Reply to post: Re: Meltdown and Spectre - a Wake-up call for HPC

Meltdown, Spectre: The password theft bugs at the heart of Intel CPUs

Ceiling Cat

Re: Meltdown and Spectre - a Wake-up call for HPC

Your comment makes a lot of sense to me, even though I'm just the bloke who cleans the lavatory.

My interests are gaming, not HPC. In either, there is a need for data to be passed as quickly as possible, with a minimum of interference/pipeling lag introduced by what I'll just simplify as "security-related functions".

I only kind of understand the idea of "speculative execution" - I'm not a coder, but it feels like a cousin of a feature that Windows used to use to pre-load certain files based on the times they were most frequently used. Some kind of pre-caching whose specific name escapes me.

Being uneducated in this aspect of computing, I can't help but think that such a function is not as necessary for code which runs from a compiled state, rather than code that is runs "just in time". Or am I wrong on this?

I am only being so serious because my brain hurts from trying to figure this out : On one hand, there are explanations that are way too technical; on the other hand, there's a lot (and I mean a *lot*) of FUD, trolling, and general "Hurr Durr intel Fucked Up", and I can't really make heads or tails of this.

All I know is, If I'm patched, my performance may take a nosedive. If I don't patch, I'm at risk of ... [variable unknown]. Do either or both vulnerabilities really affect non-productivity machines as badly as HPC/Server Class gear? And will there be improvements made to mitigate the performance hit?

Can someone get me a Tylenol? As I said earlier, my brain hurts.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019