Intel Iris Pro Graphics P580 vs NVIDIA Quadro K620 Comparison

Head-to-Head Benchmarks

The recorded data shows a clear split between the two parts, with each winning one of the two benchmark tests. In Geekbench OpenCL, the Intel Iris Pro Graphics P580 posts a score of 9,082 against 6,693 for the NVIDIA Quadro K620, a lead of 35.7%. That is the largest margin in either direction across the entire comparison. The OpenCL result places the Intel part firmly ahead in compute workloads that scale across its execution resources.

The Vulkan test tells a different story. The NVIDIA Quadro K620 scores 5,870, while the Intel Iris Pro Graphics P580 scores 5,258. That gives the NVIDIA part a 10.4% advantage in Vulkan. This is a narrower margin than the OpenCL gap, but it is still a meaningful reversal. The data indicates that the two products do not simply rank one above the other across all APIs; each has a domain where it leads.

Looking at the broader benchmark averages, the Intel part records an average score of 7,170, while the NVIDIA part records 6,282. That is a difference of roughly 14% in favor of Intel when averaging across the two recorded tests. However, the average masks the fact that the Vulkan result goes the other way. A user who relies primarily on Vulkan-based applications would see the NVIDIA part as the faster option, while a user whose workload is dominated by OpenCL would see the opposite.

The percentile data reinforces the overall positioning. The Intel Iris Pro Graphics P580 sits at the 39th percentile among all GPUs in the database, while the NVIDIA Quadro K620 sits at the 36th percentile. The three-percentile gap is modest, and both parts land in the lower-middle range of the overall GPU distribution. Neither product is a high-end part by contemporary standards, but the Intel part edges ahead in aggregate standing.

Nearest rival comparisons provide additional context for each part. The Intel Iris Pro Graphics P580 has an average score of 7,170, which is essentially tied with the NVIDIA GeForce GTX 560 SE at 7,171 (zero percent delta) and the NVIDIA GeForce GTX 970 at 7,157 (0.2% delta). It also sits 0.5% ahead of the AMD Radeon Vega 8 Mobile and 0.7% behind the NVIDIA GeForce GTX 750. These are all near-ties, meaning the Intel part performs at a level comparable to a spread of older discrete and integrated GPUs.

The NVIDIA Quadro K620, with an average score of 6,282, is within 0.2% of both the NVIDIA GeForce RTX 5070 Ti SUPER and the NVIDIA GeForce RTX 4070 Ti SUPER AD102, both at 6,270. It is 0.7% behind the AMD Radeon R7 M350 and 0.8% behind the AMD Radeon Pro WX 4100. The proximity to those two RTX parts is striking, though it likely reflects the specific benchmark workload rather than general gaming performance. In the database's recorded metrics, the Quadro K620 lands in a cluster of parts with near-identical average scores.

The Verdict

From the data, the choice depends entirely on which API matters more. The Intel Iris Pro Graphics P580 wins the OpenCL test by a decisive 35.7% margin. The NVIDIA Quadro K620 wins the Vulkan test by a more modest 10.4%. If the workload is OpenCL-heavy, the Intel part is the clear pick. If the workload is Vulkan-heavy, the NVIDIA part takes the lead.

The average benchmark score favors Intel, 7,170 versus 6,282, and the percentile ranking favors Intel as well, 39th versus 36th. That makes the Intel part the stronger choice in aggregate. However, the NVIDIA part has the advantage of being a single-slot discrete card with its own 2 GB of DDR3 memory and a dedicated 28.80 GB/s memory bus, while the Intel part relies on system shared memory with bandwidth marked as system dependent. For users who need predictable memory performance and a dedicated frame buffer, the NVIDIA part is the safer option, even if its average score is lower.

A user who wants maximum compute throughput in OpenCL should choose the Intel Iris Pro Graphics P580. A user who prioritizes Vulkan performance or needs a dedicated discrete card with its own memory should choose the NVIDIA Quadro K620. The data does not support a universal winner. It supports two conditional winners.

