Re: Does NVDA take up a lot of ram in win10?
@12, I find that... really, really weird. And it almost sounds too hard to believe. I suppose it makes sense because, if you really want to be pedantic, the processor has less data to operate on (e.g.: 4 bytes instead of 8), as well as a smaller address space. I seriously wonder if its because the microcode for 64-bit native operation might be mis-optimized some way and the 32-bit code is better optimized? I don't know... that just seems really weird to say that 32-bit code is potentially much faster than 64-bit code. I'd blame it on paging if it weren't for the fact that 64-bit mode requires paging, and X32 mode can't bypass that.
Of course, the black magic to emulate 32-bit code on 64-bit systems is... less than ideal and ridiculously messy -- so messy that kernel developers have discussed its removal.
Edit: as that article notes, the extra performance is gained by the Linux kernel exposing 64-bit CPU registers and other features to code that runs using the X32 ABI. This is not 32-bit code emulation, however: "The Linux x32 ABI as a reminder requires x86_64 processors and is engineered to support the modern x86_64 features but with using 32-bit pointers rather than 64-bit pointers. The x32 ABI allows for making use of the additional registers and other features of x86_64 but with just 32-bit pointers in order to provide faster performance when 64-bit pointers are unnecessary." In other words, I don't think that the code is necessarily 32-bit but that the pointers are. The code could even be a combination of 32-and 64-bit code.
-- Audiogames-reflector mailing list Audiogames-reflector@sabahattin-gucukoglu.com https://sabahattin-gucukoglu.com/cgi-bin/mailman/listinfo/audiogames-reflector
