AMD Radeon RX 570X vs NVIDIA GeForce GTX 1080 Ti Comparison

The database pits two end-of-life cards against each other in NVIDIA's GeForce GTX 1080 Ti and AMD's Radeon RX 570X, and the recorded data shows a contest that is more lopsided than the raw silicon counts alone suggest. The Pascal flagship wins every shared benchmark in our measurements, sweeping the head-to-head three wins to zero, with margins ranging from a meaningful quarter in 2D throughput to a near order-of-magnitude blowout in PassMark's flagship 3D test. The RX 570X fights back only on efficiency and board requirements, drawing 150 W against the 1080 Ti's 250 W and needing a single 6-pin connector instead of a 6-pin plus 8-pin pair.

The Verdict

The data leaves little room for debate on performance. Across the three benchmarks both cards share, the GTX 1080 Ti takes all of them: Geekbench OpenCL by 74.4 percent, PassMark G2D by 25.2 percent, and PassMark G3D by a staggering 867.2 percent. Anyone choosing on frame rates or compute throughput should pick the 1080 Ti without hesitation.

The RX 570X earns consideration in narrower circumstances. Its 150 W TDP is well below the 1080 Ti's 250 W, and its 450 W suggested PSU undercuts the 250 W card's 600 W recommendation, which matters in tight builds or older systems with limited cabling. It is also physically shorter at 241 mm versus 267 mm, and it carries a DVI output alongside HDMI 2.0b and three DisplayPort 1.4a connectors, where the 1080 Ti drops DVI entirely for HDMI 2.0 and three DisplayPort 1.4a ports.

Context within the database reinforces the gap. The GTX 1080 Ti sits at the 58th percentile of all recorded GPUs with an average benchmark score of 15548, while the RX 570X lands at the 55th percentile with an average of 13871. Note that the percentile figures are closer than the head-to-head scores imply, because the averages draw from different mixes of recorded tests: the 1080 Ti's average spans eleven benchmark entries, while the RX 570X's covers only three.

FAQ

Q: Which card is faster overall?

A: The GTX 1080 Ti wins all three shared benchmarks, including an 867.2 percent victory in PassMark G3D (18600 versus 1923) and a 74.4 percent win in Geekbench OpenCL (67929 versus 38939).

Q: How do their database averages compare?

A: The GTX 1080 Ti averages 15548 across its recorded tests, against 13871 for the RX 570X, placing them at the 58th and 55th percentiles respectively among all GPUs in the database.

Q: Which card needs less power?

A: The RX 570X, clearly. Its 150 W TDP, single 6-pin connector, and 450 W suggested PSU compare favorably to the 1080 Ti's 250 W TDP, 6-pin plus 8-pin connectors, and 600 W PSU recommendation.

Q: How does the GTX 1080 Ti stack up against its nearest rivals?

A: Its average score sits 1.7 percent above both the GeForce RTX 2060 and the GeForce GTX 580, 0.8 percent below the Radeon Pro W5500, and 0.2 percent above the Radeon R9 M380, making it a tightly clustered matchup in aggregate terms.

Q: Which card has the larger memory buffer and bandwidth?

A: The 1080 Ti, with 11 GB of GDDR5X on a 352-bit bus delivering 484.4 GB/s. The RX 570X has 8 GB of GDDR5 on a 256-bit bus at 224.0 GB/s.

Q: Do either of these cards support ray tracing hardware?

A: No. The database records no RT cores and no tensor cores for either card.

Architecture Differences

These cards come from opposing design philosophies of the same era. The GTX 1080 Ti uses NVIDIA's Pascal architecture on the GP102 die, fabricated at TSMC on a 16 nm process. The RX 570X uses AMD's GCN 4.0 architecture on the Polaris 20 die, fabricated at GlobalFoundries on a 14 nm node.

The scale difference is dramatic. GP102 packs 11,800 million transistors onto a 471 mm² die, while Polaris 20 fits 5,700 million into 232 mm². Transistor density is nearly identical, at 25.1M per mm² for the 1080 Ti versus 24.6M per mm² for the RX 570X, so the 1080 Ti's advantage comes from simply using more silicon rather than denser engineering.

