TLDR its reinventing the wheel coz criminal **bleep**s stole all your wheels and you been driving stuff jacked up on bricks.
Its something thats in ALL data center and networking and storage/HDD CDROM technology since FOREVER AGO.
Okay, so bring up some photos of the Xbox ONE hardware and the playstation 3 hardware..
you will see theres special RAM in a separate chips around the main processor they arent GROUPED together on one stick or slot but different islands.
Think of the PS3 as having a NORTH and SOUTH RAM .. theres the main whatever the size was its been years.. 2GB.. then theres another shared memory which is separate maybe from the uhh computer memory? basically theres a shared memory pool thats total size is like double what the specs say for its GPU or whatever. But you see, AMD hardware since long ago like some of the first x64 CPU's had their OWN INTEGRATED MEMORY CONTROLLER. and now the infinity fabric means the CPU doesnt have to wait on slow RAM timings or fake intel XMP profiles or fake sticks of RAM. But physically placing them the GPU and the RAM and the CPU around the board like that in different islands.. the boat that sails the data there.. you might have to wait for a round trip. So.. the same old trick in CD ROM or IDE spinning disc magnetic hard drives with buffering and caching was used with a type of uhh PAGE FILE. which being accessed by all the different hardware and apps for frequently used things.. is you guessed it, how drive tiering and ryzen store MI works.. AMD had that stuff for ages in business and data center storage and load balancing.. basically what you call a swap file or page file or drive buffering or caching.. is called a shared page memory pool.. Imagine having a bunch of cheap large 10TB spinning disc HDD with sata ports.. but you've got one 1TB or 2 TB stick of the fastest SSD or NVME.. so the lower latency and higher 550MEGABYTES/sec transfer speed could boot your OS or get website data quicker for servers or whatever. So the most accessed and used data gets automatically stored on the faster SSD.. well the ps3 and other AMD systems have always had this hardware cache thats a central sort of memory pool or a system for letting the apps see what the others doing with the hardware so it doesnt waste a trip by using a Microsoft one note or google groups or google docs approach.. so yeah they all call it something different.. intel tries to charge fortunes for it after disabling it.. calling it OPTANE.. or whatever.. but yeah.. its just basically letting you use a more CENTRALIZED bit of memory like a cache or buffer. so the high performance stuff can be used less and still have the same performance. its a method of saving costs like RAID redundant array of inexpensive disks.
So the smart access means the SMART access means it makes less round trips to all the islands and checks the cached middle page pool of shared data (google groups/docs one note?) and the resizeable bar was supposed to be that too in a small bit before your graphics card for uhh frequent requests. So you get a small high speed spec bit as you enter the front door of the GPU house hold and you often might get told 'we already have that go get these other things' right at the door for the GPU with resizable bar which is basically the same thing as SMART access.. but its more INSIDE the GPU as a second SMART ACCESS MEMORY. So essentially they both serve the same purpose and do a similar function.. like CPU cache. but these were often or always disabled and hadnt been enabled as hackers like to put their own google groups or microsoft one note in there and stick your recently CTRL C CTRL V copy pasted or highlighted content in there. is what i imagine, but now that our drives have no moving parts and 0 latency for all things like RAM and CPU.. you shouldnt need those buffers and page files for OLDER IDE CDROM and spinning disc drives.. as those took time to spin the disc around physically to access data inner edge or outer edge and different locations seeking it out so they waited till it had a truckload of data to drive the road train to the display or elsewhere because the display was built to pile them up like warehouse pallets before it loaded them in as CDs often had dust or scratches and read errors and maybe needed to retry or could take way longer on some parts of the disc so the rate it could read or write data could vary a lot and the 'live stream' of video or other content on the disc might not be constant or stable without a pile it up into container loads full buffering.
So you should be in fact DISABLING and deleting all your page files and pools and swap files BUFFERING queueing whatevers. I dont mean zero CPU cache as thats like coverting your coins into bank notes so its easier for all to work with.
