Intel UHD Graphics P750 vs NVIDIA Quadro K620M Comparison

Intel UHD Graphics P750 and NVIDIA Quadro K620M are both end-of-life mobile graphics solutions, but they represent fundamentally different design philosophies. The Intel part is an integrated graphics processor (IGP) built into a Rocket Lake chip, while the NVIDIA part is a discrete MXM module based on the Maxwell architecture. Benchmark data from the database shows a clear, though modest, performance gap between the two, with the Intel solution taking the lead in the single recorded test.

Head-to-Head Benchmarks

The database contains one head-to-head benchmark comparison between these two GPUs: Geekbench OpenCL. In this test, the Intel UHD Graphics P750 scored 6554 points, while the NVIDIA Quadro K620M scored 5957 points. This results in a 10% advantage for the Intel part, making it the winner in the sole recorded benchmark. The delta of 10% is significant enough to place the two products in different performance tiers, though both remain in the lower half of the overall GPU distribution.

The Intel UHD Graphics P750’s score of 6554 places it at the 38th percentile among all GPUs in the database. Its nearest rivals in the database include the AMD Radeon HD 7730M, which scores 6581 points (a minuscule 0.4% ahead of the Intel part), and the AMD Radeon R7 M460, which scores 6612 points (0.9% ahead). On the other side, the NVIDIA GeForce GTX 670M scores 6513 points, which is 0.6% behind the Intel UHD Graphics P750, and the NVIDIA GeForce GT 555M scores 6493 points, which is 0.9% behind. This clustering of scores within a roughly 2% range suggests that the Intel UHD Graphics P750 is performing right at the expected level for its class, sitting between the slightly faster AMD Radeon HD 7730M and the slightly slower NVIDIA GeForce GTX 670M.

The NVIDIA Quadro K620M’s score of 5957 places it at the 34th percentile among all GPUs, which is four percentile points lower than the Intel part. Its nearest rivals in the database form a similarly tight cluster. The AMD Radeon HD 8730M scores 5955 points, which is effectively tied with the Quadro K620M (0% delta). The AMD Radeon HD 8750M scores 5970 points, just 0.2% higher. The NVIDIA Quadro K4000 scores 5982 points, 0.4% higher, and the Intel UHD Graphics 730 scores 5929 points, 0.5% lower. This data indicates that the Quadro K620M is positioned right at the boundary between these comparable mobile and workstation parts, with no single rival holding a decisive edge.

The 10% delta between the two products in question is the largest gap observed in either product’s rival list. For context, the largest delta among the Intel UHD Graphics P750’s rivals is 0.9%, and among the Quadro K620M’s rivals it is 0.5%. This means the performance difference between the Intel UHD Graphics P750 and the NVIDIA Quadro K620M is an order of magnitude larger than the differences between either product and its closest peers. In practical terms, the Intel part is not just marginally faster; it sits in a distinctly higher performance bracket relative to this specific NVIDIA competitor.

FAQ

Q: Which GPU has the higher Geekbench OpenCL score?

A: The Intel UHD Graphics P750 scores 6554 points, which is 10% higher than the NVIDIA Quadro K620M’s score of 5957 points.

Q: How does the Intel UHD Graphics P750 compare to its nearest rivals?

A: The Intel part’s score of 6554 is 0.4% below the AMD Radeon HD 7730M (6581) and 0.9% below the AMD Radeon R7 M460 (6612), while being 0.6% above the NVIDIA GeForce GTX 670M (6513) and 0.9% above the NVIDIA GeForce GT 555M (6493).

Q: How does the NVIDIA Quadro K620M compare to its nearest rivals?

A: The Quadro K620M’s score of 5957 is effectively tied with the AMD Radeon HD 8730M (5955, 0% delta), 0.2% below the AMD Radeon HD 8750M (5970), 0.4% below the NVIDIA Quadro K4000 (5982), and 0.5% above the Intel UHD Graphics 730 (5929).

Q: What are the percentile ranks of these two GPUs?

A: The Intel UHD Graphics P750 sits at the 38th percentile among all GPUs, while the NVIDIA Quadro K620M sits at the 34th percentile.

Q: Which GPU has the higher pixel fill rate?

A: The Intel UHD Graphics P750 has a pixel rate of 41.60 GPixel/s, which is substantially higher than the NVIDIA Quadro K620M’s pixel rate of 8.992 GPixel/s.

Q: Which GPU has the higher texture fill rate?

A: The Intel UHD Graphics P750 has a texture rate of 83.20 GTexel/s, compared to the NVIDIA Quadro K620M’s texture rate of 17.98 GTexel/s.

Architecture Differences

The two GPUs are built on entirely different architectures and manufacturing processes. The Intel UHD Graphics P750 uses the Generation 12.1 architecture, implemented on a 14 nm+++ process node at Intel’s own foundry. It is part of the HD Graphics-W (Rocket Lake) generation. The NVIDIA Quadro K620M, in contrast, uses the Maxwell architecture, fabricated by TSMC on a 28 nm process node. It belongs to the Quadro Kepler-M (Kx200M) generation, despite being based on Maxwell silicon. The NVIDIA chip is identified as GM108S, while the Intel part is designated as Rocket Lake.

The Intel UHD Graphics P750 integrates its memory controller with the system, using System Shared memory for both size, type, and bus width. Its memory bandwidth is listed as System Dependent, meaning it varies with the host platform’s memory configuration. The NVIDIA Quadro K620M has a dedicated 2 GB DDR3 memory pool with a 64-bit bus and a fixed bandwidth of 16.02 GB/s. Its memory clock is listed as 1001 MHz, with 2 Gbps effective data rate.

