AMD Radeon R9 M380 vs NVIDIA GeForce GTX 660 Ti Comparison

AMD Radeon R9 M380 and NVIDIA GeForce GTX 660 Ti are two end-of-life graphics cards from different eras, with the R9 M380 targeting mobile integration and the GTX 660 Ti built for desktop towers. The data shows a clear performance gap: the GTX 660 Ti averages 15,063 across benchmarks, while the R9 M380 sits at 15,521, but those figures conceal significant architectural and feature differences. The GTX 660 Ti wins the only direct head-to-head benchmark, yet the R9 M380 holds advantages in memory capacity and API support that matter for certain workloads. Below, the facts from the benchmark database break down where each card stands.

FAQ

Q: Which card has the higher average benchmark score?

A: The AMD Radeon R9 M380 scores 15,521 on average, which is 458 points higher than the NVIDIA GeForce GTX 660 Ti’s 15,063. This places the R9 M380 at the 58th percentile of all GPUs, versus the GTX 660 Ti’s 57th percentile.

Q: How do the two cards compare in the only direct head-to-head test?

A: In the Geekbench OpenCL test, the GTX 660 Ti scores 15,113 against the R9 M380’s 12,565. That is a 16.9% margin in favor of the NVIDIA card, making it the sole winner among the recorded head-to-head benchmarks.

Q: What are the memory specifications for each card?

A: The R9 M380 comes with 4 GB of GDDR5 memory on a 128-bit bus, delivering 96.00 GB/s of bandwidth. The GTX 660 Ti has 2 GB of GDDR5 on a 192-bit bus, providing 144.2 GB/s of bandwidth.

Q: Which card supports newer graphics APIs?

A: The R9 M380 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.2.170. The GTX 660 Ti supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175, meaning the AMD part has a higher DirectX feature level.

Q: What is the transistor count and die size difference?

A: The GTX 660 Ti uses 3,540 million transistors on a 294 mm² die, while the R9 M380 packs 2,080 million transistors into 160 mm². Both are built on TSMC’s 28 nm process, but the NVIDIA chip has higher transistor density per area at 12.0M/mm² versus 13.0M/mm² for AMD.

Q: Does either card have a launch MSRP listed?

A: Only the GTX 660 Ti has a launch MSRP, which was 299 USD. The R9 M380 has no launch MSRP in the data.

Where Each One Wins

The NVIDIA GeForce GTX 660 Ti wins decisively in raw compute performance. Its Geekbench OpenCL score of 15,113 beats the R9 M380’s 12,565, a 16.9% advantage. This makes the GTX 660 Ti the better choice for compute-heavy tasks like OpenCL acceleration, where its higher shading unit count (1,344 vs 768) and larger texture unit array (112 vs 48) translate directly into faster number crunching. The GTX 660 Ti also leads in pixel throughput with 27.44 GPixel/s versus 16.00 GPixel/s, and texture fill rate at 109.8 GTexel/s versus 48.00 GTexel/s, making it superior for traditional rasterization workloads.

The AMD Radeon R9 M380 wins on memory capacity and API modernity. With 4 GB of VRAM, it doubles the GTX 660 Ti’s 2 GB, which is critical for modern games and applications that exceed 2 GB of frame buffer usage. The R9 M380’s DirectX 12 (12_0) support is a full feature level above the GTX 660 Ti’s DirectX 12 (11_0), enabling access to more advanced rendering techniques. Its higher average benchmark score of 15,521 also shows that in aggregate, it edges out the NVIDIA card across the full test suite, even though it loses the single direct comparison.

Architecture Differences

The two cards represent fundamentally different GPU architectures. The R9 M380 uses AMD’s GCN 2.0 architecture on the Strato chip, while the GTX 660 Ti uses NVIDIA’s Kepler architecture on the GK104 chip. Both are fabricated on TSMC’s 28 nm process, but the similarities end there. The GTX 660 Ti’s GK104 is a much larger chip at 294 mm² with 3,540 million transistors, versus the R9 M380’s 160 mm² die with 2,080 million transistors. This gives the NVIDIA chip a raw resource advantage: 1,344 shading units, 112 TMUs, and 24 ROPs, compared to the R9 M380’s 768 shading units, 48 TMUs, and 16 ROPs.

The architectural philosophies diverge in how those resources are used. Kepler was designed for high clock efficiency and massive parallelism, while GCN 2.0 emphasized compute throughput and asynchronous execution. The R9 M380’s transistor density is slightly higher at 13.0M/mm² versus 12.0M/mm², suggesting a more compact design despite fewer total transistors. Neither card has ray tracing or tensor cores, as both predate those features. The R9 M380 is part of the Gem System (R9 M300) generation, while the GTX 660 Ti belongs to the GeForce 600 generation, with the former succeeding the Solar System and the latter succeeding the GeForce 500 series.

