cores: 84 vs none
• Pixel rate: 459.6 GPixel/s vs 73.49 GPixel/s
• Texture rate: 804.4 GTexel/s vs 73.49 GTexel/s
• FP32: 51.48 TFLOPS vs 2.352 TFLOPS
• FP16: 103.0 TFLOPS (2:1) vs 36.74 GFLOPS (1:64)
• TDP: 300 W vs 75 W
• Power connectors: 2x 8-pin vs 1x 6-pin
• Suggested PSU: 700 W vs 250 W
• Bus interface: PCIe 4.0 x16 vs PCIe 1.0 x1
• Display outputs: 1x HDMI 2.1a, 2x DisplayPort 2.1, 1x USB Type-C vs none
• DirectX version: 12 Ultimate (12_2) vs 12 (12_1)

The RX 7900 XT also has a game clock of 2,025 MHz, which the P106-090 does not list. The RX 7900 XT has dimensions of 276 mm length, 110 mm height, and 51 mm width, while the P106-090 only lists a length of 250 mm. The RX 7900 XT supports a 2:1 FP16 ratio, while the P106-090 uses a 1:64 ratio, making the AMD card vastly more capable for half-precision compute.

FAQ

Q: Which card is faster in DirectX 12?

A: The AMD Radeon RX 7900 XT is decisively faster. In the 3DMark Steel Nomad DX12 test, it scores 5,623 versus the NVIDIA P106-090’s 509, a delta of 1,004.7%.

Q: Does the NVIDIA P106-090 win any benchmark?

A: Yes. In the Geekbench OpenCL test, the P106-090 scores 21,304 against the RX 7900 XT’s 12,756, a 40.1% advantage for the NVIDIA card.

Q: Can the NVIDIA P106-090 be used as a display adapter?

A: No. The P106-090 has no display outputs, while the RX 7900 XT has one HDMI 2.1a port, two DisplayPort 2.1 ports, and one USB Type-C port.

Q: How do the memory bandwidths compare?

A: The RX 7900 XT has 800.0 GB/s of bandwidth from 20 GB of GDDR6 on a 320-bit bus. The P106-090 has 192.2 GB/s from 3 GB of GDDR5 on a 192-bit bus.

Q: What is the transistor density difference?

A: The RX 7900 XT has a density of 109.1M transistors per mm², while the P106-090 has 22.0M per mm². The RX 7900 XT packs 57,700 million transistors into 529 mm², versus 4,400 million into 200 mm².

Q: Which card has a higher average benchmark score?

A: The RX 7900 XT has an average benchmark score of 13,745, versus the P106-090’s 13,470. The RX 7900 XT sits at the 55th percentile against all GPUs, and the P106-090 at the 54th.

The Verdict

The data points to one conclusion for almost every use case: the AMD Radeon RX 7900 XT is the superior graphics card. It wins 2 out of 3 head-to-head benchmarks, with margins of 1,004.7% in DirectX 12 and 282.2% in Vulkan. It has 20 GB of GDDR6 memory versus 3 GB of GDDR5, 84 ray tracing cores versus none, and a boost clock of 2,394 MHz versus 1,531 MHz. Its FP32 throughput of 51.48 TFLOPS is nearly 22 times the P106-090’s 2.352 TFLOPS. For gaming, rendering, or any graphics-intensive workload, the RX 7900 XT is the only reasonable choice.

The NVIDIA P106-090 does have a niche. Its 40.1% OpenCL win over the RX 7900 XT is real and reproducible in the data. Its 75 W TDP and 250 W PSU recommendation make it far easier to power than the RX 7900 XT’s 300 W TDP and 700 W PSU requirement. It also has a 1x 6-pin connector versus the RX 7900 XT’s 2x 8-pin, which matters for space-constrained systems. However, the P106-090’s lack of display outputs means it cannot be a primary GPU, and its PCIe 1.0 x1 interface severely limits data transfer to the host system.

The verdict splits cleanly: if the workload involves any form of visual output, modern graphics APIs, or ray tracing, the RX 7900 XT is the pick. If the workload is a specific OpenCL compute task that runs headless, the P106-090’s single benchmark win and lower power draw make it a defensible choice — but only for that narrow scenario. The average benchmark scores are nearly identical (13,745 vs 13,470), but that parity is misleading. The RX 7900 XT’s average is dragged down by low DirectX 9 and 10 scores, while the P106-090’s average reflects only three strong entries. The head-to-head data, not the averages, should drive the decision.