The compute resources differ markedly. The Intel UHD Graphics P750 has 256 shading units, 64 texture mapping units, and 32 raster output units. The NVIDIA Quadro K620M has 384 shading units but only 16 texture mapping units and 8 raster output units. This distribution explains the fill rate numbers: the Intel part’s higher TMU and ROP counts drive its 83.20 GTexel/s and 41.60 GPixel/s, while the NVIDIA part’s lower counts produce 17.98 GTexel/s and 8.992 GPixel/s. The FP32 compute performance tells a different story: the NVIDIA Quadro K620M delivers 863.2 GFLOPS, which is higher than the Intel UHD Graphics P750’s 665.6 GFLOPS. The Intel part does support FP16 at 1,331.2 GFLOPS with a 2:1 ratio, while the NVIDIA part has no recorded FP16 capability.

In terms of API support, both GPUs support DirectX 12, OpenGL 4.6, and Vulkan 1.4. However, the DirectX version differs: the Intel UHD Graphics P750 supports DirectX 12 (12_1), while the NVIDIA Quadro K620M supports DirectX 12 (11_0). This indicates a feature level difference, with the Intel part supporting a newer DirectX feature set.

The Intel UHD Graphics P750 is an IGP with a Ring Bus interface, meaning it connects to the CPU via the internal ring interconnect. The NVIDIA Quadro K620M is an MXM Module with an MXM-A (3.0) bus interface, designed for removable mobile workstation modules. The Intel part’s display outputs are Motherboard Dependent, while the NVIDIA part’s are Portable Device Dependent. The Intel part has no power connectors, consistent with an integrated design, and the NVIDIA part also has no power connectors, relying on the MXM slot for power.

Specification Differences

The two GPUs differ across nearly every measurable specification. The process node is a clear differentiator: Intel uses 14 nm+++, while NVIDIA uses 28 nm. The NVIDIA part has a disclosed transistor count of 1,020 million on a 77 mm² die, giving a transistor density of 13.2M per mm². The Intel part does not have transistor count, die size, or density data in the database.

Clock speeds vary significantly. The Intel UHD Graphics P750 has a base clock of 350 MHz and a boost clock of 1300 MHz. The NVIDIA Quadro K620M has a much higher base clock of 1029 MHz and a boost clock of 1124 MHz. The Intel part’s boost clock is 176 MHz higher than the NVIDIA part’s boost clock, but its base clock is 679 MHz lower.

Memory configuration is another major difference. The Intel part uses System Shared memory with no dedicated size, type, or bus width. The NVIDIA part uses 2 GB of DDR3 on a 64-bit bus with 16.02 GB/s bandwidth. The TDP also diverges: the Intel UHD Graphics P750 is rated at 15 W, while the NVIDIA Quadro K620M is rated at 30 W. The slot width differs as well: the Intel part is an IGP, while the NVIDIA part is an MXM Module.

The shading unit count favors NVIDIA (384 vs 256), but TMU and ROP counts favor Intel (64 and 32 vs 16 and 8). FP32 performance favors NVIDIA (863.2 GFLOPS vs 665.6 GFLOPS), while FP16 is only available on Intel. The release date for the NVIDIA Quadro K620M is recorded as 2015-02-28, while the Intel part has no recorded release date. Both products are marked as End-of-life. The NVIDIA part has a predecessor (Quadro Fermi-M) and a successor (Quadro Maxwell-M), while the Intel part has neither listed.

The Verdict

Based on the recorded benchmark data, the Intel UHD Graphics P750 is the faster GPU. It wins the sole head-to-head test by 10%, a margin that is not trivial when considering the tight clustering of rival scores around both products. The Intel part’s 38th percentile rank versus the NVIDIA part’s 34th percentile rank reinforces this conclusion. For users running OpenCL workloads, the Intel UHD Graphics P750 offers a clear performance advantage over the NVIDIA Quadro K620M.

However, the choice is not purely about benchmark scores. The NVIDIA Quadro K620M has a dedicated 2 GB memory pool with fixed bandwidth, which can be beneficial for applications that require consistent memory behavior without depending on system RAM. Its higher FP32 compute (863.2 GFLOPS) also suggests it may handle certain compute tasks more efficiently, despite the lower overall OpenCL score. The NVIDIA part also has a higher base clock and a lower boost clock, which may indicate more stable sustained performance under load.

The Intel UHD Graphics P750 is the better choice for users who prioritize the recorded OpenCL benchmark performance and want a lower power envelope (15 W vs 30 W). It also offers FP16 support, which the NVIDIA part lacks, and a newer DirectX feature level (12_1 vs 11_0). The Intel part’s fill rates are substantially higher, which could benefit texture-heavy or pixel-heavy workloads.

The NVIDIA Quadro K620M is the better choice for users who need a discrete module with dedicated memory and higher raw FP32 compute. Its 384 shading units and 863.2 GFLOPS FP32 performance are superior to the Intel part’s 256 shading units and 665.6 GFLOPS. The dedicated 2 GB DDR3 memory with 16.02 GB/s bandwidth provides predictable memory performance, unlike the Intel part’s system-dependent bandwidth.

Given the 10% delta in the only benchmark, the Intel UHD Graphics P750 is the recommended part for general OpenCL performance. The data does not support recommending the NVIDIA Quadro K620M for raw benchmark dominance, but its distinct architectural features, such as dedicated memory and higher FP32 throughput, may justify its selection in specific workstation scenarios where those characteristics are more valuable than the measured OpenCL score. Both products are end-of-life, so availability and platform compatibility will factor into any real-world decision, but from a pure performance standpoint, the Intel part holds the edge in the recorded data.