Intel Iris Pro Graphics P6300 vs NVIDIA GeForce GTX 670M Comparison

Where Each One Wins

The benchmark data splits cleanly along a single axis: the NVIDIA GeForce GTX 670M wins the only recorded compute test, while the Intel Iris Pro Graphics P6300 has no victories in any measured category. In the Geekbench OpenCL test, the GTX 670M scores 6513 against 5712 for the Intel part, a 14% advantage. That is the entire competitive margin in this comparison, and it reflects where each design concentrates its resources.

The GTX 670M occupies the 38th percentile among all GPUs in the database, while the Iris Pro P6300 sits at the 33rd percentile. The five-point percentile gap is modest, but it translates into a clear hierarchy in raw compute throughput. The GTX 670M's nearest rivals include the GeForce GT 555M (6493, 0.3% behind), the Quadro M5000M (6481, 0.5% behind), and the Radeon Vega 10 Mobile (6476, 0.6% behind). The Intel part's nearest rivals are the GeForce GTX 670MX (5721, 0.1% ahead of Intel), the GeForce GTX 550 Ti (5731, 0.3% ahead), and the Radeon HD 8790M (5691, 0.4% behind Intel). These groupings show that the GTX 670M competes in a higher performance tier, while the Iris Pro P6300 trades blows with older discrete mobile parts and desktop cards from the previous generation.

For workloads that stress OpenCL compute, the GTX 670M is the clear choice. The 14% delta is not enormous, but it is consistent and measurable. For any task where the Intel IGP would be considered, the data suggests the NVIDIA part delivers more usable throughput per unit of work. However, the Iris Pro P6300 does not trail by a wide margin in absolute terms; its 5712 score is within striking distance of several discrete GPUs, which is notable for an integrated solution.

Architecture Differences

The two parts come from different foundries, different process nodes, and fundamentally different design philosophies. The GTX 670M uses the GF114 chip built on TSMC's 40 nm process, packing 1,950 million transistors into a 332 mm² die. That yields a transistor density of 5.9 million per square millimeter. The Iris Pro P6300 uses Intel's Broadwell GT3e die on a 14 nm process with Intel as the foundry. Die size and transistor counts are not recorded for the Intel part, but the process node advantage is substantial: 14 nm versus 40 nm represents a generational leap in manufacturing.

The compute resources tell a nuanced story. The GTX 670M fields 336 shading units, 56 texture mapping units, and 24 ROPs. The Iris Pro P6300 counters with 384 shading units, 48 TMUs, and only 6 ROPs. The Intel part has more shaders but fewer texture units and dramatically fewer ROPs. This explains the raw throughput numbers: the GTX 670M reaches 803.7 GFLOPS FP32 and 33.49 GTexel/s, while the Iris Pro P6300 manages 614.4 GFLOPS and 38.40 GTexel/s. The Intel IGP actually wins the texture rate race by a meaningful 15% (38.40 versus 33.49 GTexel/s), but its pixel rate suffers badly: 4.800 GPixel/s versus 8.372 GPixel/s for the NVIDIA part. The 74% advantage in pixel throughput for the GTX 670M is a direct consequence of having four times the ROP count.

Memory subsystems diverge completely. The GTX 670M uses 1536 MB of dedicated GDDR5 on a 192-bit bus, delivering 72.00 GB/s of bandwidth at an effective 3 Gbps memory clock. The Iris Pro P6300 shares system memory, with bus width and bandwidth described as system dependent. This is a structural difference: dedicated VRAM versus shared memory. The GTX 670M's 72 GB/s is a fixed, guaranteed figure, while the Intel part's bandwidth varies with the host system's memory configuration.

Clock behavior also differs. The Iris Pro P6300 has a base clock of 300 MHz and a boost of 800 MHz, which is typical for an IGP that ramps based on thermal and power headroom. The GTX 670M's base and boost clocks are not recorded, only its 750 MHz memory clock. Power envelopes reflect the design intent: the GTX 670M is rated at 75 W TDP and comes as an MXM module, while the Iris Pro P6300 is a 15 W IGP on the Ring Bus interface. The NVIDIA part requires no power connectors, but its MXM-B (3.0) form factor is a discrete module. The Intel part is an integrated graphics processor with motherboard-dependent display outputs; the GTX 670M's display outputs are portable-device dependent.

API support shows a split. The GTX 670M supports DirectX 12 (11_0) and OpenGL 4.6, but has no Vulkan entry. The Iris Pro P6300 supports DirectX 12 (11_1), OpenGL 4.4, and Vulkan 1.0. The Intel part has broader API coverage in the database, including Vulkan, while the NVIDIA part reaches a higher OpenGL version. Neither supports FP16 in the recorded data.

The Verdict

The data points to a straightforward recommendation for compute-oriented workloads: the NVIDIA GeForce GTX 670M is the stronger part. Its 6513 OpenCL score beats the Iris Pro P6300's 5712 by 14%, and its 38th percentile placement versus the Intel part's 33rd confirms the hierarchy. The GTX 670M also delivers substantially higher pixel throughput (8.372 GPixel/s versus 4.800 GPixel/s) and higher FP32 compute (803.7 GFLOPS versus 614.4 GFLOPS). Its dedicated 72.00 GB/s memory bandwidth is a structural advantage over system-shared memory, which is inherently variable.

