AMD Radeon Pro 555X vs NVIDIA GeForce GTX 1660 Comparison

Head-to-Head Benchmarks

The recorded data contains two direct comparisons between the AMD Radeon Pro 555X and the NVIDIA GeForce GTX 1660, and both are decisive victories for the NVIDIA part. In the Geekbench OpenCL test, the GTX 1660 scores 47,850 against the Radeon Pro 555X's 12,628, a margin of 73.6 percent. That is not a close race; it is a generational gap expressed in raw compute throughput. The Vulkan result tells the same story: the GTX 1660 posts 50,137, while the Radeon Pro 555X manages 12,855, a 74.4 percent deficit. In both API environments, the NVIDIA card more than triples the AMD part's output.

The broader database averages reinforce this separation. The Radeon Pro 555X carries an average benchmark score of 13,321, while the GTX 1660 averages 11,680. That is an unusual situation: the older, lower-power AMD part actually holds a higher aggregate average than its newer rival. This is explained by the different benchmark sets recorded for each card. The AMD part's three recorded tests are all Geekbench compute workloads (Metal, OpenCL, Vulkan), which favor its GCN architecture's compute throughput. The GTX 1660's ten recorded tests include Passmark DX9, DX10, DX11, DX12, G2D, G3D, and GPU compute, which pull its average down. The GTX 1660's Passmark G3D score of 11,646 is strong, but its Passmark DX9 score of 177 and DX12 score of 49 are low absolute numbers that dilute the average.

The percentile rankings are close despite the head-to-head blowout. The Radeon Pro 555X sits at the 54th percentile of all GPUs in the database, while the GTX 1660 sits at the 51st. This suggests that in the full context of every recorded GPU, both cards occupy similar mid-tier territory. The nearest rivals for the Radeon Pro 555X include the NVIDIA GeForce GTX 480 (average score 13,300, a 0.2 percent difference) and the AMD FirePro M6100 (13,354, 0.2 percent behind the 555X). The GTX 1660's nearest neighbors include the AMD Radeon RX 7800 XT (11,627, only 0.5 percent behind) and the AMD Radeon RX 6500 XT (11,842, 1.4 percent ahead). The data shows that the Radeon Pro 555X is competitive with a 2010-era flagship, while the GTX 1660 trades blows with much newer Radeon parts in the aggregate database, even though its raw compute scores dwarf the 555X.

FAQ

Q: Which card wins the direct benchmark comparisons?

A: The NVIDIA GeForce GTX 1660 wins both recorded head-to-head tests. It scores 47,850 in Geekbench OpenCL versus 12,628 for the AMD Radeon Pro 555X, and 50,137 in Geekbench Vulkan versus 12,855. The deltas are 73.6 percent and 74.4 percent, respectively.

Q: Why does the Radeon Pro 555X have a higher average benchmark score than the GTX 1660?

A: The Radeon Pro 555X's average is 13,321, while the GTX 1660's is 11,680. This occurs because the two cards have different recorded benchmark sets. The AMD card's three tests are all Geekbench compute workloads, while the GTX 1660's ten tests include several low-scoring Passmark legacy API tests, such as 177 in DX9 and 49 in DX12, which lower its aggregate.

Q: What is the memory configuration difference?

A: The Radeon Pro 555X has 4 GB of GDDR5 on a 128-bit bus, yielding 94.08 GB/s of bandwidth. The GTX 1660 has 6 GB of GDDR5 on a 192-bit bus, yielding 192.1 GB/s, which is roughly double the bandwidth.

Q: Do both cards support the same graphics APIs?

A: No. Both support DirectX 12 and OpenGL 4.6, but the Radeon Pro 555X is rated for DirectX 12 (12_0) and Vulkan 1.3, while the GTX 1660 supports DirectX 12 (12_1) and Vulkan 1.4. The NVIDIA part also has a higher feature level for DirectX 12.

Q: How do their transistor counts and die sizes compare?

