Intel Iris Pro Graphics 5200 vs NVIDIA GeForce GT 735M Comparison

Intel Iris Pro Graphics 5200 and NVIDIA GeForce GT 735M are both end-of-life mobile graphics solutions from the 2013 era, but they approach their tasks from fundamentally different directions. The Intel part is an integrated processor graphics solution with a large embedded DRAM cache, while the NVIDIA part is a discrete mobile GPU built on a smaller process node. Benchmark data in the database shows a clear overall winner, but the specifications reveal why each part exists and where one might be preferable in specific scenarios.

Where Each One Wins

The recorded data shows a single head-to-head benchmark result, and it is decisive. In the Geekbench OpenCL test, the Intel Iris Pro Graphics 5200 scores 5042 against the NVIDIA GeForce GT 735M’s 3616. That is a 39.4% advantage for the Intel part. This is not a marginal win; it is a substantial performance gap that places the two parts in different performance tiers despite their similar era and target market.

The Intel part wins the only direct comparison available, so the use-case split is largely defined by what each part offers beyond raw compute. The GT 735M has a dedicated 2 GB memory pool with a 64-bit bus and 14.40 GB/s bandwidth. The Iris Pro relies on system shared memory, with bandwidth described as “System Dependent.” In a laptop with fast dual-channel DDR3, the Iris Pro could approach or exceed the GT 735M’s bandwidth, but in a system with slower memory, the GT 735M’s dedicated VRAM would be more consistent. For applications that are sensitive to memory latency or that need to stream textures without competing with the CPU for bandwidth, the GT 735M’s dedicated memory is a structural advantage.

Conversely, the Iris Pro wins on raw compute throughput. Its FP32 rate is 736.0 GFLOPS versus 482.3 GFLOPS for the GT 735M. That is roughly 52.6% higher theoretical floating-point performance. For OpenCL workloads, which often scale with FP32 throughput, the Iris Pro is the clear choice. The GT 735M has more shading units (384 vs 320) but runs at much lower clocks (575 MHz base, 628 MHz boost vs 200 MHz base, 1150 MHz boost for the Intel part). The Intel part’s high boost clock and larger texture rate (46.00 GTexel/s vs 20.10 GTexel/s) make it better suited for compute-heavy tasks and texture-bound scenarios.

The GT 735M wins in pixel rate. Its 5.024 GPixel/s exceeds the Intel part’s 4.600 GPixel/s by about 9.2%. For fill-rate-bound tasks, such as simple 2D rendering or certain older games at low resolutions, the GT 735M has a slight edge. Also, the GT 735M supports Vulkan 1.2.175 and OpenGL 4.6, while the Iris Pro supports Vulkan 1.0 and OpenGL 4.3. For modern applications that rely on newer API features, the GT 735M is more future-proof.

Architecture Differences

The Intel Iris Pro Graphics 5200 is built on the Haswell GT3e chip using Intel’s 22 nm process. Its architecture is Generation 7.5, part of the HD Graphics (Haswell) family. The key differentiator is the embedded DRAM (eDRAM) that the GT3e variant includes, which provides a large cache for the graphics core. This is why the Intel part achieves high texture and compute rates despite being an IGP. The bus interface is a Ring Bus, which ties it directly to the CPU’s memory controller, but the eDRAM helps mitigate the latency and bandwidth limitations of system shared memory.

The NVIDIA GeForce GT 735M is based on the GK208 chip, built on a 28 nm process at TSMC. This is a Kepler 2.0 architecture, part of the GeForce 700M generation. The chip contains 1,020 million transistors on an 87 mm² die, giving a transistor density of 11.7M per mm². The GT 735M uses a PCIe 3.0 x8 bus interface, which is a standard discrete GPU connection. It has no power connectors, drawing all power from the motherboard slot, which is typical for a low-power mobile GPU.

The transistor counts are not directly comparable because the Iris Pro’s transistor count is not listed in the database. However, the die size difference is notable: the Intel part is an IGP integrated into the CPU die, while the GT 735M is a separate chip with its own memory bus. The GT 735M’s dedicated memory controller and 2 GB DDR3 frame buffer are fundamental architectural differences. The Iris Pro’s memory is system shared, meaning the CPU and GPU share the same memory pool, which can cause contention.

