Intel HD Graphics P4600 vs NVIDIA GeForce GT 640 Comparison

Head-to-Head Benchmarks

The sole head-to-head benchmark available in this comparison is Geekbench OpenCL, and the data shows a decisive victory for the NVIDIA GeForce GT 640. The GT 640 scores 3736, while the Intel HD Graphics P4600 scores 3389. This represents a delta of -9.3% for the Intel part, meaning the NVIDIA GPU is roughly 10% faster in raw compute throughput under OpenCL. While a single benchmark does not tell the entire story, the margin is consistent with the underlying hardware specifications, which will be examined in later sections.

The GT 640’s advantage is not overwhelming, but it is clear. A 9.3% lead in a compute-oriented workload suggests that the NVIDIA part has a meaningful edge in shader-heavy tasks, though the Intel integrated solution is not far behind. The P4600’s performance places it in a narrow band where small architectural differences can flip the result; for instance, the P4600 sits within 1.7% of the Intel HD Graphics 530 and 2.2% ahead of the NVIDIA GeForce GT 730M in its own rival list. Conversely, the GT 640’s nearest rivals include the NVIDIA GeForce 920M at -2.3% and the NVIDIA Quadro P1000 at +1.5%, showing that the GT 640 is itself tightly clustered with other entry-level discrete GPUs.

It is worth remembering the GT 640 has three benchmark entries in its full record — Geekbench Metal at 2086, Geekbench OpenCL at 3736, and Geekbench Vulkan at 3809 — while the P4600 only lists a single OpenCL score. The GT 640’s Vulkan score is the highest of its three, suggesting that the Kepler architecture responds well to modern low-level APIs, but since the P4600 lacks comparable Vulkan or Metal data, a direct cross-API comparison is not possible from the fact pack.

FAQ

Q: Which GPU is faster in the only shared benchmark?

A: The NVIDIA GeForce GT 640 wins the Geekbench OpenCL test with a score of 3736 against the Intel HD Graphics P4600’s 3389, a delta of -9.3% for the Intel part.

Q: How does the Intel HD Graphics P4600 compare to its closest rivals?

A: The P4600 is 1.2% behind the NVIDIA GeForce GT 740 (3431), 1.3% behind the NVIDIA Quadro 2000M (3434), 1.7% ahead of the Intel HD Graphics 530 (3332), and 2.2% ahead of the NVIDIA GeForce GT 730M (3316). Its average benchmark score is 3389.

Q: What is the average benchmark score for the NVIDIA GeForce GT 640, and how does it sit among its peers?

A: The GT 640 averages 3210 across its benchmark entries. It is 0.9% ahead of the Intel Arc Pro B60 (3182), 1.5% ahead of the NVIDIA Quadro P1000 (3163), 2.3% behind the NVIDIA GeForce 920M (3287), and 3.2% behind the NVIDIA GeForce GT 730M (3316).

Q: Does the GT 640 have any performance advantage beyond OpenCL?

A: The GT 640’s own benchmark suite includes Vulkan (3809) and Metal (2086) scores, with Vulkan being its strongest result. The Intel P4600 has no Vulkan or Metal scores listed, so no direct comparison can be made on those APIs.

Q: Which GPU has a higher pixel and texture throughput?

A: The GT 640 leads significantly: it delivers 7.216 GPixel/s and 28.86 GTexel/s, while the P4600 manages 2.400 GPixel/s and 24.00 GTexel/s. The pixel rate advantage is nearly 3x, though the texture rate gap is smaller.

Q: What are the FP32 compute figures for each GPU?

A: The GT 640 outputs 692.7 GFLOPS of FP32 compute, while the P4600 delivers 384.0 GFLOPS. This translates to an 80% advantage for the NVIDIA part in raw single-precision throughput.

The Verdict

The data is unambiguous: for any workload that stresses raw compute, pixel fill, or texture throughput, the NVIDIA GeForce GT 640 is the superior choice. Its FP32 output of 692.7 GFLOPS is 80% higher than the P4600’s 384.0 GFLOPS, and its pixel rate of 7.216 GPixel/s nearly triples the Intel part’s 2.400 GPixel/s. The GT 640 also benefits from dedicated 2 GB of DDR3 memory on a 128-bit bus with 28.51 GB/s of bandwidth, whereas the P4600 relies on system-shared memory with bandwidth described as "System Dependent."

