VMware Working On Emulated Coherent Graphics Memory - Needed For GL 4.4 / Vulkan
For ironing out the OpenGL 4.4+ support within their VMWGFX virtual graphics driver stack and/or for starting out work on bringing up Vulkan support to guest VMs running VMware virtualization products, their longtime graphics driver team is working on emulated coherent graphics memory support.
Longtime Mesa contributor Thomas Hellstrom, who had been with Tungsten Graphics before being acquired by VMware, posted their latest code on Wednesday for emulating coherent graphics memory support as needed by the latest OpenGL revisions and Vulkan.
This is the functionality by which the GPU sees any content written to the coherent memory in time for the next GPU operation and the CPU sees any content written by the GPU to that memory immediately after any fence object trailing the GPU operation has been signaled. Providing this coherent graphics memory support is trickier for setups like VMWGFX but the VMware developers have been ironing out the necessary bits.
There is a performance penalty under this emulated support of around 5%. Interested parties can learn more via this dri-devel patch series.
Longtime Mesa contributor Thomas Hellstrom, who had been with Tungsten Graphics before being acquired by VMware, posted their latest code on Wednesday for emulating coherent graphics memory support as needed by the latest OpenGL revisions and Vulkan.
This is the functionality by which the GPU sees any content written to the coherent memory in time for the next GPU operation and the CPU sees any content written by the GPU to that memory immediately after any fence object trailing the GPU operation has been signaled. Providing this coherent graphics memory support is trickier for setups like VMWGFX but the VMware developers have been ironing out the necessary bits.
There is a performance penalty under this emulated support of around 5%. Interested parties can learn more via this dri-devel patch series.
2 Comments