NVIDIA GeForce GTX 590 vs NVIDIA P106-090 Comparison

FAQ

Q: Which GPU is faster in the Geekbench OpenCL benchmark?

A: The NVIDIA P106-090 wins decisively, scoring 21,304 versus the GTX 590’s 12,830, a 66% advantage in this head-to-head test.

Q: How does the P106-090 compare to its closest rivals in average benchmark score?

A: The P106-090’s average score of 13,470 places it within 0.3% of the GeForce GTX 570 (13,515) and 0.5% above the AMD Radeon Pro 555 (13,407). It sits at the 54th percentile of all GPUs.

Q: What is the GTX 590’s standing relative to its nearest competitors?

A: The GTX 590 averages 12,830, matching the AMD Radeon Pro 455 (12,831) within 0.0% and trailing the AMD Radeon 740M (12,870) by 0.3%. It ranks at the 52nd percentile of all GPUs.

Q: Do both cards support DirectX 12?

A: Yes, both list DirectX 12 support, but the P106-090 supports feature level 12_1, while the GTX 590 is limited to feature level 11_0.

Q: Which card has more shading units and what does that imply?

A: The P106-090 has 768 shading units versus the GTX 590’s 512. This contributes to its higher FP32 throughput of 2.352 TFLOPS compared to 1,244.2 GFLOPS.

Q: Are both cards currently in production?

A: No, both are marked as end-of-life. The P106-090 was released in July 2017, while the GTX 590 launched in March 2011.

Where Each One Wins

The data paints a clear picture: the P106-090 wins the only direct benchmark comparison, taking the Geekbench OpenCL test by 66%. However, the GTX 590 does not go down without arguments in specific usage scenarios.

The P106-090 is the stronger compute-oriented card. Its FP32 performance of 2.352 TFLOPS is nearly double the GTX 590’s 1,244.2 GFLOPS, making it the obvious choice for OpenCL workloads that scale with raw shading throughput. It also has double the memory capacity at 3 GB versus 1536 MB, which matters for larger datasets. The P106-090’s 16 nm process node and 4,400 million transistors in a 200 mm² die suggest significantly better efficiency per watt, though the numbers are not directly comparable.

The GTX 590, despite its age, holds advantages in specific architectural features. Its 64 texture mapping units (TMUs) outnumber the P106-090’s 48, giving it a theoretical texture rate advantage in certain fill-rate-bound scenarios, though the actual benchmark data does not test this. It also has a wider 384-bit memory bus versus 192-bit, which can help in memory-bandwidth-sensitive applications that do not exceed its 164.0 GB/s bandwidth. The GTX 590’s 3x DVI and 1x mini-DisplayPort outputs make it the only one of the two that can drive a display; the P106-090 has no display outputs at all.

For gaming, the GTX 590 supports DirectX 12 (11_0) but lacks Vulkan support entirely, while the P106-090 supports both DirectX 12 (12_1) and Vulkan 1.4. The P106-090 is the better choice for modern API workloads. The GTX 590’s 365 W TDP and requirement for a 750 W PSU (versus 75 W and 250 W for the P106-090) make it impractical in power-constrained builds, but its PCIe 2.0 x16 interface is far more standard than the P106-090’s PCIe 1.0 x1.

Architecture Differences

The two GPUs come from entirely different architectural eras. The P106-090 is built on the Pascal architecture using the GP106 chip, fabricated at 16 nm by TSMC. The GTX 590 uses the Fermi 2.0 architecture with the GF110 chip, fabricated at 40 nm, also by TSMC. This generational gap shows in transistor density: the P106-090 packs 4,400 million transistors into 200 mm² for a density of 22.0M per mm², while the GTX 590 crams 3,000 million transistors into a much larger 520 mm² die, yielding just 5.8M per mm².

