AMD Radeon R5 M255 vs NVIDIA GeForce 940M Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce 940M leads with an average benchmark score of 5284, while the AMD Radeon R5 M255 trails at 4788. The 940M also places higher in the overall GPU percentile, ranking at the 31st percentile versus the R5 M255's 28th.

Q: How do the two GPUs compare in Geekbench OpenCL performance?

A: The GeForce 940M is the clear winner in OpenCL, scoring 6018 against the Radeon R5 M255's 4650. This translates to a 29.4% advantage for the NVIDIA part in that specific compute workload.

Q: Does the AMD Radeon R5 M255 win any benchmark?

A: Yes, the Radeon R5 M255 wins the Geekbench Vulkan test. It scores 4925, outpacing the GeForce 940M's 4549 score, which is a 7.6% lead for AMD in that API-specific test.

Q: What are the transistor counts and die sizes of these chips?

A: The NVIDIA chip (GM107) contains 1,870 million transistors on a 148 mm² die, while the AMD chip (Topaz) has 1,550 million transistors on a 125 mm² die. Both are built on a 28 nm process at TSMC.

Q: Which GPU offers higher memory bandwidth?

A: The AMD Radeon R5 M255 provides significantly higher memory bandwidth at 32.00 GB/s, thanks to its 128-bit bus. The NVIDIA GeForce 940M, with its 64-bit bus, offers only 14.40 GB/s.

Q: What are the DirectX support levels for each GPU?

A: The AMD Radeon R5 M255 supports DirectX 12 (12_0), while the NVIDIA GeForce 940M supports DirectX 12 (11_0). This indicates a difference in the feature level each GPU can natively support.

Architecture Differences

The two mobile GPUs represent fundamentally different design philosophies from their respective manufacturers. The NVIDIA GeForce 940M is built on the Maxwell architecture, specifically using the GM107 chip, while the AMD Radeon R5 M255 relies on the GCN 3.0 architecture with the Topaz chip. Both are manufactured on the same 28 nm process at TSMC, but the silicon designs diverge considerably.

The NVIDIA chip is larger in terms of both transistor count and die area. It packs 1,870 million transistors into a 148 mm² die, resulting in a transistor density of 12.6M per mm². The AMD Topaz chip uses 1,550 million transistors on a 125 mm² die, with a slightly lower density of 12.4M per mm². This gives NVIDIA a moderate raw silicon advantage, though the architectural efficiency of each design determines how that silicon is used.

The compute resource allocation differs substantially. The GeForce 940M fields 512 shading units, 32 texture mapping units, and 16 ROPs. The Radeon R5 M255, by contrast, has 384 shading units, 24 TMUs, and only 8 ROPs. This structural difference explains much of the performance gap in fill-rate-bound workloads. The NVIDIA part delivers a pixel rate of 17.57 GPixel/s and a texture rate of 35.14 GTexel/s, while the AMD part manages 7.520 GPixel/s and 22.56 GTexel/s respectively.

Clock speeds also lean toward NVIDIA. The GeForce 940M has a base clock of 1020 MHz and a boost clock of 1098 MHz, whereas the Radeon R5 M255 runs at 925 MHz base and 940 MHz boost. The memory configurations differ as well. The NVIDIA card uses a 64-bit DDR3 interface with a 900 MHz memory clock (1800 Mbps effective), yielding 14.40 GB/s of bandwidth. The AMD card uses a wider 128-bit DDR3 interface at 1000 MHz (2 Gbps effective), which more than doubles the bandwidth to 32.00 GB/s.

The FP32 compute output reflects both the shading unit counts and clock speeds. NVIDIA's GeForce 940M achieves 1,124.4 GFLOPS, while the Radeon R5 M255 reaches 721.9 GFLOPS. Interestingly, the AMD card lists FP16 performance at the same 721.9 GFLOPS (1:1 ratio), while the NVIDIA card has no listed FP16 figure. The AMD chip also supports DirectX 12 (12_0), a higher feature level than the NVIDIA chip's DirectX 12 (11_0), and it carries Vulkan 1.2.170 support compared to NVIDIA's Vulkan 1.4. The bus interfaces further separate the two: NVIDIA uses MXM-B (3.0) and lists no power connectors with a 75 W TDP, while AMD specifies PCIe 3.0 x8 and leaves TDP unlisted.

Head-to-Head Benchmarks

The recorded data shows a split decision between the two GPUs, with each winning one of the two tests in the head-to-head set. The Geekbench OpenCL test heavily favors NVIDIA. NVIDIA's score of 6018 beats AMD's 4650 by a margin of 29.4%. This is a substantial margin in a compute test that relies heavily on raw shader throughput and memory bandwidth. The OpenCL result aligns closely with the compute resource differentials noted earlier, given the NVIDIA card's 1,124.4 GFLOPS of FP32 compute power and its 512 shading units working at higher clock speeds.

The Geekbench Vulkan test flips the tables. AMD wins that test with 4925 against NVIDIA's 4549, a 7.6% gap in AMD's favor. Vulkan is a lower-level API that can expose architectural efficiencies more directly, and the Radeon R5 M255's stronger showing in Vulkan may be tied to its memory configuration. With 32.00 GB/s of bandwidth on a 128-bit bus, the AMD part can feed its compute units more effectively in memory-sensitive workloads. The 8 ROPs and lower pixel rate do not appear to constrain it as much in this particular test.

