AMD Radeon 660M vs NVIDIA GeForce GTX 580 Comparison

Head-to-Head Benchmarks

The benchmark database records a single direct head-to-head comparison between these two GPUs, and the result is decisive. In the Geekbench OpenCL test, the NVIDIA GeForce GTX 580 scores 15,283 points against the AMD Radeon 660M's 12,876 points. That is an 18.7% advantage for the older discrete card, a significant margin that places the GTX 580 clearly ahead in raw compute workloads as measured by this test.

The GTX 580's score of 15,283 places it at the 58th percentile among all GPUs in the database, while the Radeon 660M's average benchmark score of 13,812 sits at the 55th percentile. When considering the Radeon 660M's best recorded result, its Vulkan score of 14,748 is substantially higher than its OpenCL result, yet it still falls short of the GTX 580's OpenCL number. The gap between the two in their respective strongest tests is roughly 3.5%, a much tighter margin than the OpenCL head-to-head suggests.

The head-to-head data shows one win for the GTX 580 and zero for the Radeon 660M. However, the Radeon 660M does not participate in the same test suite across the board, as it has an additional Vulkan benchmark entry that the GTX 580 lacks entirely. This absence of a Vulkan score for the GTX 580 means the database cannot directly compare their performance under that API, leaving the OpenCL result as the only common metric.

Looking at the nearest rivals in the database provides useful context. The GTX 580's score of 15,283 is virtually identical to the NVIDIA GeForce RTX 2060, which averages 15,290, a delta of 0%. It also edges out the AMD Radeon 680M (15,270, 0.1% slower), the NVIDIA GeForce RTX 3050 OEM (15,199, 0.6% slower), and the AMD Radeon RX 7600 (15,171, 0.7% slower). This clustering suggests the GTX 580, despite its age, delivers compute performance on par with much newer mid-range hardware.

The Radeon 660M's average score of 13,812 places it near the NVIDIA RTX A2000 Mobile, which averages 13,821, a delta of -0.1% in the Radeon's favor. It also sits close to the AMD Radeon RX 570X (13,871, 0.4% ahead), the AMD Radeon RX 7900 XT (13,745, 0.5% behind), and the NVIDIA Tesla K10 (14,029, 1.5% behind). This positioning shows the 660M performing in a mid-range bracket, well below the GTX 580's tier.

Where Each One Wins

The GTX 580 wins the only direct benchmark comparison available, taking the Geekbench OpenCL test by 18.7%. This is a clear victory in general-purpose compute workloads, and the score distribution around 15,283 shows it performing at a level comparable to modern discrete GPUs like the RTX 2060 and RX 7600. For any application relying on OpenCL compute, the data firmly favors the GTX 580.

The Radeon 660M, despite losing the OpenCL head-to-head, demonstrates a notable strength in Vulkan workloads. Its Geekbench Vulkan score of 14,748 is 14.5% higher than its own OpenCL score of 12,876. This suggests the RDNA 2.0 architecture handles Vulkan more efficiently than OpenCL, making the 660M a more competitive option in Vulkan-based applications. However, since the GTX 580 has no recorded Vulkan score, the database cannot confirm whether this advantage would hold in a direct comparison.

For gaming and graphics workloads, the Radeon 660M's feature set, including DirectX 12 Ultimate support and Vulkan 1.4 compatibility, gives it a modern API advantage. The GTX 580 only supports DirectX 12 (11_0) and has no recorded Vulkan support. This means the 660M can run titles that require newer API features, while the GTX 580 may face compatibility limitations.

The Radeon 660M's integrated design, with a 40 W TDP and no power connectors, makes it suitable for portable devices where power draw and physical space are constrained. The GTX 580, with a 244 W TDP, dual-slot cooler, and external power connectors, is a different class of hardware entirely. For users prioritizing low power consumption and compact systems, the 660M is the only viable option in this pairing.

FAQ

Q: Which GPU has the higher benchmark score in the direct comparison?

A: The NVIDIA GeForce GTX 580 scores 15,283 in Geekbench OpenCL, which is 18.7% higher than the AMD Radeon 660M's 12,876 in the same test.

Q: How does the Radeon 660M perform in Vulkan compared to OpenCL?

A: The Radeon 660M records a Vulkan score of 14,748, which is significantly higher than its OpenCL score of 12,876, indicating stronger performance under the Vulkan API.

Q: What is the percentile ranking of each GPU?

A: The GTX 580 sits at the 58th percentile among all GPUs, while the Radeon 660M holds the 55th percentile based on its average benchmark score.

Q: Does the GTX 580 have a Vulkan benchmark result?

A: No, the database records no Vulkan score for the GTX 580. It only has a Geekbench OpenCL result of 15,283.

Q: How close is the GTX 580 to modern GPUs in performance?

A: The GTX 580's score of 15,283 is within 0.7% of the AMD Radeon RX 7600 (15,171) and within 0.6% of the NVIDIA GeForce RTX 3050 OEM (15,199), placing it in the same performance tier.

Q: What are the power requirements for each GPU?

