well nvidia just announced hardware on their new cards to enhance data transfers
I want to know more about DirectStorage and how it's implemented. Curious if it requires other hardware like a compatible chipset or something.
I want to know more about DirectStorage and how it's implemented. Curious if it requires other hardware like a compatible chipset or something.
Plus, wouldn't this require new motherboard chipsets? Or a BIOS update? How does it know to send the data to the GPU?
Same questions as I had as well. It'll be interesting to see how this plays out. Though maybe eastman has it right, where the software is modified, so it loads the data compressed to the GPU and then does processing there. I think that would require direct support by the program as it alters the typical data flow.
Nvidia's RTX IO has integration with Microsoft DirectStorage to accelerate loading into the GPU.
Here's the Nvidia RTX IO slide:
Same questions as I had as well. It'll be interesting to see how this plays out. Though maybe eastman has it right, where the software is modified, so it loads the data compressed to the GPU and then does processing there. I think that would require direct support by the program as it alters the typical data flow.
In that case the chart they showed doesn't make sense as it shows the data going directly to the GPU. Another artistic impression?
Andrew Burnes Admin • a few seconds ago Featured by GeForce.com
Hi. RTX IO is supported on all GeForce RTX Turing and NVIDIA Ampere-architecture GPUs.
@Shortbread DIrectStorage is an API, so software support is a must. In terms of bypassing the cpu to access the pcie lanes for the SSD, that's where I'm wondering if it'll require chipset support.
@Shortbread DIrectStorage is an API, so software support is a must. In terms of bypassing the cpu to access the pcie lanes for the SSD, that's where I'm wondering if it'll require chipset support.
So I guess this means either we should expect very different IO performance between the two architectures or the aforementioned on-GPU decompression meant just general GPU compute without dedicated hardware, meaning the IO performance on either new console will still be a generational leap over the RTX30.Here's some good news: RTX IO will be supported by ALL RTX GPUs:
https://www.nvidia.com/en-us/geforce/news/rtx-io-gpu-accelerated-storage-technology/
Why do you say this?So I guess this means either we should expect very different IO performance between the two architectures or the aforementioned on-GPU decompression meant just general GPU compute without dedicated hardware, meaning the IO performance on either new console will still be a generational leap over the RTX30.
also MS confirmed that a new update of Windows 10 is going to add those features soonHere's some good news: RTX IO will be supported by ALL RTX GPUs:
https://www.nvidia.com/en-us/geforce/news/rtx-io-gpu-accelerated-storage-technology/