AMD Radeon R5 M335 vs AMD Radeon R7 M260X Comparison

The AMD Radeon R7 M260X and AMD Radeon R5 M335 are both end-of-life mobile graphics solutions from AMD, built on the same GCN 1.0 architecture and 28 nm process at TSMC. The database records show a near-even split in benchmark victories, with each card taking one of the two head-to-head tests. The R7 M260X, built on the Opal chip with 950 million transistors and a 77 mm² die, carries a 30th percentile ranking among all GPUs. The R5 M335, using the Exo chip with 690 million transistors and a 56 mm² die, sits at the 28th percentile. These are closely matched parts, but the underlying specifications reveal distinct design priorities that translate into different workload strengths.

Head-to-Head Benchmarks

The two recorded head-to-head tests are Geekbench OpenCL and Geekbench Vulkan, and the results are split exactly one win apiece. In the OpenCL test, the R7 M260X posts a score of 5690 against 4745 for the R5 M335, a decisive 19.9% margin in favor of the R7. This is the largest performance gap in either direction across the entire comparison. The R7’s advantage in this compute-oriented API aligns with its hardware configuration: it has 384 shading units, 24 texture mapping units, and 8 ROPs, compared to 320 shading units, 20 TMUs, and 8 ROPs on the R5. The R7 also runs its memory at 1000 MHz with a 128-bit bus, delivering 64.00 GB/s of bandwidth, while the R5 operates at 900 MHz with a 64-bit bus and only 14.40 GB/s. That bandwidth difference is stark, and it shows up clearly in OpenCL throughput.

In the Vulkan test, the tables turn. The R5 M335 scores 4758, narrowly edging out the R7 M260X’s 4631, a 2.7% delta in favor of the R5. This is a much smaller margin than the OpenCL gap, but it is still a win. The R5’s higher pixel rate and texture rate likely contribute here: it reaches 8.240 GPixel/s and 20.60 GTexel/s, whereas the R7 manages 5.720 GPixel/s and 17.16 GTexel/s. Even though the R7 has more shading units and far more memory bandwidth, the R5’s higher clock-derived throughput rates give it an edge in this particular API workload. The R5’s FP32 performance is also higher at 659.2 GFLOPS versus 549.1 GFLOPS for the R7, which may explain why Vulkan, with its lower-level draw call overhead, favors the smaller chip.

Looking at average benchmark scores across all recorded tests, the R7 M260X sits at 5161, while the R5 M335 averages 4752. That is an 8.6% overall lead for the R7, driven almost entirely by its dominant OpenCL result. The Vulkan win for the R5 is real but narrow, and it does not fully offset the R7’s larger OpenCL advantage. The head-to-head record is 1-1, but the magnitude of the R7’s OpenCL win carries more weight in the aggregate data.

Where Each One Wins

The R7 M260X is the clear winner in OpenCL compute workloads. Its 19.9% lead over the R5 in that specific test is the single largest delta recorded between the two cards. This makes sense given its hardware profile: 384 shading units, a 128-bit memory bus, and 64.00 GB/s of bandwidth. The R7 is also built on a larger die with more transistors, 950 million versus 690 million, which suggests it has more raw compute resources to draw upon. For applications that leverage OpenCL for general-purpose GPU computing, rendering, or physics simulation, the R7 is the stronger choice.

The R5 M335 takes the Vulkan test, and while the margin is only 2.7%, it is a consistent result in the database. The R5’s higher pixel rate and texture rate, 8.240 GPixel/s and 20.60 GTexel/s respectively, indicate that its clock speeds are effectively higher despite the fact that no base or boost clocks are recorded for the card. The R5 also has double the memory capacity at 2 GB versus 1024 MB on the R7, although it uses slower DDR3 memory on a narrower 64-bit bus. In Vulkan-based games or applications that are less bandwidth-sensitive and more dependent on shader throughput, the R5’s higher FP32 output (659.2 GFLOPS) gives it a slight edge.

For gaming specifically, the data is mixed. The R7 wins the compute-heavy OpenCL path, which many older games and emulators use, while the R5 wins the modern Vulkan path, which is increasingly common in current titles and translation layers. Neither card is dominant across the board, and the choice depends entirely on the API in use. The R7’s memory bandwidth advantage suggests it will handle texture-heavy scenes better, while the R5’s higher pixel rate points to better fill-rate-bound performance at lower resolutions.

The R7 also posts a higher percentile ranking at 30 versus 28 for the R5, and its average benchmark score is 8.6% higher. That makes the R7 the better overall performer in the database’s aggregate metrics, despite the R5’s Vulkan win.

Architecture Differences

Both cards share the same GCN 1.0 architecture and are fabricated on TSMC’s 28 nm process, but they are built around different chips with distinct configurations. The R7 M260X uses the Opal chip, which packs 950 million transistors into a 77 mm² die, yielding a transistor density of 12.3M per mm². The R5 M335 uses the Exo chip, with 690 million transistors on a 56 mm² die, also at 12.3M per mm². The identical transistor density confirms that both chips come from the same process generation, but the Opal chip is physically larger and more complex.

