GPU Comparison

The NVIDIA GeForce GT 735M and NVIDIA GeForce 920M are both end-of-life mobile graphics solutions built on the same Kepler 2.0 architecture. Despite their generational separation, the data reveals a surprisingly narrow performance gap, with the newer 920M holding a consistent, though modest, lead in the available benchmarks. The GT 735M appears in the GeForce 700M generation, while the 920M belongs to the GeForce 900M series, but both are pinned to the same 28 nm TSMC process node.

Where Each One Wins

The benchmark data presents a clear, single-test picture: the NVIDIA GeForce 920M wins the only head-to-head comparison available. In the Geekbench OpenCL test, the 920M scores 3725 against the GT 735M’s 3616, a delta of -2.9% from the perspective of the older card. This means the 920M is approximately 3% faster in this compute-oriented workload. There are zero wins recorded for the GT 735M across the head-to-head benchmarks, underscoring that the newer chip’s advantage is not a fluke of a single metric but a consistent outcome in the data provided.

The two cards are otherwise closely matched in their positioning. The GT 735M holds a 21st percentile rank among all GPUs, while the 920M sits at the 20th percentile. This means both are entry-level parts, but the data suggests the 920M is slightly less capable relative to the broader GPU landscape, even though it beats the 735M directly. The 920M’s average benchmark score across all its tests is 3287, which is dragged down by its additional Vulkan result of 2849, a workload where it performs significantly worse than in OpenCL. The GT 735M’s average score is simply its single OpenCL result of 3616.

For a user choosing between these two, the use case is straightforward: the 920M is the better pick for OpenCL-accelerated tasks, which often include video encoding, physics simulations, and certain compute offloads. The 735M, however, has no test where it outperforms the 920M, so it does not win any specific use-case category in this dataset. The only nuance is that the 920M’s Vulkan score is notably lower, but since the 735M has no Vulkan result recorded, a direct comparison there is impossible.

FAQ

Q: Which GPU is faster in the Geekbench OpenCL benchmark?

A: The NVIDIA GeForce 920M scores 3725, while the NVIDIA GeForce GT 735M scores 3616. The 920M wins this test with a delta of -2.9% relative to the 735M.

Q: How do the two cards compare in terms of memory bandwidth?

A: Both GPUs feature 2 GB of DDR3 memory on a 64-bit bus, providing identical bandwidth of 14.40 GB/s. There is no difference in this specification.

Q: What is the transistor count for each chip?

A: Both the GT 735M’s GK208 chip and the 920M’s GK208B chip contain exactly 1,020 million transistors on an 87 mm² die, yielding the same transistor density of 11.7M / mm².

Q: Do these GPUs support the same graphics APIs?

A: Yes, both list DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175 as supported APIs. The feature set is identical in this regard.

Q: Which card has a higher boost clock?

A: The GeForce 920M has a boost clock of 954 MHz, which is also its base clock. The GT 735M has a boost clock of 628 MHz, significantly lower than the 920M’s.

Q: What is the percentile ranking difference between the two?

A: The GT 735M ranks in the 21st percentile of all GPUs, while the 920M ranks in the 20th percentile. Despite the 920M winning the direct comparison, it sits one percentile lower in the global distribution.

Head-to-Head Benchmarks

The only head-to-head benchmark provided is the Geekbench OpenCL test, and it tells a clear story of generational improvement within the same architectural family. The GeForce 920M produces a score of 3725, edging out the GT 735M’s 3616. The delta of -2.9% indicates that the 735M trails by roughly 3 percentage points. While this is not a massive gap, it is consistent with the 920M’s higher clock speeds, which will be detailed in the specification section. In absolute terms, the 920M is 109 points ahead in this synthetic compute workload.

There are no other benchmark results to compare across the two cards. The 920M has an additional Geekbench Vulkan score of 2849, but no equivalent Vulkan result exists for the GT 735M. This absence means the dataset cannot confirm whether the 735M would outperform the 920M in Vulkan, especially given the 920M’s relatively weak Vulkan showing compared to its own OpenCL result. The wins tally is 0 for the GT 735M and 1 for the 920M, making the outcome unambiguous within the provided data.

For context, the nearest rivals for each card show how they sit in the broader market. The GT 735M’s average score of 3616 places it near the NVIDIA GeForce GTX 1050 (3629, -0.3%) and the NVIDIA RTX 5000 Mobile Ada Generation (3596, 0.6%), indicating that benchmark scores in this range are crowded. The 920M’s average of 3287 is lower, but its OpenCL score alone is competitive with the GT 735M’s. Its nearest rivals include the NVIDIA GeForce GT 730M (3316, -0.9%) and Intel HD Graphics 530 (3332, -1.4%), showing that the 920M’s average positions it against older discrete parts and modern integrated graphics.

