![]() ![]() The ROG RAMDisk can form ability to junction with exiting files hosted on the SSD or HDD. While less than ever with the prolific use of modern SSDs, the SATA bus is still a number of times slower than main memory access latency and bandwidth. In many cases, a application will still have its performance limited due to the need to access or cache data hosted by the SSD or HDDs into memory. However, Windows 7 or Windows 8 features a better use of memory so most of it remains empty, a large amount of the time. Modern ROG PCs usually feature a memory capacity of between 4~64GB in total. This software was developed mainly for the following usage scenario: The basic concept behind the ROG RAMDisk software is to provide a quick and easy way for users to further enhance the performance of their system and durability of their SSD(s), by utilizing the system memory (DRAM) that is not in use. You could think of putting everything in RAM but so that you can put a whole game in RAM in BC format or uncompressed instructions only for the GPU you would occupy many hundreds of GB in a AAA and it would not be economically and physically viable, it is much cheaper to synchronize either a decompressor and a transformer to GPU format via dedicated hardware or via CPU with good bandwidth.The RAMDisk software is exclusively offered on selected ASUS ROG product lines and enables users to utilize the available system DRAM to its full capability. That is why RTX IO will perform the decompression and apply logic, it is only necessary that the I / O controller requests do not saturate the OS and an SSD capable of dumping all that memory. all this process of decompression and transformation is the problem because current CPUs are multitasking to do several different tasks is part of the way a PC is and not all hardware is in the same formats for dedicated hardware can perform the function. When you take a file out of the SSD it goes to the RAM generally compressed when it is going to be used it decompresses passing through the CPU and returning to the RAM, it is generally decompressed and transformed into a format that the GPU understands BC for textures and instructions a different one etc. The main problem is that the RAM, the SSD, the GDDR and the OS do not speak the same language. This is possible, correct? Why don't devs do this more. otherwise they would just need to make due with normal storage options. A game in the future might even support both options and let you either load the entire game into RAM (if you have enough) or just use your pcie-4.0 storage the same as PS5 for example (if you have fast storage). you don't even need to have a system that has pcie-4.0 or own a super fast storage drive, you can just buy more RAM and upgrade that way (useful for systems that are a little older or don't support pcie-4.0 like my 10700k). This would also help with porting some next-gen games to PC that rely on high IO/SSD speed. ![]() (Think about the loading screens when you initially start up GTAV story mode, except faster) If loading data from a pci-e 4.0 SSD is fast, imagine if you could start up a game and then it takes like a minute or two to load the ENTIRE GAME into your system RAM, and then from that point forward it's just instantaneous loading. Some system even have 1TB these days, though that is expensive. I have 32gb but you could easily go for 64gb, 128gb, etc. Consoles have super fast SSD's but on PC you have the ability to have much more system RAM. ![]()
0 Comments
Leave a Reply. |