AMD Radeon 660M vs NVIDIA GeForce GTX 570 Comparison

The AMD Radeon 660M and the NVIDIA GeForce GTX 570 represent two distinct eras of GPU design, separated by over a decade of architectural evolution. The 660M is a modern integrated graphics processor built on TSMC's 6 nm process, while the GTX 570 is a legacy discrete card from the Fermi generation on a 40 nm node. Despite their differences, benchmark data places them remarkably close in overall compute performance, with the GTX 570 holding a narrow edge in the single available head-to-head test.

Head-to-Head Benchmarks

The only direct comparison available is the Geekbench OpenCL test. In this benchmark, the NVIDIA GeForce GTX 570 scores 13,515 points, while the AMD Radeon 660M scores 12,876 points. The delta percentage shows the GTX 570 leading by 4.7%. This is a decisive, though not overwhelming, victory for the older discrete card. The 639-point gap places the GTX 570 clearly ahead, but the margin suggests the two are in the same performance tier rather than different leagues.

Looking at the broader benchmark landscape, the average scores reinforce this close competition. The Radeon 660M has an average benchmark score of 13,812, which is slightly higher than its single OpenCL result. The GTX 570's average score of 13,515 is identical to its OpenCL score, as that is its only listed benchmark. The difference between their averages is 297 points, favoring the AMD part. This discrepancy highlights that the 660M may perform better in other workloads, such as the Geekbench Vulkan test where it scores 14,748, a result not available for the GTX 570.

When placed against their respective peer groups, both GPUs occupy similar percentile rankings. The Radeon 660M sits at the 55th percentile of all GPUs, while the GTX 570 is at the 54th percentile. This one-percentile difference is negligible in practical terms. The nearest rivals for the 660M include the NVIDIA RTX A2000 Mobile with an average score of 13,821 (a 0.1% difference) and the AMD Radeon RX 570X at 13,871 (a 0.4% difference). The GTX 570's closest competitors include the AMD Radeon HD 7770M at 13,536 (a 0.2% difference) and the AMD Radeon RX 9070 XT at 13,543 (also a 0.2% difference). These tight clusters indicate that both cards are positioned in a crowded mid-range performance band.

The Verdict

The data presents a nuanced picture for potential users. The NVIDIA GeForce GTX 570 wins the sole direct head-to-head benchmark, the Geekbench OpenCL test, by a 4.7% margin. For applications that rely heavily on OpenCL compute, the GTX 570 is the safer choice based strictly on this metric. Its 480 shading units and 60 texture mapping units provide a raw compute foundation that still holds up against modern integrated graphics.

However, the AMD Radeon 660M counters with a superior average benchmark score of 13,812 versus 13,515, and it offers a Vulkan score of 14,748 that the GTX 570 cannot match, as the older card lists no Vulkan support. For users working with Vulkan-based applications or games, the 660M is the only viable option between the two. The 660M also achieves this performance with a 40 W TDP, in stark contrast to the GTX 570's 219 W TDP, making it dramatically more power-efficient.

The verdict depends entirely on the use case. For legacy OpenCL compute tasks where power consumption is not a concern, the GTX 570's 4.7% lead in that specific test makes it the nominal winner. For modern workloads, API compatibility, and efficiency, the Radeon 660M is the more versatile and future-proof part. The GTX 570's lack of Vulkan support is a critical disqualifier for any modern graphics workload, despite its slight edge in OpenCL.

Architecture Differences

The architectural gap between these two GPUs is vast. The AMD Radeon 660M is built on the RDNA 2.0 architecture, specifically the Rembrandt chip, which is a Navi II IGP designed for mobile integration. It uses a 6 nm process from TSMC. The NVIDIA GeForce GTX 570 uses the Fermi 2.0 architecture with the GF110 chip, a discrete design from the GeForce 500 generation, fabricated on a 40 nm process.

The transistor counts tell a story of process efficiency. The 660M packs 13,100 million transistors into a 208 mm² die, yielding a transistor density of 63.0 million per square millimeter. The GTX 570 has only 3,000 million transistors on a much larger 520 mm² die, resulting in a density of just 5.8 million per square millimeter. This 10.9x difference in density is a direct consequence of the 34 nm process node advantage.

Feature support diverges sharply. The Radeon 660M includes 6 ray tracing cores and supports DirectX 12 Ultimate (12_2) as well as Vulkan 1.4. The GTX 570 has no ray tracing cores and only supports DirectX 12 (11_0), with no Vulkan support listed. The 660M also supports OpenGL 4.6, matching the GTX 570 in that regard, but its API list is far more modern and comprehensive.

