News for hackers. Stuff that matters.

A comment on After 50 years of effort, researchers made silicon emit light. Parent comment

Re:After 50 years of effort, researchers made silicon emit l... by MrBuddyCasino on Saturday April 11, 2020 @09:30AM (#22841574)
> In most architectures, the caches are physically addressed, so TLB lookups occur before even L1 cache access.

So to see if a memory location is contained in a cache line, a TLB lookup is needed to first get the physical address? I wouldn't have expected this, can you expand on why this is the case?

[ Parent | on HN ]
Re:After 50 years of effort, researchers made silicon emit l... by dooglius on Saturday April 11, 2020 @11:09AM (#22842038)
See Linus Torvalds' thesis, section 4.3.3 "The case against virtual data caches": https://www.cs.helsinki.fi/u/kutvonen/index_files/linus.pdf
[ Parent | on HN ]
Re:After 50 years of effort, researchers made silicon emit l... by pwildani on Saturday April 11, 2020 @11:03AM (#22842008)
If you use virtual addresses to index your caches, you have to clear them all on every process context switch.

With physical addresses, you just have to clear your TLB cache.

[ Parent | on HN ]
Re:After 50 years of effort, researchers made silicon emit l... by adwn on Saturday April 11, 2020 @11:43AM (#22842235)
Two reasons: 1) a virtual address might refer to different physical addresses (see pwildani's comment), and 2) a physical address can be mapped to different virtual addresses – a virtually addresses cache has to keep track of that somehow, otherwise the cache will become incoherent.
[ Parent | on HN ]