The P106-090 belongs to the "Mining GPUs" generation, which explains its lack of display outputs and its unusual PCIe 1.0 x1 bus interface. The GTX 590 is part of the GeForce 500 series, with a standard PCIe 2.0 x16 interface and full display connectivity. The GTX 590 has a stated predecessor (GeForce 400) and successor (GeForce 600), while the P106-090 lists neither.

Shader architecture differs significantly. The P106-090 has 768 shading units and 48 ROPs, matching the GTX 590’s 48 ROPs but exceeding its 512 shading units. The GTX 590 counters with 64 TMUs versus 48. The P106-090’s FP16 throughput is 36.74 GFLOPS at a 1:64 ratio, while the GTX 590 does not list FP16 support. Both cards lack ray tracing and tensor cores.

The GTX 590 supports DirectX 12 (11_0) and OpenGL 4.6 but has no Vulkan entry. The P106-090 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. This makes the P106-090 the more future-proof option for API compatibility.

Specification Differences

The two cards differ across nearly every major specification. The P106-090 has a base clock of 1354 MHz and a boost clock of 1531 MHz, while the GTX 590 lists no base or boost clocks. Memory clocks also diverge: the P106-090 runs at 2002 MHz (8 Gbps effective), while the GTX 590 runs at 854 MHz (3.4 Gbps effective).

Memory configuration shows the P106-090 with 3 GB GDDR5 on a 192-bit bus delivering 192.2 GB/s bandwidth. The GTX 590 has 1536 MB GDDR5 on a 384-bit bus delivering 164.0 GB/s. The GTX 590’s wider bus does not compensate for its lower clock speed.

Compute rates favor the P106-090 heavily: 73.49 GPixel/s pixel rate and 73.49 GTexel/s texture rate versus the GTX 590’s 19.46 GPixel/s and 38.91 GTexel/s. FP32 performance is 2.352 TFLOPS versus 1,244.2 GFLOPS.

Power and physical specs could not be more different. The P106-090 draws 75 W with a single 6-pin connector and a suggested 250 W PSU. The GTX 590 draws 365 W with dual 8-pin connectors and a suggested 750 W PSU. Both are dual-slot cards, but the GTX 590 is longer at 279 mm (11 inches) versus 250 mm (9.8 inches) and has listed height (111 mm) and width (40 mm) dimensions, while the P106-090 only lists length.

The GTX 590 had a launch MSRP of 699 USD. The P106-090 has no launch MSRP on record.

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test, and it is not close. The P106-090 scores 21,304 against the GTX 590’s 12,830, a 66% delta in favor of the Pascal part. This is the largest margin in the entire comparison and reflects the P106-090’s superior compute architecture and higher clock speeds.

Contextualizing these scores against their respective rival pools adds nuance. The P106-090’s OpenCL result is roughly 58% above its own average benchmark score of 13,470, showing that it excels in OpenCL specifically. The GTX 590’s OpenCL score of 12,830 matches its average exactly, suggesting this is representative of its overall performance level.

In the broader field, the P106-090’s average score of 13,470 puts it within 0.3% of the GeForce GTX 570 (13,515) and 0.5% above the AMD Radeon Pro 555 (13,407). It trails the AMD Radeon RX 9070 XT (13,543) by 0.5%. The GTX 590’s average of 12,830 is essentially tied with the AMD Radeon Pro 455 (12,831) and within 0.4% of the GeForce GTX 670 (12,773).

The benchmark data does not include a Vulkan or 3DMark test for the GTX 590, but the P106-090 has additional results: 18,596 in Geekbench Vulkan and 509 in 3DMark Steel Nomad DX12. These indicate that the P106-090’s lead is not limited to OpenCL, though direct GTX 590 comparisons are unavailable.

The wins tally is 1-0 in favor of the P106-090. Given the 66% delta in the sole head-to-head test and the P106-090’s higher percentile ranking (54th versus 52nd), the data consistently points to the newer card as the stronger performer. The GTX 590’s advantages are limited to legacy features like display outputs and a standard PCIe interface, none of which appear in benchmark scoring.