Discussing PCIe generations and their data rates
Discussing PCIe generations and their data rates
Someone can clarify what PCIe bandwidth means. It refers to how much data your system can transfer over the PCIe lanes, not just about which devices use which lanes. Different generations do have varying speeds and capabilities, but it doesn’t mean your GPU is always hogging all the bandwidth. Your assumption might be off—PCIe lanes are shared among multiple devices, so a high-end GPU won’t necessarily block other components. The drive’s performance depends on how efficiently it uses those lanes, not just its internal bus speed.
GPU positioned in the top x16 port typically connects via PCIe lanes from the CPU. If the CPU supports up to 16 lanes, only GPUs with that capacity can fully utilize them. For instance, an i5-10400 has a maximum of 16 lanes, while an i5-12400 offers up to 20 lanes. This means the 12400 isn't restricted to single 1x16 devices—it can also accommodate a 1x4 device simultaneously. Conversely, the 10400 is limited to either 1x16 or 4x4 configurations. The top M.2 slot links directly to the CPU, whereas lower slots connect to the motherboard chipset. The chipset itself imposes constraints; for example, an NVMe drive in an M.2 slot may restrict certain SATA ports due to bandwidth limits. Your setup ultimately depends on your CPU model and the type of NVMe SSD you install.
Visualize the tracks as pins representing physical links. Each generation handles varying data needs, with newer ones supporting higher throughput per track. PCI-E 4.0 offers about double the bandwidth compared to PCI-E 3.0, which in turn is double that of PCI-E 2.0. Placing an x8 PCI-E 4.0 card in a 16-lane slot on a PCI-E 3.0 board will utilize only 8 lanes while delivering the bandwidth equivalent to 8 PCI-E 3.0 lanes.