Architecture Differences

The two parts come from different manufacturers and use fundamentally different designs. The Intel Iris Pro Graphics P580 is built on the Skylake GT4e chip, part of Intel's Generation 9.0 architecture, and is fabricated on a 14 nm+ process at Intel's own foundry. The NVIDIA Quadro K620 uses the GM107 chip, built on NVIDIA's Maxwell architecture, and is fabricated on a 28 nm process at TSMC. The process node difference is significant: 14 nm+ versus 28 nm means Intel's part uses a much newer manufacturing process.

The Intel part is an integrated graphics processor, with a slot width listed as IGP and a bus interface of Ring Bus. It has no dedicated memory; both memory size and type are marked as System Shared, and its bandwidth is listed as System Dependent. The NVIDIA part is a discrete single-slot card, 160 mm long and 69 mm high, with a PCIe 2.0 x16 interface. It carries 2 GB of DDR3 memory on a 128-bit bus, delivering 28.80 GB/s of bandwidth. This is a fundamental difference in how the two parts access data.

The Intel part has 576 shading units, 72 texture mapping units, and 9 raster output units. The NVIDIA part has 384 shading units, 24 texture mapping units, and 16 raster output units. Intel has more shading units and far more TMUs, while NVIDIA has more ROPs. The pixel rate reflects this: NVIDIA delivers 17.98 GPixel/s versus 9.00 GPixel/s for Intel. The texture rate goes the other way: Intel delivers 72.00 GTexel/s versus 26.98 GTexel/s for NVIDIA. These are complementary strengths, with Intel favoring texture-heavy compute and NVIDIA favoring pixel throughput.

Clock speeds also differ. The Intel part has a base clock of 350 MHz and a boost clock of 1000 MHz. The NVIDIA part has a base clock of 1058 MHz and a boost clock of 1124 MHz. NVIDIA runs at much higher clocks, which helps explain its higher pixel rate despite fewer ROPs. Intel compensates with more shading units and TMUs, which explains its FP32 advantage. The Intel part delivers 1,152.0 GFLOPS of FP32 performance, while the NVIDIA part delivers 863.2 GFLOPS. Intel also lists FP16 performance of 2.304 TFLOPS with a 2:1 ratio; NVIDIA does not list FP16 at all.

The Intel part has a TDP of 15 W, while the NVIDIA part has a TDP of 45 W. The NVIDIA part also lists a suggested PSU of 200 W, though it requires no power connectors. The Intel part, as an IGP, draws much less power. The NVIDIA part is the more power-hungry of the two, but neither is a high-consumption part.

API support differs as well. The Intel part supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3. The NVIDIA part supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. The DirectX feature level is higher on the Intel part, while the Vulkan version is higher on the NVIDIA part. Both support OpenGL 4.6.

The NVIDIA part has a known transistor count of 1,870 million on a 148 mm² die, with a transistor density of 12.6 million per square millimeter. The Intel part does not list transistor count, die size, or density in the database. The NVIDIA part also lists a predecessor (Quadro Fermi) and a successor (Quadro Maxwell), while the Intel part lists neither.

FAQ

Q: Which GPU has the higher OpenCL score?

A: The Intel Iris Pro Graphics P580 scores 9,082 in Geekbench OpenCL, which is 35.7% higher than the NVIDIA Quadro K620's 6,693.

Q: Which GPU has the higher Vulkan score?

A: The NVIDIA Quadro K620 scores 5,870 in Geekbench Vulkan, which is 10.4% higher than the Intel Iris Pro Graphics P580's 5,258.

Q: How much memory does each GPU have?

A: The NVIDIA Quadro K620 has 2 GB of DDR3 memory on a 128-bit bus with 28.80 GB/s bandwidth. The Intel Iris Pro Graphics P580 uses system shared memory, with bandwidth listed as system dependent.

Q: What are the TDP ratings?

A: The Intel Iris Pro Graphics P580 has a TDP of 15 W. The NVIDIA Quadro K620 has a TDP of 45 W and lists a suggested PSU of 200 W.

