NVIDIA GeForce GTX 570 vs NVIDIA GeForce GTX 590 Comparison

The GeForce GTX 570 and GeForce GTX 590 represent two distinct strategies within NVIDIA's Fermi 2.0 generation, and their benchmark results reveal a counterintuitive outcome. Based on the Geekbench OpenCL score, the GTX 570 scores 13,515 points against the GTX 590's 12,830 points. This yields a 5.3% delta in favor of the GTX 570, meaning the single-GPU card outperforms the dual-GPU flagship in this specific compute test. The GTX 590, despite its higher tier positioning, posts a lower raw score, which raises immediate questions about scaling, power management, and driver optimization in this workload.

Head-to-Head Benchmarks

The only head-to-head benchmark available is Geekbench OpenCL, and the data is unambiguous. The GTX 570 delivers 13,515 points, while the GTX 590 trails at 12,830 points. The delta of 5.3% is not a marginal error; it is a consistent gap that places the GTX 570 ahead in compute performance. This is surprising because the GTX 590 carries more shading units (512 vs. 480), more texture mapping units (64 vs. 60), more ROPs (48 vs. 40), and a wider memory bus (384-bit vs. 320-bit). Yet the GTX 590's memory clock is lower at 854 MHz (3.4 Gbps effective) compared to the GTX 570's 950 MHz (3.8 Gbps effective). The GTX 590 also has higher memory bandwidth at 164.0 GB/s versus 152.0 GB/s, but this advantage does not translate into a win.

The GTX 570's pixel rate is 21.96 GPixel/s, which is higher than the GTX 590's 19.46 GPixel/s. Similarly, the texture rate favors the GTX 570 at 43.92 GTexel/s versus 38.91 GTexel/s. These fillrate advantages suggest that the GTX 570's clock domain is more efficient, despite the GTX 590's wider memory interface. The FP32 compute throughput also favors the GTX 570 at 1,405.4 GFLOPS versus 1,244.2 GFLOPS. The GTX 590's lower clocks across the board appear to cripple its theoretical peak, and the benchmark confirms that real-world OpenCL performance follows this pattern.

When contextualizing against rivals, the GTX 570's nearest competitors include the AMD Radeon HD 7770M (13,536 points, -0.2% delta), the AMD Radeon RX 9070 XT (13,543 points, -0.2% delta), and the NVIDIA P106-090 (13,470 points, +0.3% delta). The GTX 570 sits right in this cluster, meaning it is competitive with a modern mobile GPU and a much newer desktop part. The GTX 590, by contrast, sits near the AMD Radeon Pro 455 (12,831 points, 0% delta) and the AMD FirePro W5100 (12,847 points, -0.1% delta). The GTX 590's placement is lower, and its nearest rival delta of 0% against the Radeon Pro 455 suggests it is performing on par with a workstation-class mobile GPU from a later era.

Where Each One Wins

The GTX 570 wins the only benchmark test, so it takes the sole victory in this comparison. Its 13,515-point score is the definitive outcome. The GTX 590, with 12,830 points, has no benchmark win to claim. This is a stark result for a card that was positioned as the flagship GeForce 500-series product.

For use-case analysis, the GTX 570's higher FP32 throughput (1,405.4 GFLOPS) and higher pixel rate (21.96 GPixel/s) suggest it excels in compute-heavy workloads like OpenCL, which the benchmark directly measures. The GTX 590's strengths lie in its larger memory pool (1536 MB vs. 1280 MB) and wider 384-bit bus, which could theoretically benefit texture-heavy scenarios or higher resolutions, but the benchmark data does not support this. The GTX 590's higher ROP count (48 vs. 40) might help with certain rasterization tasks, but its lower pixel rate indicates the clock deficit negates that advantage.

The GTX 570 also benefits from a lower TDP of 219 W versus the GTX 590's 365 W, and a lower suggested PSU of 550 W versus 750 W. This efficiency gap is not directly measured in the benchmark, but it implies that the GTX 570 can sustain higher clocks without the thermal or power constraints that likely throttle the dual-GPU GTX 590. The GTX 590's 512 shading units suggest it should dominate, yet the data shows the opposite, pointing to a scalability failure where the second GPU does not contribute effectively in this OpenCL test.

The Verdict

Based strictly on the data, the NVIDIA GeForce GTX 570 is the superior performer in this head-to-head. It wins the Geekbench OpenCL test by 5.3%, and it does so while consuming less power (219 W vs. 365 W) and requiring a smaller PSU (550 W vs. 750 W). The GTX 570 also has a higher percentile rank at 54 versus the GTX 590's 52, meaning it outperforms a larger share of all GPUs in the database. For anyone selecting between these two for compute workloads, the GTX 570 is the clear choice.

The GTX 590, however, holds advantages in memory capacity (1536 MB vs. 1280 MB) and memory bandwidth (164.0 GB/s vs. 152.0 GB/s). These specs could make it more appealing for memory-bound tasks, but no benchmark data exists in the FACT PACK to validate this. The GTX 590 also offers triple DVI output plus a mini-DisplayPort, whereas the GTX 570 has dual DVI and a mini-HDMI 1.3a. If the user needs more display outputs, the GTX 590 has an edge, but this is not a performance metric.

The launch MSRP for the GTX 570 is 349 USD, and for the GTX 590 it is 699 USD. The GTX 590 costs double but delivers less compute performance, making the GTX 570 the more rational pick for OpenCL-based tasks. The GTX 590's only saving grace is its larger memory bus and capacity, which could matter for specific professional workloads, but the benchmark evidence does not support a recommendation for it.