Specification Differences

The most striking specification gap is in memory. The R9 M380 offers 4 GB of GDDR5 on a 128-bit bus with 96.00 GB/s bandwidth, while the GTX 660 Ti has 2 GB of GDDR5 on a 192-bit bus with 144.2 GB/s bandwidth. The NVIDIA card’s wider bus gives it 50% more bandwidth, but the AMD card’s double capacity is more future-proof for texture-heavy workloads.

Clock speeds are close but not identical. The R9 M380 runs at 900 MHz base and 1000 MHz boost, while the GTX 660 Ti runs at 915 MHz base and 980 MHz boost. The R9 M380’s boost clock is 20 MHz higher, but the GTX 660 Ti’s base clock is 15 MHz higher, resulting in a near-tie in raw frequency. Memory clocks are effectively the same at 1500 MHz / 1502 MHz, both delivering 6 Gbps effective.

Compute rates favor the GTX 660 Ti significantly: 2.634 TFLOPS FP32 versus 1.536 TFLOPS for the R9 M380. The NVIDIA card also has a 150 W TDP, dual-slot cooler, and 2x 6-pin power connectors, requiring a 450 W PSU, while the R9 M380 has no TDP or power specifications listed. The GTX 660 Ti is 241 mm (9.5 inches) long with 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2 outputs; the R9 M380 has no listed dimensions or display outputs. The R9 M380 was released on May 4, 2015, while the GTX 660 Ti came out on August 15, 2012.

Head-to-Head Benchmarks

The database records only one direct benchmark between these two cards: Geekbench OpenCL. In that test, the GTX 660 Ti scores 15,113, defeating the R9 M380’s 12,565 by a 16.9% margin. This is the sole head-to-head result, and it goes to NVIDIA. The R9 M380’s average benchmark score of 15,521 is higher than the GTX 660 Ti’s 15,063, but that average includes different tests (Geekbench Metal and Geekbench OpenCL for AMD, versus Geekbench OpenCL and Geekbench Vulkan for NVIDIA). The only shared test is OpenCL, where NVIDIA wins outright.

Looking at the nearest rivals provides context. The R9 M380’s 15,521 average puts it within 0.2% of the NVIDIA GeForce GTX 1080 Ti (15,548) and 1% behind the AMD Radeon Pro W5500 (15,679), while sitting 1.5% ahead of the RTX 2060 (15,290) and 1.6% ahead of the GTX 580 (15,283). The GTX 660 Ti’s 15,063 average places it 0.1% behind the AMD Radeon Pro 560X (15,082), 0.7% behind the RX 7600 (15,171), 0.9% behind the RTX 3050 OEM (15,199), and 1.4% behind the AMD Radeon 680M (15,270). These deltas show both cards performing in a similar tier, but the R9 M380’s placement among much newer hardware suggests its compute features age better.

The Verdict

Pick the NVIDIA GeForce GTX 660 Ti if you need raw compute performance and are working within OpenCL or Vulkan workloads. Its 16.9% lead in the direct OpenCL test, combined with 2.634 TFLOPS FP32 versus 1.536 TFLOPS, makes it the stronger choice for number-crunching tasks. The higher pixel rate (27.44 GPixel/s) and texture rate (109.8 GTexel/s) also favor it for traditional rendering pipelines. Its 2 GB frame buffer and 144.2 GB/s bandwidth are sufficient for older titles and lighter loads, and its 150 W TDP with dual-slot cooling is well-documented for desktop builds.

Pick the AMD Radeon R9 M380 if memory capacity and API support matter more than raw speed. Its 4 GB VRAM is double the GTX 660 Ti’s, which prevents stuttering in modern games that exceed 2 GB. The DirectX 12 (12_0) feature level is a generation ahead of the GTX 660 Ti’s DirectX 12 (11_0), offering access to newer rendering features. Its higher average benchmark score (15,521 vs 15,063) and superior percentile ranking (58th vs 57th) indicate better overall performance across diverse test suites, even if it loses the only direct comparison. The R9 M380’s smaller die (160 mm² vs 294 mm²) and lower transistor count (2,080M vs 3,540M) also suggest better power efficiency per millimeter, though no TDP is listed to confirm.

The data does not support a universal winner. The GTX 660 Ti is a compute specialist with clear victories in fill rates and OpenCL, while the R9 M380 is a capacity-focused card with better memory headroom and modern API support. For users prioritizing frame rates in older DirectX 11 titles, the GTX 660 Ti’s higher texture and pixel throughput will show tangible gains. For users running newer software that demands more than 2 GB of VRAM or leverages DirectX 12 features, the R9 M380 is the safer bet. Both are end-of-life products, so the choice comes down to your specific workload requirements rather than future upgrade potential.