Tag GPU

Vulkan memory management in PasVulkan

Memory management in Vulkan is a complex and intricate process that gives developers control over how graphics memory is allocated and used. The primary principle of Vulkan’s memory management is that it offers a lower-level interface to the GPU than APIs like OpenGL, entrusting developers with the responsibility of managing memory allocations themselves. This can lead to more efficient use of memory and enhanced performance, but it also necessitates a deeper understanding of how the GPU operates and the optimal way to manage its resources.

In Vulkan, memory is organized in a hierarchy that includes heaps, memory types, and memory property flags. Heaps are fixed-size memory resources exposed by the device, and each heap can support different memory types. When creating a resource like a buffer, Vulkan provides information about which memory types are compatible with that resource. Depending on the resource’s usage flags, the developer must select the right memory type and, based on that type, the … (read more)

The problems with the Intel Iris Xe (i)GPU archiecture in connection with USB-C and Thunderbolt docking stations.

I bought a new notebook (a gaming Asus TUF Dash F15 notebook, which I want to use as a workstation notebook), which also has an Intel Iris Xe iGPU (alongside an NVIDIA Geforce RTX3070). And that has exactly the same problems as my Samsung Galaxy Book Pro (also with Iris Xe iGPU) with USB-C and Thunderbolt docks with external monitors, with random dropouts and display timing mismatch problem. So there seems to be a general problem with the new Intel Iris Xe GPU architecture regarding the DisplayPort protocol over USB-C and Thunderbolt. Because I have six different USB-C and Thunderbolt docks, and with all of them the problems occur with these two different notebooks, each with an Intel Iris Xe (i)GPU. And of course, I've also tested through different cables and as well as different monitors, with absolutely no improvement.