However, the Iris Pro P6300 is not without merits. Its texture rate of 38.40 GTexel/s exceeds the GTX 670M's 33.49 GTexel/s, which could benefit texture-heavy workloads. Its 15 W TDP versus 75 W means it fits in power-constrained designs where the GTX 670M cannot. The Intel part also supports Vulkan 1.0 and DirectX 12 (11_1), while the NVIDIA part lacks Vulkan in the recorded data. For an integrated solution, the Iris Pro P6300's 5712 score is competitive with discrete parts like the GeForce GTX 670MX (5721) and GeForce GTX 550 Ti (5731), both of which score within 0.3% of the Intel IGP.

Who should pick which? Users who need consistent, dedicated graphics memory and higher raw compute should choose the GTX 670M. Users who require an integrated solution with lower power draw, Vulkan support, and a higher texture rate should consider the Iris Pro P6300. The database shows the GTX 670M as the overall compute winner, but the Intel part wins the efficiency and API coverage arguments.

FAQ

Q: Which GPU has the higher OpenCL score?

A: The NVIDIA GeForce GTX 670M scores 6513, which is 14% higher than the Intel Iris Pro P6300's 5712.

Q: How do these GPUs compare to their nearest rivals?

A: The GTX 670M's closest rival is the GeForce GT 555M at 6493, a 0.3% difference. The Iris Pro P6300's closest rival is the GeForce GTX 670MX at 5721, which is 0.1% ahead of the Intel part.

Q: Does the Intel Iris Pro P6300 have any performance advantage?

A: Yes, its texture rate is 38.40 GTexel/s versus 33.49 GTexel/s for the GTX 670M, a 15% advantage.

Q: What is the memory bandwidth of each GPU?

A: The GTX 670M has a fixed 72.00 GB/s from 1536 MB of GDDR5 on a 192-bit bus. The Iris Pro P6300 uses system shared memory with bandwidth described as system dependent.

Q: Which GPU supports Vulkan?

A: The Intel Iris Pro P6300 supports Vulkan 1.0. The NVIDIA GTX 670M has no Vulkan entry in the database, but supports DirectX 12 (11_0) and OpenGL 4.6.

Q: What are the power ratings?

A: The GTX 670M is rated at 75 W TDP, while the Iris Pro P6300 is rated at 15 W TDP.

Head-to-Head Benchmarks

The only recorded head-to-head benchmark is Geekbench OpenCL, and it favors the NVIDIA GeForce GTX 670M decisively. The GTX 670M scores 6513 against 5712 for the Intel Iris Pro P6300, a 14% delta in favor of the NVIDIA part. This is the single largest measurable gap between these two GPUs in the database.

The 14% margin places the GTX 670M comfortably ahead, but context from the nearest rivals clarifies the competitive landscape. The GTX 670M's 6513 score puts it 0.3% ahead of the GeForce GT 555M (6493), 0.5% ahead of the Quadro M5000M (6481), and 0.6% ahead of the Radeon Vega 10 Mobile (6476). It trails the Intel UHD Graphics P750 (6554) by 0.6%. This means the GTX 670M sits in a tight cluster of similarly performing parts, with the entire spread from 6476 to 6554 covering less than 1.2% total variance. The Iris Pro P6300's 5712 score, by contrast, sits in a different cluster: it trails the GeForce GTX 670MX (5721) by 0.1%, trails the GeForce GTX 550 Ti (5731) by 0.3%, beats the Radeon HD 8790M (5691) by 0.4%, and beats the Quadro M500M (5604) by 1.9%. The distance between these two clusters is the story: the GTX 670M's cluster sits roughly 800 points higher, which is the 14% gap.

Looking beyond the single benchmark, the architecture data reinforces the GTX 670M's advantage in certain metrics. The pixel rate of 8.372 GPixel/s versus 4.800 GPixel/s gives the NVIDIA part a 74% edge in fill-rate-bound scenarios. The FP32 compute of 803.7 GFLOPS versus 614.4 GFLOPS is a 31% advantage. The dedicated memory bandwidth of 72.00 GB/s versus system-shared memory is not directly comparable, but the fixed nature of the GDDR5 allocation is an advantage for consistent performance.

The Iris Pro P6300 does claim one outright win in the recorded specifications: texture rate. Its 38.40 GTexel/s beats the GTX 670M's 33.49 GTexel/s by 15%. This is a real, if narrow, victory for the Intel part, driven by its 48 TMUs versus 56 TMUs but higher clock behavior. The Intel part also has more shading units (384 versus 336), but its lower FP32 throughput suggests those shaders are not being fed as effectively, likely due to the shared memory subsystem.

The verdict from the head-to-head data is unambiguous on compute: the GTX 670M wins the only benchmark that matters in this comparison. But the texture rate result and the power envelope (15 W versus 75 W) mean the Intel part is not a pure also-ran. It wins where texture throughput and power efficiency are the priorities. For general compute, the GTX 670M's 14% lead and its 38th percentile placement versus the 33rd percentile for the Intel part make it the recommended choice in the database.