The R7 has 384 shading units, 24 TMUs, and 8 ROPs, while the R5 has 320 shading units, 20 TMUs, and 8 ROPs. The R7 therefore has 20% more shading units and 20% more texture units than the R5, which explains its higher raw compute ceiling. However, the R5 compensates with higher clock-derived rates: its pixel rate of 8.240 GPixel/s is 44% higher than the R7’s 5.720 GPixel/s, and its texture rate of 20.60 GTexel/s is 20% higher than the R7’s 17.16 GTexel/s. This is a classic trade-off between more execution units at lower effective clocks versus fewer units at higher effective clocks.

Memory is where the two cards diverge most sharply. The R7 uses 1024 MB of GDDR5 on a 128-bit bus, running at 1000 MHz with 4 Gbps effective, for a bandwidth of 64.00 GB/s. The R5 uses 2 GB of DDR3 on a 64-bit bus, running at 900 MHz with 1800 Mbps effective, for a bandwidth of only 14.40 GB/s. That is a 4.4x difference in memory bandwidth in favor of the R7, and it is the single most important architectural distinction between the two. The R5’s larger capacity (2 GB versus 1 GB) is a capacity advantage, but its bandwidth limitation severely constrains performance in memory-intensive workloads.

Both cards support DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170, and both use a PCIe 3.0 x8 bus interface. Neither requires external power connectors, and both have display outputs that are portable device dependent, meaning they are designed for laptops. The R7 was released on December 5, 2015, while the R5 came slightly earlier on October 20, 2015. Both belong to the Gem System generation, with the R7 in the R7 M200 series and the R5 in the R5 M300 series. Both list Solar System as their predecessor and Polaris Mobile as their successor.

The Verdict

The data points to a nuanced conclusion. If the workload is OpenCL-based, the R7 M260X is unequivocally the better card, with a 19.9% lead in that benchmark and a higher average score overall. Its 64.00 GB/s memory bandwidth and 384 shading units give it a substantial compute advantage that the R5 cannot match. For users running OpenCL applications, GPU-accelerated tasks, or older games that use this API, the R7 is the clear pick.

If the workload is Vulkan-based, the R5 M335 wins by a slim 2.7% margin. Its higher pixel rate and texture rate, along with its 659.2 GFLOPS FP32 performance, make it slightly faster in this modern API. For users playing current Vulkan titles or using Vulkan-based emulators, the R5 is the marginally better option, though the gap is small enough that real-world differences may be imperceptible.

The R7 also wins on aggregate metrics. Its average benchmark score of 5161 is 8.6% higher than the R5’s 4752, and its 30th percentile ranking beats the R5’s 28th. The R7’s nearest rivals include the NVIDIA Quadro K3100M (0.1% ahead), the NVIDIA Quadro 4000M (1% behind), and the NVIDIA GeForce GTX 760M (1.4% behind), placing it in a competitive mid-range tier. The R5’s rivals include the AMD Radeon R8 M445DX (0.5% ahead) and the NVIDIA Quadro P400 (1.5% behind), a slightly lower tier.

The R7 M260X is the better overall GPU according to the database, despite the 1-1 head-to-head split. Its OpenCL dominance and memory bandwidth advantage outweigh the R5’s narrow Vulkan win. However, for users who specifically target Vulkan workloads and value the R5’s 2 GB memory capacity, the R5 remains a viable alternative. The verdict is straightforward: choose the R7 for compute-heavy tasks and overall performance, choose the R5 only if Vulkan is your primary API and you prefer the larger memory pool.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon R7 M260X has an average benchmark score of 5161, while the AMD Radeon R5 M335 averages 4752, giving the R7 an 8.6% lead.

Q: How big is the OpenCL performance gap between the two cards?

A: In the Geekbench OpenCL test, the R7 M260X scores 5690 versus 4745 for the R5 M335, a 19.9% difference in favor of the R7.

Q: Does the R5 M335 win any benchmark against the R7 M260X?

A: Yes, the R5 M335 wins the Geekbench Vulkan test with a score of 4758 against 4631 for the R7, a 2.7% margin.

Q: What is the memory bandwidth difference between the two cards?

A: The R7 M260X has a memory bandwidth of 64.00 GB/s using GDDR5 on a 128-bit bus, while the R5 M335 has 14.40 GB/s using DDR3 on a 64-bit bus.

Q: Why does the R5 M335 have a higher pixel rate than the R7 M260X?

A: The R5 M335 records a pixel rate of 8.240 GPixel/s and a texture rate of 20.60 GTexel/s, both higher than the R7’s 5.720 GPixel/s and 17.16 GTexel/s, despite having fewer shading units.

Q: How do these cards rank among all GPUs in the database?

A: The R7 M260X is at the 30th percentile, while the R5 M335 is at the 28th percentile, indicating that the R7 sits slightly higher in the overall performance distribution.