Intel Iris Pro Graphics 5200 vs NVIDIA GeForce GTX 650 Comparison

The Intel Iris Pro Graphics 5200 and NVIDIA GeForce GTX 650 represent two very different approaches to graphics processing from the same era. The Iris Pro is an integrated solution built into a Haswell processor, while the GTX 650 is a discrete desktop graphics card based on the Kepler architecture. The benchmark data shows a split decision: each part wins one of the two recorded tests, with the overall averages landing close together but with distinct implications for different workloads.

Where Each One Wins

The recorded data splits the two contenders almost perfectly down the middle. The Intel Iris Pro Graphics 5200 wins the Geekbench OpenCL test with a score of 5042, beating the GTX 650’s 4545 by 10.9%. This is the more general-purpose compute workload, and the Iris Pro’s advantage here is notable. The GTX 650, meanwhile, takes the Geekbench Vulkan test with 4524 against the Iris Pro’s 3677, a margin of 18.7%. Vulkan is a lower-level API that favors the discrete card’s dedicated memory and driver maturity.

Looking at the broader picture, the Iris Pro’s average benchmark score is 4360, which places it in the 26th percentile of all GPUs in the database. The GTX 650’s average is 3823, sitting in the 22nd percentile. That means the Iris Pro has a higher overall average by roughly 537 points, or about 14%, but the GTX 650’s Vulkan advantage is substantial enough to matter for specific applications.

For compute-heavy tasks that use OpenCL, the Iris Pro is the winner. Its 320 shading units and 40 texture mapping units, paired with 4 ROPs, deliver 736.0 GFLOPS of FP32 throughput. The GTX 650 counters with 384 shading units, 32 TMUs, and 16 ROPs, producing 812.5 GFLOPS. Despite having fewer shading units, the Iris Pro wins the OpenCL test, which suggests that its architecture or drivers handle that workload more efficiently. For Vulkan, the GTX 650’s dedicated 1024 MB GDDR5 memory with 80.00 GB/s bandwidth gives it a clear edge, as the Iris Pro relies on system shared memory with bandwidth that is system dependent.

The Verdict

The choice between these two comes down to what software you plan to run. If your primary workload is OpenCL compute, the Intel Iris Pro Graphics 5200 is the stronger option. Its 10.9% lead in that test is the largest single margin either card holds in the head-to-head benchmarks. The data indicates that for general compute tasks, the integrated solution outperforms the discrete card.

If you need Vulkan support, the NVIDIA GeForce GTX 650 is the clear choice. Its 18.7% advantage in the Vulkan test is nearly double the Iris Pro’s OpenCL lead. The GTX 650 also has a higher API version for Vulkan at 1.2.175 compared to the Iris Pro’s 1.0, which means better compatibility with modern Vulkan applications.

For users who want a single card that handles both compute APIs reasonably well, the average scores tell the story. The Iris Pro’s average of 4360 is 0.6% ahead of the NVIDIA GeForce RTX 4070 GDDR6’s 4335 and 1.5% ahead of the AMD FirePro W2100’s 4295, while sitting 0.6% behind the NVIDIA GeForce 930M’s 4388 and 1.2% behind the NVIDIA GeForce GT 645M’s 4411. The GTX 650’s average of 3823 is 0.3% behind the NVIDIA GeForce MX110’s 3834 and 1.5% behind the AMD Radeon R5 Graphics’ 3883, while staying 0.8% ahead of the Intel UHD Graphics 710’s 3792 and 1.9% ahead of the NVIDIA Quadro 2000’s 3898. The Iris Pro’s percentile ranking of 26 versus the GTX 650’s 22 confirms that the integrated part has a slight overall edge in the database’s accumulated measurements.

Head-to-Head Benchmarks

The first recorded test, Geekbench OpenCL, shows the Intel Iris Pro Graphics 5200 at 5042 against the NVIDIA GeForce GTX 650 at 4545. The delta is 10.9% in favor of Intel. This is a significant margin for integrated graphics. The Iris Pro achieves this with only 320 shading units and 4 ROPs, whereas the GTX 650 has 384 shading units and 16 ROPs. The Iris Pro’s texture rate is 46.00 GTexel/s, which is higher than the GTX 650’s 33.86 GTexel/s, and its pixel rate of 4.600 GPixel/s is lower than the GTX 650’s 8.464 GPixel/s. The OpenCL score suggests that the Iris Pro’s higher texture throughput and its 22 nm Haswell GT3e chip compensate for its lower pixel output.

The second test, Geekbench Vulkan, reverses the outcome. The GTX 650 scores 4524, while the Iris Pro manages only 3677, a delta of 18.7% in NVIDIA’s favor. The GTX 650’s dedicated GDDR5 memory with a 128-bit bus and 80.00 GB/s bandwidth is the obvious factor here. Vulkan workloads often benefit from low-latency, dedicated memory access, which the Iris Pro cannot provide because its memory is system shared and system dependent. The GTX 650 also supports Vulkan 1.2.175, while the Iris Pro is limited to Vulkan 1.0, which may affect driver optimizations and feature support.