Specification Differences

The specifications reveal how each card approaches its workload. The AMD Radeon 660M has 384 shading units, 24 texture mapping units, and 16 raster operations pipelines. The NVIDIA GeForce GTX 570 has 480 shading units, 60 TMUs, and 40 ROPs. The GTX 570 has a clear advantage in all three categories, with 25% more shaders, 150% more TMUs, and 150% more ROPs.

Clock speeds differ significantly. The 660M operates at a base clock of 1500 MHz with a boost clock of 1900 MHz. The GTX 570 has no base or boost clock listed, but its memory runs at 950 MHz (3.8 Gbps effective). In terms of throughput, the 660M achieves a pixel rate of 30.40 GPixel/s and a texture rate of 45.60 GTexel/s. The GTX 570 has a lower pixel rate of 21.96 GPixel/s but a comparable texture rate of 43.92 GTexel/s.

Memory configurations are fundamentally different. The 660M uses system shared memory, meaning its size, type, bus width, and bandwidth are all system dependent. The GTX 570 has 1280 MB of dedicated GDDR5 memory on a 320-bit bus, delivering 152.0 GB/s of bandwidth. The 660M's FP32 performance is 1,459.2 GFLOPS, slightly higher than the GTX 570's 1,405.4 GFLOPS. The 660M also supports FP16 at 2.918 TFLOPS (2:1), while the GTX 570 has no listed FP16 capability.

Physical and power specifications are starkly contrasted. The 660M is an IGP with no slot width, no power connectors, and a 40 W TDP. The GTX 570 is a dual-slot card measuring 267 mm (10.5 inches) long, requiring two 6-pin power connectors and a 550 W suggested PSU, with a 219 W TDP. The bus interfaces also differ: PCIe 4.0 x8 for the 660M versus PCIe 2.0 x16 for the GTX 570. The 660M's display outputs are portable device dependent, while the GTX 570 has 2x DVI and 1x mini-HDMI 1.3a.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon 660M has an average benchmark score of 13,812, while the NVIDIA GeForce GTX 570 has an average score of 13,515. The 660M leads by 297 points.

Q: Does the NVIDIA GeForce GTX 570 support Vulkan?

A: No, the GTX 570 has no Vulkan support listed in its API specifications. The AMD Radeon 660M supports Vulkan 1.4.

Q: What is the TDP difference between the two cards?

A: The AMD Radeon 660M has a TDP of 40 W, while the NVIDIA GeForce GTX 570 has a TDP of 219 W. The 660M consumes significantly less power.

Q: How much memory does each GPU have?

A: The AMD Radeon 660M uses system shared memory, so its size is system dependent. The NVIDIA GeForce GTX 570 has 1280 MB of dedicated GDDR5 memory with a 320-bit bus.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce GTX 570 has 480 shading units, while the AMD Radeon 660M has 384. The GTX 570 has 25% more shaders.

Q: What is the launch MSRP of the NVIDIA GeForce GTX 570?

A: The launch MSRP of the NVIDIA GeForce GTX 570 is 349 USD. The AMD Radeon 660M has no listed launch MSRP.

Where Each One Wins

The AMD Radeon 660M wins in scenarios that demand modern API support and efficiency. Its Vulkan 1.4 support and DirectX 12 Ultimate (12_2) make it compatible with current and future software. The 660M's 14,748 Vulkan score demonstrates strong performance in that API, a test the GTX 570 cannot even run. Its superior FP32 throughput of 1,459.2 GFLOPS and FP16 capability give it an edge in compute tasks that leverage these features. The 660M's 6 nm process and 40 W TDP make it ideal for portable devices and power-constrained environments. Its higher average benchmark score of 13,812 indicates more consistent performance across multiple test types.

The NVIDIA GeForce GTX 570 wins in the specific OpenCL benchmark, scoring 13,515 versus 12,876, a 4.7% advantage. This makes it the better choice for legacy OpenCL applications that do not use Vulkan. Its 480 shading units, 60 TMUs, and 40 ROPs provide raw rasterization power that exceeds the 660M in traditional pixel-pushing workloads. The dedicated 1280 MB GDDR5 memory with 152.0 GB/s bandwidth offers predictable, low-latency memory access that system shared memory cannot guarantee. The GTX 570's 40 ROPs deliver a fill rate advantage in high-resolution rendering scenarios, despite its lower pixel rate of 21.96 GPixel/s compared to the 660M's 30.40 GPixel/s.

For users building a modern system with a focus on energy efficiency and API compatibility, the AMD Radeon 660M is the clear winner. For those running older OpenCL-centric software on a desktop with ample power delivery, the NVIDIA GeForce GTX 570 provides a slight performance edge in that specific test. The data ultimately shows two capable GPUs that, despite their 12-year release gap, land within 5% of each other in raw compute, with the modern part offering substantially better features and efficiency.