Specification Differences

The raw numbers show that the GeForce 920M is essentially a higher-clocked revision of the GT 735M, with all other core specifications remaining identical. The most significant difference is in clock speeds: the GT 735M runs at a base clock of 575 MHz with a boost of 628 MHz, while the 920M operates at a flat 954 MHz for both base and boost. This represents a 66% increase in base clock and a 52% increase in boost clock, directly driving higher compute throughput.

The consequences of these clock changes are visible in the derived performance metrics. The GT 735M achieves a pixel rate of 5.024 GPixel/s and a texture rate of 20.10 GTexel/s, while the 920M reaches 7.632 GPixel/s and 30.53 GTexel/s, respectively. Similarly, FP32 compute rises from 482.3 GFLOPS on the 735M to 732.7 GFLOPS on the 920M. These are the only areas where the two cards differ in the specification sheet. Memory size (2 GB), type (DDR3), bus width (64 bit), bandwidth (14.40 GB/s), shading units (384), TMUs (32), ROPs (8), TDP (33 W), and bus interface (PCIe 3.0 x8) are all identical.

The chip designation also differs, with the GT 735M using the GK208 and the 920M using the GK208B. This is a minor revision, but the underlying transistor count and die size remain the same at 1,020 million and 87 mm², respectively. The process node is also unchanged at 28 nm from TSMC. The release dates are different, with the 735M launching in 2013 and the 920M in 2015, but this does not affect the measured specifications.

Architecture Differences

Both GPUs are built on the Kepler 2.0 architecture, which means the fundamental design philosophy is shared. The shading unit count is identical at 384, as are the 32 texture mapping units and 8 render output units. Neither card supports ray tracing or tensor cores, and both lack FP16 compute capabilities. The API support is also identical, with DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175 listed for both.

The primary architectural difference is the chip revision: GK208 versus GK208B. The data does not specify what changes the "B" suffix entails, but the clock behavior suggests a more mature or refined process bin that allows for higher operating frequencies at the same 33 W TDP. The fact that the 920M’s base clock equals its boost clock at 954 MHz indicates a stable, fixed operating point, whereas the 735M’s boost range from 575 MHz to 628 MHz shows a more variable clocking scheme.

In terms of feature sets, the two are functionally identical. Both have the same transistor density of 11.7M / mm², confirming they are cut from the same architectural cloth. The generation label differs, GeForce 700M versus GeForce 900M, but this appears to be a marketing distinction rather than a technical one, given the shared node, transistor count, and memory subsystem. The predecessor and successor chains differ, with the 735M following the GeForce 600M and preceding the GeForce 800M, while the 920M follows the GeForce 800M and precedes the GeForce 10 Mobile, but this does not alter the on-paper capabilities.

The Verdict

The data makes the choice straightforward for anyone comparing these two specific SKUs. The NVIDIA GeForce 920M is the superior performer in the only benchmark where both are measured, winning the Geekbench OpenCL test by 2.9% with a score of 3725 versus 3616. The 920M also offers substantially higher clock speeds, with a 954 MHz base and boost against the 735M’s 575 MHz base and 628 MHz boost, which translates into a 52% higher pixel rate and a 52% higher texture rate. For compute-heavy tasks, the 920M’s FP32 output of 732.7 GFLOPS is meaningfully ahead of the 735M’s 482.3 GFLOPS.

However, the choice is not universally in favor of the 920M when considering the broader context. The 920M’s average benchmark score across all its tests is 3287, which is lower than the GT 735M’s average of 3616, due to the 920M’s weak Vulkan score of 2849. If a workload relies heavily on Vulkan, the 920M may underperform its OpenCL showing, and the lack of a Vulkan result for the 735M leaves this comparison unresolved. Furthermore, the 920M ranks in the 20th percentile of all GPUs, one spot below the 735M’s 21st percentile, suggesting that the newer card does not elevate the user experience relative to the entire GPU market.

For a user prioritizing raw OpenCL compute and higher clock stability, the 920M is the clear pick. For a user who values a more consistent average performance across heterogeneous workloads, the 735M’s single-score profile is less volatile, though it lacks the 920M’s peak performance. Both cards share identical memory bandwidth, TDP, and API support, so there is no advantage in those areas. The verdict from the data: the 920M wins on peak performance and clock speed, but the 735M’s percentile ranking and lack of a weak Vulkan score make it a more balanced, if slower, option. The only benchmark result where the 735M does not lose is one that does not exist for it.