The wins are split at one apiece. That means neither part dominates the other across the board. The Iris Pro’s OpenCL win is worth 10.9%, and the GTX 650’s Vulkan win is worth 18.7%. If you weight the tests equally, the GTX 650’s win is larger in magnitude, but the Iris Pro’s overall average score is higher because its OpenCL score of 5042 is much higher than its Vulkan score of 3677, while the GTX 650 stays consistent at 4545 and 4524.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The Intel Iris Pro Graphics 5200 has an average benchmark score of 4360, while the NVIDIA GeForce GTX 650 has an average of 3823. That puts the Iris Pro about 14% higher overall.

Q: Does the GTX 650 win any benchmark test?

A: Yes, the GTX 650 wins the Geekbench Vulkan test with a score of 4524 compared to the Iris Pro’s 3677, an 18.7% margin.

Q: Is the Iris Pro better for OpenCL workloads?

A: Yes, the Iris Pro scores 5042 in Geekbench OpenCL versus the GTX 650’s 4545, a 10.9% advantage. This is its only head-to-head win, but it is a clear one.

Q: How do their percentile rankings compare?

A: The Iris Pro is in the 26th percentile of all GPUs in the database, while the GTX 650 is in the 22nd percentile.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce GTX 650 has 384 shading units, while the Intel Iris Pro Graphics 5200 has 320 shading units.

Q: What is the memory configuration difference?

A: The GTX 650 has 1024 MB of GDDR5 memory on a 128-bit bus with 80.00 GB/s bandwidth. The Iris Pro uses system shared memory with system dependent bandwidth, meaning it has no dedicated graphics memory.

Architecture Differences

The two GPUs come from fundamentally different design philosophies. The Intel Iris Pro Graphics 5200 is an integrated graphics processor built on the Haswell GT3e chip, using Intel’s 22 nm process node. Its architecture is Generation 7.5, part of the HD Graphics (Haswell) generation. The chip is connected via a Ring Bus interface, and it is classified as an IGP, meaning it shares system memory and has no dedicated VRAM.

The NVIDIA GeForce GTX 650, by contrast, is a discrete GPU based on the GK106 chip, using the Kepler architecture. It is manufactured by TSMC on a 28 nm process node. The chip contains 2,540 million transistors on a 221 mm² die, giving it a transistor density of 11.5 million per square millimeter. The GTX 650 is part of the GeForce 600 generation and is the successor to the GeForce 500 series, with the GeForce 700 series following it.

The Iris Pro has a base clock of 200 MHz and a boost clock of 1150 MHz, while the GTX 650 has no recorded base or boost clock values in the database. The Iris Pro’s memory clock is listed as system shared, while the GTX 650 runs its GDDR5 memory at 1250 MHz, or 5 Gbps effective. The Iris Pro supports DirectX 12 (11_1), OpenGL 4.3, and Vulkan 1.0. The GTX 650 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175, which is a newer Vulkan version.

The Iris Pro is a 45 W part with a single-slot IGP form factor, while the GTX 650 is a 65 W card that is single-slot, requires one 6-pin power connector, and has a suggested power supply of 250 W. The GTX 650 also has a physical length of 147 mm or 5.8 inches, while the Iris Pro’s dimensions are not recorded. Display outputs differ as well: the Iris Pro is motherboard dependent, while the GTX 650 offers 1x DVI and 2x DisplayPort 1.2.

Specification Differences

The specification sheet reveals several clear differences between the two GPUs. The process node differs: Intel uses 22 nm, while NVIDIA uses 28 nm. The Iris Pro has 320 shading units, 40 TMUs, and 4 ROPs, while the GTX 650 has 384 shading units, 32 TMUs, and 16 ROPs. The Iris Pro’s pixel rate is 4.600 GPixel/s, and its texture rate is 46.00 GTexel/s. The GTX 650’s pixel rate is 8.464 GPixel/s, and its texture rate is 33.86 GTexel/s. FP32 performance is 736.0 GFLOPS for the Iris Pro and 812.5 GFLOPS for the GTX 650.

Memory is the most pronounced difference. The Iris Pro has no dedicated memory: its size, type, and bus width are all system shared. The GTX 650 has 1024 MB of GDDR5 on a 128-bit bus with 80.00 GB/s bandwidth. Power consumption also differs: the Iris Pro draws 45 W, while the GTX 650 draws 65 W and needs an auxiliary 6-pin power connector plus a 250 W power supply recommendation.

The bus interface is another split. The Iris Pro uses a Ring Bus, which is typical for integrated parts, while the GTX 650 uses PCIe 3.0 x16. The GTX 650’s display outputs are specified as 1x DVI and 2x DisplayPort 1.2, whereas the Iris Pro’s outputs are motherboard dependent. Both are end-of-life products, with the Iris Pro released on June 2, 2013, and the GTX 650 released on November 26, 2013. The GTX 650 has a predecessor in the GeForce 500 series and a successor in the GeForce 700 series; the Iris Pro has no recorded predecessor or successor. The GTX 650 is also the only one with recorded transistor and die size data: 2,540 million transistors on a 221 mm² die.