The shading units are 320 for the Intel part and 384 for the NVIDIA part. Despite having fewer units, the Intel part achieves higher FP32 throughput because its boost clock is nearly double the NVIDIA part’s boost clock (1150 MHz vs 628 MHz). The Intel part also has more TMUs (40 vs 32) but fewer ROPs (4 vs 8). This is a classic trade-off: the Intel part is optimized for texture-heavy compute, while the NVIDIA part is optimized for pixel output.

Head-to-Head Benchmarks

The only head-to-head benchmark in the database is Geekbench OpenCL. The Intel Iris Pro Graphics 5200 scores 5042, and the NVIDIA GeForce GT 735M scores 3616. The Intel part wins by 39.4%. This is a large margin in the context of GPU benchmarks. To put it in perspective, the Intel part’s average benchmark score is 4360, while the GT 735M’s average is 3616. The Intel part’s percentile rank among all GPUs is 26, versus 21 for the GT 735M. That means the Intel part sits in a higher performance bracket overall.

The Geekbench OpenCL test is compute-intensive, often involving floating-point operations, memory access patterns, and parallel execution. The Intel part’s higher FP32 rate (736.0 GFLOPS vs 482.3 GFLOPS) and higher texture rate (46.00 GTexel/s vs 20.10 GTexel/s) directly contribute to its superior score. The GT 735M’s higher pixel rate (5.024 GPixel/s vs 4.600 GPixel/s) is less relevant in this particular workload, which is why it loses so decisively.

There is no Geekbench Vulkan score for the GT 735M in the database, but the Intel part records a Vulkan score of 3677. This indicates the Intel part can handle Vulkan workloads, though with an older API version (1.0). The GT 735M supports a much newer Vulkan version (1.2.175), which could allow it to run newer Vulkan applications that the Intel part cannot, but its raw performance in OpenCL suggests it would still be slower in compute-bound Vulkan tasks.

FAQ

Q: Which GPU has higher peak FP32 performance?

A: The Intel Iris Pro Graphics 5200 has a peak FP32 rate of 736.0 GFLOPS, compared to 482.3 GFLOPS for the NVIDIA GeForce GT 735M. This is a 52.6% advantage for the Intel part.

Q: Does the NVIDIA GPU have more video memory?

A: Yes. The GT 735M has 2 GB of dedicated DDR3 memory with a 64-bit bus and 14.40 GB/s bandwidth. The Iris Pro uses system shared memory, with bandwidth described as “System Dependent.”

Q: Which GPU supports newer graphics APIs?

A: The NVIDIA GeForce GT 735M supports Vulkan 1.2.175 and OpenGL 4.6. The Intel Iris Pro Graphics 5200 supports Vulkan 1.0 and OpenGL 4.3. Both support DirectX 12, but the Intel part is limited to feature level 11_1, while the NVIDIA part supports 11_0.

Q: What is the GT 735M’s transistor count and die size?

A: The GT 735M contains 1,020 million transistors on an 87 mm² die, with a transistor density of 11.7M per mm². This data is not listed for the Iris Pro.

Q: How does the average benchmark score compare?

A: The Intel Iris Pro has an average benchmark score of 4360, while the GT 735M has an average of 3616. The Intel part’s percentile rank is 26, versus 21 for the NVIDIA part.

Q: Which GPU has a higher base clock?

A: The NVIDIA GT 735M has a base clock of 575 MHz, while the Intel Iris Pro has a base clock of 200 MHz. However, the Intel part’s boost clock is 1150 MHz, much higher than the GT 735M’s 628 MHz boost.

Specification Differences

The two GPUs differ in nearly every key specification. The process node is 22 nm for Intel and 28 nm for NVIDIA. The Intel part is fabricated by Intel, the NVIDIA part by TSMC. The GT 735M has a listed transistor count of 1,020 million and a die size of 87 mm², while the Iris Pro has no such figures in the database.