However, the P4600 is not without merit. As an integrated graphics processor (IGP) rated at 84 W TDP, it consumes more power than the GT 640’s 65 W, but it requires no separate card, no power connectors, and uses the Ring Bus interface rather than PCIe 3.0 x16. The P4600’s OpenCL score of 3389 is within striking distance of the GT 640’s 3736, and its nearest-rival deltas are all within ±2.2%, indicating that it competes in the same performance class as low-end discrete GPUs. For users constrained by motherboard slots or expansion limitations, the P4600 might be the only option. Yet, for anyone who can install a single-slot card with no auxiliary power, the GT 640 wins on every measurable performance metric in this comparison.

Specification Differences

The two GPUs differ across nearly every hardware dimension. The process node is a key distinction: the Intel P4600 uses a 22 nm process at Intel’s foundry, while the GT 640 uses a 28 nm process at TSMC. The GT 640 has a known transistor count of 1,270 million on a 118 mm² die, yielding a density of 10.8M per mm²; the P4600’s transistor count and die size are not listed. Clock behavior also differs: the P4600 has a base clock of 350 MHz and a boost of 1200 MHz, whereas the GT 640 lists no base or boost clock, only a memory clock of 891 MHz (1782 Mbps effective). Memory allocation is polar opposite — the P4600 uses system-shared memory with a dependent bandwidth, while the GT 640 has 2 GB of dedicated DDR3 on a 128-bit bus at 28.51 GB/s.

Compute unit counts heavily favor the GT 640: it has 384 shading units, 32 TMUs, and 16 ROPs, against the P4600’s 160 shading units, 20 TMUs, and 2 ROPs. This explains the fill-rate and compute disparities. The GT 640 also has a much higher pixel rate (7.216 vs 2.400 GPixel/s) and texture rate (28.86 vs 24.00 GTexel/s). Power and physical design diverge as well: the GT 640 is a single-slot, 145 mm (5.7 inches) card with a 65 W TDP, a 250 W suggested PSU, and no power connectors; the P4600 is an IGP with an 84 W TDP and no slot width, relying on the motherboard for display output. The GT 640 uses PCIe 3.0 x16, while the P4600 uses a Ring Bus interface. Display outputs are also distinct — the GT 640 offers 1x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, whereas the P4600’s outputs are motherboard dependent.

Architecture Differences

Architecturally, these are entirely different designs. The Intel HD Graphics P4600 is built on the Haswell GT2 chip using Intel’s Generation 7.5 architecture, with a codename of "HD Graphics-W (Haswell)." The NVIDIA GeForce GT 640 is based on the GK107 chip using the Kepler architecture from the GeForce 600 generation. The foundries and processes reflect their respective manufacturers: Intel uses its own 22 nm node, while NVIDIA relies on TSMC’s 28 nm process.

The API support reveals generational differences. The P4600 supports DirectX 12 (11_1), OpenGL 4.3, and Vulkan 1.0. The GT 640 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The GT 640 has newer OpenGL and Vulkan versions, which may matter for compatibility with modern software, even though its DirectX feature level is slightly older (11_0 vs 11_1). Neither GPU lists ray tracing or tensor cores, confirming that both are legacy designs without dedicated acceleration for those workloads.

The GT 640’s predecessor is the GeForce 500 series, and its successor is the GeForce 700 series; the P4600 lists no predecessor or successor. Release timing is close: the P4600 launched on 2013-05-31, while the GT 640 launched about a year earlier on 2012-06-04. Both are marked as end-of-life. The GT 640 has a launch MSRP of 99 USD, while the P4600 has no listed MSRP, consistent with it being an integrated part not sold separately.

The transistor density and die size data further highlight the design philosophy: the GT 640 packs 1,270 million transistors into 118 mm², whereas the P4600’s integrated nature likely means it shares die space with CPU cores, though the fact pack does not specify its transistor count. The GT 640’s dedicated 2 GB VRAM is a structural advantage for texture-heavy workloads, while the P4600’s system-shared memory creates variable performance dependent on the host system’s RAM bandwidth. In summary, the GT 640 is a purpose-built discrete GPU with superior throughput, while the P4600 is a capable integrated fallback that sacrifices raw speed for motherboard integration.