Video random-access memory: Difference between revisions

Revision as of 08:36, 4 June 2024

Revision as of 08:37, 4 June 2024

Line 15:

[[File:Nvidia@8nm@Ampere@GA102@GeForce RTX 3090@S TW 2032A1 SNNB9W.000 GA102-300-A1 DSC06025-DSC06107 (50740715646).jpg|thumb|A [[GPU]] die surrounded by VRAM chips]]

Many modern GPUs rely on VRAMkire khar. In contrast, a GPU that does ''not'' use VRAM, and relies instead on system RAM, is said to have a [[unified memory architecture]], or [[shared graphics memory]].

System RAM and VRAM have been segregated due to the bandwidth requirements of GPUs,<ref>{{Cite web|title=What is VRAM: The Memory Power Behind Real-time Ray-Tracing|url=https://digitaladvisor.com/graphics-cards/what-is-vram/}}</ref><ref>{{Cite web|url=https://itigic.com/relationship-between-ram-and-vram-bandwidth-and-their-latency/|title=Relationship Between RAM and VRAM Bandwidth and Their Latency|date=17 May 2021 }}</ref> and to achieve lower latency, since VRAM is physically closer to the GPU die.<ref>{{Cite web|url=https://www.makeuseof.com/ram-vs-vram-differences/|title=RAM vs. VRAM: What's the Difference?|website=makeuseof.com|date=16 July 2021 }}</ref>