A: The GTX 580 has a 244 W TDP and requires both a 6-pin and 8-pin power connector, while the Radeon 660M has a 40 W TDP and uses no external power connectors.

Specification Differences

The two GPUs differ substantially across nearly every specification category. The GTX 580 uses 1,536 MB of dedicated GDDR5 memory on a 384-bit bus, delivering 192.4 GB/s of bandwidth. The Radeon 660M instead relies on system shared memory, with bandwidth described as system dependent. This is a fundamental architectural difference: one has dedicated VRAM, the other borrows from the host system.

Clock speeds also diverge. The GTX 580's memory runs at 1,002 MHz (4 Gbps effective), while the Radeon 660M has a base clock of 1,500 MHz and a boost clock of 1,900 MHz. The Radeon's higher clock speeds reflect its modern process node and design priorities, but they operate within a shared memory context.

The compute unit counts favor the GTX 580 in raw numbers. It has 512 shading units, 64 texture mapping units, and 48 raster operation units. The Radeon 660M has 384 shading units, 24 TMUs, and 16 ROPs. Despite fewer units, the Radeon achieves a higher pixel rate of 30.40 GPixel/s compared to the GTX 580's 24.70 GPixel/s, though the GTX 580's texture rate of 49.41 GTexel/s edges out the Radeon's 45.60 GTexel/s.

In floating-point performance, the GTX 580 delivers 1.581 TFLOPS of FP32 compute, while the Radeon 660M provides 1,459.2 GFLOPS (approximately 1.459 TFLOPS). The Radeon also supports FP16 at 2.918 TFLOPS (2:1), a capability the GTX 580 lacks entirely.

The bus interface and physical design differ completely. The GTX 580 uses PCIe 2.0 x16, measures 267 mm in length, and occupies a dual-slot profile. The Radeon 660M uses PCIe 4.0 x8 and is an integrated GPU (IGP) with no standalone dimensions. Display outputs also differ: the GTX 580 offers 2x DVI and 1x mini-HDMI 1.3a, while the Radeon 660M's outputs are portable device dependent.

Architecture Differences

The GTX 580 is built on NVIDIA's Fermi 2.0 architecture using the GF110 chip, manufactured on a 40 nm process at TSMC. It contains 3,000 million transistors on a 520 mm² die, yielding a transistor density of 5.8 million transistors per square millimeter. This is a large, power-hungry design from the GeForce 500 generation.

The Radeon 660M uses AMD's RDNA 2.0 architecture with the Rembrandt chip, fabricated on a 6 nm process, also at TSMC. It packs 13,100 million transistors onto a 208 mm² die, achieving a transistor density of 63.0 million per square millimeter. This is over ten times the density of the GTX 580, reflecting the massive process node advantage.

The Radeon 660M includes 6 ray tracing cores, a feature entirely absent from the GTX 580. It also supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the GTX 580 is limited to DirectX 12 (11_0) with no Vulkan support. Both GPUs support OpenGL 4.6.

The Radeon 660M belongs to the Navi II IGP generation (Rembrandt Mobile), succeeding the Vega II IGP and preceding the Navi III IGP. The GTX 580 comes from the GeForce 500 generation, with the GeForce 400 as its predecessor and GeForce 600 as its successor. The GTX 580 launched in November 2010, while the Radeon 660M arrived in January 2022, a gap of over a decade.

Power efficiency is a stark differentiator. The GTX 580 draws 244 W with a suggested power supply of 550 W, while the Radeon 660M consumes only 40 W and requires no external power. This makes the 660M suitable for integration into laptops and compact devices, whereas the GTX 580 demands a full desktop power delivery system.

The Verdict

The data points to a clear winner for raw compute performance: the NVIDIA GeForce GTX 580. Its 18.7% lead in the OpenCL benchmark is substantial, and its score of 15,283 places it alongside modern discrete GPUs like the RTX 2060 and RX 7600. For any user prioritizing OpenCL compute throughput, the GTX 580 is the superior choice based on recorded measurements.

However, the AMD Radeon 660M should not be dismissed. Its Vulkan score of 14,748 demonstrates strong performance under that API, and its modern feature set, including ray tracing cores and DirectX 12 Ultimate, makes it more future-proof for graphics workloads. The 660M also offers vastly better efficiency at 40 W versus 244 W, which is critical for portable or low-power systems.

The specification differences reinforce a generational divide. The GTX 580 uses dedicated GDDR5 memory with high bandwidth, while the 660M relies on system shared memory. The GTX 580 has more shading units and TMUs, but the 660M achieves a higher pixel rate and supports FP16 computation. The GTX 580's 40 nm process and 3,000 million transistors contrast sharply with the 660M's 6 nm process and 13,100 million transistors.

For users building a desktop system focused on compute tasks and willing to accommodate a 244 W power draw, the GTX 580 is the proven performer. For users needing an integrated solution with modern API support and minimal power consumption, the Radeon 660M is the only practical option. The database records one head-to-head victory for the GTX 580, and that result should guide any decision where OpenCL performance is the primary criterion.