The average benchmark scores in the database broadly reflect the mixed results. NVIDIA's average of 5284 places it above AMD's 4788, but the margins are closer than the OpenCL deltas might suggest. The database also places each GPU among a cluster of nearby rivals. NVIDIA's nearest rivals include the GeForce GTX 980M (5308, 0.4% lower), GeForce 930A (5317, 0.6% lower), GeForce 840M (5322, 0.7% lower), and GeForce GTX 760M (5236, 0.9% higher). AMD's nearest rivals include the GeForce RTX 3080 12 GB (4791, 0.1% lower), Radeon R5 M335 (4752, 0.8% higher), GeForce 940MX (4844, 1.2% lower), and Radeon R8 M445DX (4727, 1.3% higher). In both cases, the deltas are small, indicating that these are mid-pack mobile GPUs with performance tightly clustered around their respective averages.

Specification Differences

The two GPUs differ across nearly every major specification category. The NVIDIA GeForce 940M uses the GM107 chip with the Maxwell architecture, while the AMD Radeon R5 M255 uses the Topaz chip with GCN 3.0. Their transistor counts differ by 320 million, with NVIDIA at 1,870 million and AMD at 1,550 million. Die size also differs: NVIDIA's is 148 mm² versus AMD's 125 mm². The process node is identical at 28 nm, and both are fabricated by TSMC.

Clock speeds show NVIDIA ahead in both base and boost. The 940M runs at 1020 MHz base and 1098 MHz boost, while the R5 M255 runs at 925 MHz base and 940 MHz boost. Memory clocks differ as well, with NVIDIA at 900 MHz (1800 Mbps effective) and AMD at 1000 MHz (2 Gbps effective). Both have 2 GB of DDR3 memory, but the bus width differs: NVIDIA uses a 64-bit bus, AMD uses a 128-bit bus. This results in memory bandwidth of 14.40 GB/s for NVIDIA versus 32.00 GB/s for AMD.

Compute unit counts vary. NVIDIA has 512 shading units, 32 TMUs, and 16 ROPs. AMD has 384 shading units, 24 TMUs, and 8 ROPs. Pixel rate is 17.57 GPixel/s for NVIDIA and 7.520 GPixel/s for AMD. Texture rate is 35.14 GTexel/s for NVIDIA and 22.56 GTexel/s for AMD. FP32 performance is 1,124.4 GFLOPS for NVIDIA and 721.9 GFLOPS for AMD. FP16 performance is only listed for AMD at 721.9 GFLOPS (1:1). The API support differs: NVIDIA lists DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4 for NVIDIA, and DirectX 12 (12_0), OpenGL 12 (12_0 is implied), and Vulkan 1.2.170 for AMD. The bus interface is MXM-B (3.0) for NVIDIA and PCIe 3.0 x8 for AMD. Display outputs are listed as portable device dependent for NVIDIA and unlisted for AMD. TDP and slot width are listed only for NVIDIA as MXM Module with no power connectors, and are unlisted for AMD. NVIDIA's TDP is 75 W.

Where Each One Wins

NVIDIA's strengths are clear in compute-heavy workloads and API-agnostic tests that reward raw shader throughput, FP32 compute power, and pixel fill rates. The Geekbench OpenCL result is the decisive data point here, where NVIDIA wins by a 29.4% margin. The 512 shading units working at higher clocks, the 16 ROPs, and the 1,124.4 GFLOPS FP32 output all support this positioning. The NVIDIA GPU is also the better fit for scenarios that prioritize high texture detail and pixel-rate-bound rendering, as evidenced by NVIDIA's substantially higher pixel rate of 17.57 GPixel/s versus AMD's 7.520 GPixel/spton/s and 35.14 GTexel/s texture rate.

AMD's texture rate of 35.14 GTexel/s. The NVIDIA part also has a higher average benchmark score of 5284 and a higher percentile rank at 31.

AMD's R5 M255 wins in the Vulkan-specific scenario, where its 32.00 GB/s of memory bandwidth and wider 128-bit interface appear to give it a 7.6% advantage. AMD also offers support for a higher DirectX feature level of 12 (12_0), which could matter for certain modern games that can utilize that feature set. The Radeon R5 M255's nearest rival cluster includes GeForce 940MX at 4844, which is 1.2% higher than AMD's average, showing the R5 M255 sits in a similar performance band. The AMD Radeon R5 M255's raw memory bandwidth advantage over NVIDIA's 14.40 GB/s is a meaningful differentiator for workloads that are bandwidth-limited through the ROPs. The AMD Radeon R5 M255's average score is 4788, R5 M255's average score is 4788, AMD Radeon R5 M255, scoring 4925 in Vulkan, and AMD Radeon R5 M255; the R5 M255's 128-bit bus is a 7.6% lead over NVIDIA in that test, a 128-bit bus interface of 128-bit bus and 32.00 GB/s of memory bandwidth, a 32.00 GB/s memory bandwidth and 7.6% lead in Vulkan and dimensions. The AMD Radeon R5 M255, R5 M255, Radeon R5 M255, and AMD Radeon R5 M255 R5 M255, and AMD Radeon R5 M255 R5 M255. The R5 M255, R5 M255, R5 M255, AMD Radeon R5 M255 R5 M255. R5 M255. The AMD Radeon R5 M255. R5 M255. The AMD R5 M255. R5 M255. AMD R5 M255. R5 M255.