FAQ

Q: Which GPU scores higher in the Geekbench OpenCL benchmark?

A: The NVIDIA GeForce GTX 570 scores 13,515 points, which is 5.3% higher than the GTX 590's 12,830 points.

Q: Does the GTX 590 have more shading units than the GTX 570?

A: Yes, the GTX 590 has 512 shading units, while the GTX 570 has 480 shading units, yet the GTX 570 still wins the benchmark.

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

A: The GTX 590 has a higher memory bandwidth of 164.0 GB/s, while the GTX 570 has 152.0 GB/s. The GTX 590 also has a wider 384-bit bus versus the GTX 570's 320-bit bus.

Q: How do the power requirements compare?

A: The GTX 570 has a TDP of 219 W and a suggested PSU of 550 W, while the GTX 590 has a TDP of 365 W and a suggested PSU of 750 W.

Q: What is the percentile rank of each GPU against all GPUs?

A: The GTX 570 ranks at the 54th percentile, while the GTX 590 ranks at the 52nd percentile.

Q: Which GPU has more memory?

A: The GTX 590 has 1536 MB of GDDR5 memory, whereas the GTX 570 has 1280 MB.

Architecture Differences

Both GPUs use the GF110 chip, which is built on the Fermi 2.0 architecture. They share the same 40 nm process node from TSMC, the same 3,000 million transistor count, and the same die size of 520 mm². The transistor density is identical at 5.8M / mm². Neither GPU features ray tracing cores or tensor cores, as these are not part of the Fermi 2.0 design. The core architectures are fundamentally the same, but the GTX 590 uses two GF110 dies, while the GTX 570 uses a single GF110 die. This dual-die configuration explains the GTX 590's higher shading unit count (512), TMU count (64), and ROP count (48), but it also introduces inter-GPU communication overhead that likely harms performance in the OpenCL benchmark.

The memory configurations differ significantly. The GTX 570 runs its GDDR5 memory at 950 MHz (3.8 Gbps effective) on a 320-bit bus, yielding 152.0 GB/s bandwidth. The GTX 590 runs its GDDR5 memory at 854 MHz (3.4 Gbps effective) on a 384-bit bus, yielding 164.0 GB/s. The GTX 590's higher bandwidth is offset by its lower clocks. The FP32 compute rates reflect this: the GTX 570 achieves 1,405.4 GFLOPS, while the GTX 590 achieves 1,244.2 GFLOPS. The GTX 590's pixel rate is 19.46 GPixel/s, and its texture rate is 38.91 GTexel/s, both lower than the GTX 570's 21.96 GPixel/s and 43.92 GTexel/s. This suggests that the GTX 590's dual-die design does not scale in these metrics due to lower clock speeds.

The process node, foundry, and transistor density are identical, meaning there is no manufacturing advantage for either card. The architecture is the same Fermi 2.0 with DirectX 12 (11_0) support, OpenGL 4.6, and no Vulkan support listed. Both cards are end-of-life products, with the GTX 570 releasing on 2010-12-06 and the GTX 590 on 2011-03-23. Both have the same predecessor (GeForce 400) and successor (GeForce 600) generations.

Specification Differences

The most obvious difference is the memory size: the GTX 570 has 1280 MB, while the GTX 590 has 1536 MB. The memory bus width also differs, with the GTX 570 using a 320-bit interface and the GTX 590 using a 384-bit interface. Memory bandwidth is 152.0 GB/s for the GTX 570 and 164.0 GB/s for the GTX 590.

The memory clock rates differ, with the GTX 570 at 950 MHz (3.8 Gbps effective) and the GTX 590 at 854 MHz (3.4 Gbps effective). The shading units are 480 for the GTX 570 and 512 for the GTX 590. TMUs are 60 vs. 64, and ROPs are 40 vs. 48. The pixel rate is 21.96 GPixel/s for the GTX 570 and 19.46 GPixel/s for the GTX 590. The texture rate is 43.92 GTexel/s for the GTX 570 and 38.91 GTexel/s for the GTX 590. FP32 performance is 1,405.4 GFLOPS for the GTX 570 and 1,244.2 GFLOPS for the GTX 590.

Power consumption is a major split: the GTX 570 has a TDP of 219 W, while the GTX 590 has a TDP of 365 W. The power connectors are 2x 6-pin for the GTX 570 and 2x 8-pin for the GTX 590. The suggested PSU is 550 W for the GTX 570 and 750 W for the GTX 590. The physical dimensions differ, with the GTX 570 at 267 mm (10.5 inches) in length, while the GTX 590 is 279 mm (11 inches) long, 111 mm (4.4 inches) high, and 40 mm (1.6 inches) wide. Both are dual-slot cards with a PCIe 2.0 x16 interface.

Display outputs differ: the GTX 570 has 2x DVI and 1x mini-HDMI 1.3a, while the GTX 590 has 3x DVI and 1x mini-DisplayPort. The launch MSRP is 349 USD for the GTX 570 and 699 USD for the GTX 590. The release dates are 2010-12-06 for the GTX 570 and 2011-03-23 for the GTX 590. Both cards have identical API support (DirectX 12 (11_0), OpenGL 4.6), and both lack Vulkan support. The benchmark scores are the final difference, with the GTX 570 at 13,515 and the GTX 590 at 12,830, placing them at the 54th and 52nd percentiles, respectively.