Clock speeds: Intel base is 200 MHz, boost is 1150 MHz. NVIDIA base is 575 MHz, boost is 628 MHz. The memory clock for the GT 735M is 900 MHz (1800 Mbps effective), while the Iris Pro’s memory clock is listed as “System Shared.”

Memory: The GT 735M has 2 GB DDR3 on a 64-bit bus with 14.40 GB/s bandwidth. The Iris Pro has system shared memory with system-dependent bandwidth.

Compute units: Intel has 320 shading units, 40 TMUs, and 4 ROPs. NVIDIA has 384 shading units, 32 TMUs, and 8 ROPs. Pixel rate: Intel 4.600 GPixel/s, NVIDIA 5.024 GPixel/s. Texture rate: Intel 46.00 GTexel/s, NVIDIA 20.10 GTexel/s. FP32: Intel 736.0 GFLOPS, NVIDIA 482.3 GFLOPS.

TDP: Intel is 45 W, NVIDIA is 33 W. The GT 735M has no power connectors, while the Iris Pro is an IGP with no separate power requirement listed. Bus interface: Intel uses Ring Bus, NVIDIA uses PCIe 3.0 x8. Display outputs: Intel is motherboard dependent, NVIDIA is portable device dependent.

APIs: Intel supports DirectX 12 (11_1), OpenGL 4.3, Vulkan 1.0. NVIDIA supports DirectX 12 (11_0), OpenGL 4.6, Vulkan 1.2.175. Release dates: Intel released on 2013-06-02, NVIDIA on 2013-03-31. The GT 735M has a predecessor (GeForce 600M) and successor (GeForce 800M), while the Iris Pro has none listed.

The Verdict

The data is unambiguous in raw compute: the Intel Iris Pro Graphics 5200 outperforms the NVIDIA GeForce GT 735M by 39.4% in Geekbench OpenCL, and its FP32 rate is 52.6% higher. For any workload that is compute-bound, such as OpenCL acceleration, video encoding, or GPGPU tasks, the Intel part is the better choice. Its higher texture rate (46.00 GTexel/s) and boost clock (1150 MHz) give it a clear edge in parallel throughput.

The NVIDIA GeForce GT 735M has its own strengths. Its pixel rate is 9.2% higher (5.024 GPixel/s vs 4.600 GPixel/s), and it has twice the ROPs (8 vs 4). For fill-rate-bound scenarios, such as older games at low resolutions or simple 2D compositing, the GT 735M could be more responsive. Its dedicated 2 GB memory with a fixed 14.40 GB/s bandwidth provides more predictable performance than the Iris Pro’s system-dependent shared memory. The GT 735M also supports newer API versions, with Vulkan 1.2.175 and OpenGL 4.6, which may be necessary for certain modern software.

The TDP difference is notable: the GT 735M consumes 33 W versus the Intel part’s 45 W. In a laptop, this lower power draw could translate to longer battery life and less heat, though the Intel part’s integration into the CPU package means it does not require separate cooling for a discrete chip.

The percentile ranks tell the story: the Iris Pro sits at the 26th percentile of all GPUs, the GT 735M at the 21st. Both are low-tier by modern standards, but the Intel part is demonstrably faster in the most common compute benchmark. The GT 735M’s nearest rivals include the GeForce GTX 1050 with an average score of 3629, essentially tied with the GT 735M at 3616. The Iris Pro’s nearest rivals include the GeForce RTX 4070 GDDR6 at 4335, which is only 0.6% slower than the Iris Pro’s 4360 average. This places the Iris Pro in an odd position: it competes with modern midrange GPUs in synthetic compute, despite being an end-of-life integrated solution.

For a user who needs raw compute performance in a power-constrained environment, the Intel Iris Pro Graphics 5200 is the better pick. For a user who needs a dedicated memory pool, higher pixel throughput, and newer API support, the NVIDIA GeForce GT 735M has its place. The database shows that in the one direct comparison, the Intel part wins decisively, and that is the strongest evidence available for choosing it over the GT 735M.