A: The Radeon Pro 555X contains 3,000 million transistors on a 123 mm² die fabricated by GlobalFoundries. The GTX 1660 contains 6,600 million transistors on a 284 mm² die fabricated by TSMC. The transistor densities are close: 24.4 million per mm² for the AMD chip and 23.2 million per mm² for the NVIDIA chip.

Q: What is the power draw difference?

A: The Radeon Pro 555X is rated at 75 W TDP, while the GTX 1660 is rated at 120 W TDP. The GTX 1660 also requires a single 8-pin power connector and a 300 W suggested power supply, whereas the Radeon Pro 555X uses no power connectors and is an integrated graphics package (IGP).

Architecture Differences

The two GPUs come from entirely different architectural lineages. The AMD Radeon Pro 555X uses the Polaris 21 chip built on GCN 4.0, fabricated on a 14 nm process at GlobalFoundries. The NVIDIA GeForce GTX 1660 uses the TU116 chip built on Turing, fabricated on a 12 nm process at TSMC. The node difference is small in absolute terms, but the design philosophies diverge sharply. GCN 4.0 is a compute-oriented architecture with a 1:1 FP16 ratio, meaning the 555X's FP16 throughput equals its FP32 throughput at 1,393.2 GFLOPS. Turing, by contrast, offers a 2:1 FP16 ratio on the GTX 1660, delivering 10.05 TFLOPS of FP16 against 5.027 TFLOPS of FP32. The NVIDIA card can process half-precision data at twice the rate of full precision, which benefits workloads that tolerate reduced precision.

The shader configurations are very different. The Radeon Pro 555X has 768 shading units, 48 texture mapping units, and 16 render output units. The GTX 1660 has 1,408 shading units, 88 TMUs, and 48 ROPs. Those raw counts translate directly into throughput limits. The AMD card achieves a pixel rate of 14.51 GPixel/s and a texture rate of 43.54 GTexel/s. The GTX 1660 achieves 85.68 GPixel/s and 157.1 GTexel/s. The NVIDIA part has roughly 5.9 times the pixel fill rate and 3.6 times the texture fill rate. These are not subtle differences; they show up in any workload that stresses rasterization or texture filtering.

The memory subsystems also reflect different priorities. The Radeon Pro 555X runs GDDR5 at 1470 MHz with 5.9 Gbps effective, on a 128-bit bus, for 94.08 GB/s. The GTX 1660 runs GDDR5 at 2001 MHz with 8 Gbps effective, on a 192-bit bus, for 192.1 GB/s. The NVIDIA card has twice the memory bandwidth, which is critical for modern games and compute kernels that stream large amounts of data. The AMD card's smaller bus and slower memory clock cap its ability to feed its shader array.

The bus interfaces differ as well. The Radeon Pro 555X connects via PCIe 3.0 x8, while the GTX 1660 uses PCIe 3.0 x16. For a card with 94.08 GB/s of memory bandwidth, the x8 link is less of a bottleneck, but the GTX 1660's higher bandwidth demands and larger frame buffer make the x16 interface more appropriate. The GTX 1660 also supports DirectX 12 (12_1) and Vulkan 1.4, while the Radeon Pro 555X is limited to DirectX 12 (12_0) and Vulkan 1.3. The feature level difference matters for games that use Tier 2 or higher DirectX 12 features.

Specification Differences

The two cards differ in almost every measurable specification. The Radeon Pro 555X is a 75 W integrated GPU package with no power connectors, while the GTX 1660 is a 120 W dual-slot card requiring a single 8-pin connector and a 300 W suggested power supply. The AMD part is built on a 14 nm GlobalFoundries process with 3,000 million transistors on a 123 mm² die. The NVIDIA part is built on a 12 nm TSMC process with 6,600 million transistors on a 284 mm² die. Transistor density is nearly identical (24.4M per mm² versus 23.2M per mm²), but the NVIDIA chip is more than twice as large.