One architectural bright spot for the Polaris part is FP16 throughput. The RX 570X computes half-precision floats at full rate, 5.095 TFLOPS, matching its FP32 output at a 1:1 ratio. The GTX 1080 Ti throttles FP16 to a 1:64 ratio, producing just 177.2 GFLOPS, a small fraction of its 11.34 TFLOPS FP32 peak. Neither card offers dedicated AI or ray tracing hardware, so these figures only matter for workloads that lean on half-precision math.

Software support diverges modestly. Both cards cover DirectX 12 and OpenGL 4.6, but the 1080 Ti claims DirectX 12 Feature Level 12_1 versus 12_0 for the RX 570X, and Vulkan 1.4 versus 1.3. Lineage also differs: the 1080 Ti bridges NVIDIA's GeForce 900 and GeForce 20 generations, while the RX 570X belongs to the Polaris RX 500X line, preceded by Polaris and succeeded by Vega.

Specification Differences

The core counts tell the story before any benchmark runs. The GTX 1080 Ti fields 3584 shading units, 224 TMUs, and 88 ROPs; the RX 570X counters with 2048, 128, and 32 respectively. That imbalance flows directly into theoretical rates: 139.2 GPixel/s versus 39.81 GPixel/s for pixel fill, and 354.4 GTexel/s versus 159.2 GTexel/s for texture fill.

Clocks favor NVIDIA as well, with a 1481 MHz base and 1582 MHz boost against AMD's 1168 MHz and 1244 MHz. Memory follows suit: the 1080 Ti's GDDR5X runs at 1376 MHz, 11 Gbps effective, while the RX 570X's GDDR5 runs at 1750 MHz, 7 Gbps effective, on a narrower bus.

Physical and electrical specs split the other way. The 1080 Ti measures 267 mm long, 112 mm tall, and 40 mm wide in a dual-slot format, drawing power through one 6-pin and one 8-pin connector. The RX 570X is shorter at 241 mm, also dual-slot, and needs only one 6-pin. Both use PCIe 3.0 x16. Display output is the other spec-page difference: the RX 570X keeps a DVI port, while the 1080 Ti offers HDMI 2.0 plus three DisplayPort 1.4a connections. Release timing differs too, with the 1080 Ti arriving on 2017-03-09 and the RX 570X on 2018-04-10, and the 1080 Ti carries a launch MSRP of 699 USD.

Head-to-Head Benchmarks

Only three tests appear in both cards' records, and the GTX 1080 Ti sweeps them.

PassMark G3D is the headline number and it is not close: 18600 for the GTX 1080 Ti against 1923 for the RX 570X, a delta of 867.2 percent. That figure dwarfs every other result on the page and signals that in PassMark's primary graphics metric the two cards occupy entirely different performance classes. Even the RX 570X's own database percentile of 55, which benefits from a broader mix of tests, cannot soften a near tenfold gap on this shared metric.

Geekbench OpenCL narrows the gap considerably but does not flip it. The 1080 Ti scores 67929 to the RX 570X's 38939, a 74.4 percent advantage. This is the compute-focused test of the three, and interestingly it is where the RX 570X posts its strongest relative showing against Pascal, consistent with its 1:1 FP16 capability noted in the architecture section. Still, the 1080 Ti's raw compute throughput prevails by a wide margin.

PassMark G2D, the 2D test, goes to NVIDIA by 25.2 percent, 939 versus 750. This is the smallest recorded gap and the least relevant to modern 3D gaming, but it completes the sweep.

The wider benchmark record underscores the 1080 Ti's depth. Beyond the shared tests, its results include a Geekbench Vulkan score of 40511, Geekbench Metal at 30624, a PassMark GPU Compute score of 9632, and 3DMark Steel Nomad DX12 at 2231, results for which the RX 570X has no recorded counterpart in the database. Against its nearest rivals, the 1080 Ti's average of 15548 sits within 1.7 percent of the RTX 2060, GTX 580, Radeon Pro W5500, and Radeon R9 M380, while the RX 570X's 13871 keeps it within about one percent of the RTX A2000 Mobile, Radeon 660M, RX 7900 XT, and Tesla K10.

The final tally: three wins for the GTX 1080 Ti, zero for the RX 570X. Performance buyers land on Pascal; only the power and physical constraints argue for Polaris.