Q: Which GPU has more shading units?

A: The Intel Iris Pro Graphics P580 has 576 shading units, while the NVIDIA Quadro K620 has 384 shading units.

Q: Which GPU has a higher pixel rate?

A: The NVIDIA Quadro K620 has a pixel rate of 17.98 GPixel/s, compared to 9.00 GPixel/s for the Intel Iris Pro Graphics P580.

Where Each One Wins

The Intel Iris Pro Graphics P580 wins in compute-heavy OpenCL workloads. Its OpenCL score of 9,082 is 35.7% ahead of the NVIDIA part, and its FP32 throughput of 1,152.0 GFLOPS is substantially higher than the NVIDIA part's 863.2 GFLOPS. It also has far more texture mapping units, 72 versus 24, and a much higher texture rate of 72.00 GTexel/s. For tasks that stress shading units, TMUs, and FP32 compute, the Intel part is the better performer. Its 15 W TDP also makes it far more power-efficient, and as an IGP it requires no additional power connectors or slot space.

The NVIDIA Quadro K620 wins in Vulkan workloads. Its Vulkan score of 5,870 is 10.4% ahead of the Intel part, and it supports Vulkan 1.4 versus Vulkan 1.3 on the Intel part. It also delivers a much higher pixel rate of 17.98 GPixel/s, more than double the Intel part's 9.00 GPixel/s, thanks to its 16 ROPs and higher clock speeds. The NVIDIA part has dedicated 2 GB DDR3 memory with a fixed 28.80 GB/s bandwidth, which is a clear advantage for applications that need consistent memory performance rather than relying on system shared memory. It is also a discrete single-slot card with its own cooling and power delivery, making it suitable for systems that cannot rely on integrated graphics.

Specification Differences

The two parts differ across nearly every specification category. The Intel Iris Pro Graphics P580 is built on a 14 nm+ process at Intel, while the NVIDIA Quadro K620 is built on a 28 nm process at TSMC. The Intel part uses the Skylake GT4e chip with Generation 9.0 architecture; the NVIDIA part uses the GM107 chip with Maxwell architecture.

Clock speeds differ significantly. The Intel part has a base clock of 350 MHz and a boost clock of 1000 MHz. The NVIDIA part has a base clock of 1058 MHz and a boost clock of 1124 MHz. Memory also differs: the Intel part uses system shared memory with system-dependent bandwidth, while the NVIDIA part has 2 GB of DDR3 on a 128-bit bus with 28.80 GB/s bandwidth.

Core counts differ. The Intel part has 576 shading units, 72 TMUs, and 9 ROPs. The NVIDIA part has 384 shading units, 24 TMUs, and 16 ROPs. Pixel rate favors NVIDIA at 17.98 GPixel/s versus 9.00 GPixel/s. Texture rate favors Intel at 72.00 GTexel/s versus 26.98 GTexel/s. FP32 performance favors Intel at 1,152.0 GFLOPS versus 863.2 GFLOPS. Intel lists FP16 at 2.304 TFLOPS with a 2:1 ratio; NVIDIA lists no FP16.

Power and form factor differ. The Intel part is an IGP with 15 W TDP and a Ring Bus interface. The NVIDIA part is a single-slot card with 45 W TDP, no power connectors, a suggested PSU of 200 W, and a PCIe 2.0 x16 interface. The NVIDIA card is 160 mm long and 69 mm high. Display outputs differ: Intel is motherboard dependent, while NVIDIA offers 1x DVI and 1x DisplayPort 1.2.

API support differs. Intel supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3. NVIDIA supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4. Release dates differ: Intel launched on 2015-08-31, NVIDIA on 2014-07-21. Both are end-of-life. NVIDIA lists a predecessor (Quadro Fermi) and successor (Quadro Maxwell); Intel lists neither. NVIDIA also lists transistor count (1,870 million), die size (148 mm²), and transistor density (12.6M / mm²); Intel lists none of these.