Memory differs in capacity, bus width, and bandwidth. The Radeon Pro 555X has 4 GB of GDDR5 on a 128-bit bus at 1470 MHz (5.9 Gbps effective), delivering 94.08 GB/s. The GTX 1660 has 6 GB of GDDR5 on a 192-bit bus at 2001 MHz (8 Gbps effective), delivering 192.1 GB/s. The compute configuration is also far apart: 768 shading units, 48 TMUs, and 16 ROPs on the AMD side versus 1,408 shading units, 88 TMUs, and 48 ROPs on the NVIDIA side. The GTX 1660's boost clock is 1785 MHz with a base of 1530 MHz; the Radeon Pro 555X has no recorded base or boost clock in the database.

The FP32 throughput tells the story plainly. The Radeon Pro 555X manages 1,393.2 GFLOPS. The GTX 1660 manages 5.027 TFLOPS, which is 5,027 GFLOPS. That is a 3.6 times difference in raw single-precision compute. The FP16 numbers are even more lopsided because of the 2:1 ratio on Turing: 1,393.2 GFLOPS versus 10.05 TFLOPS. The GTX 1660 also has a longer physical footprint at 229 mm, 111 mm height, and 35 mm width, while the Radeon Pro 555X is an IGP with no recorded dimensions. Display outputs also differ: the Radeon Pro 555X is "Portable Device Dependent," while the GTX 1660 offers 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a.

Where Each One Wins

The Radeon Pro 555X wins in power efficiency and physical integration. At 75 W with no power connectors and an IGP form factor, it is designed for portable devices where space and thermal headroom are constrained. It is also an older, end-of-life product with a release date of July 2018, and its compute-heavy GCN architecture still produces respectable Geekbench Metal scores (14,479) and OpenCL scores (12,628) for its class. Its average benchmark score of 13,321 places it at the 54th percentile, slightly above the GTX 1660's 51st percentile, though this is due to the different test suites rather than raw performance superiority.

The GTX 1660 wins everywhere performance matters. It dominates in OpenCL and Vulkan compute, has 3.6 times the FP32 throughput, 5.9 times the pixel rate, 3.6 times the texture rate, and more than double the memory bandwidth. It has more VRAM (6 GB versus 4 GB), a wider memory bus, a higher DirectX feature level (12_1 versus 12_0), and a newer Vulkan version (1.4 versus 1.3). Its Passmark G3D score of 11,646 and GPU compute score of 4,963 indicate strong rasterization and compute performance for its era. The GTX 1660 also has a defined physical design: a dual-slot card with standard display outputs, making it a drop-in desktop GPU, whereas the Radeon Pro 555X is tied to a portable device's integrated design.

The Verdict

The data is unambiguous. The NVIDIA GeForce GTX 1660 is the far faster GPU in every recorded head-to-head test, with 73.6 percent and 74.4 percent leads in OpenCL and Vulkan, respectively. Its compute, fill rate, memory bandwidth, and feature set all point to a card designed for real gaming and general-purpose GPU work. The AMD Radeon Pro 555X, by contrast, is a low-power integrated solution for portable Mac systems, with a 75 W TDP, no power connectors, and a portable-device-dependent display output. Its 54th percentile ranking and 13,321 average score are respectable for its niche, but that niche is not high-performance rendering.

Anyone choosing between these two for a desktop build should pick the GTX 1660 without hesitation. It offers 6 GB of VRAM, 192.1 GB/s of bandwidth, 5.027 TFLOPS of FP32, and a 1785 MHz boost clock, all within a 120 W envelope that a 300 W power supply can handle. The Radeon Pro 555X is only relevant if the requirement is a low-power integrated GPU for a portable device, where its 75 W TDP and connector-free design are advantages. Even then, its 1,393.2 GFLOPS of FP32 and 94.08 GB/s of bandwidth put it firmly in entry-level territory. The GTX 1660 is the clear performance winner, and the benchmark records show no scenario where the Radeon Pro 555X comes out ahead in a direct comparison.