Soldato
I'm not sure how accurate that table is but yeah a redesign is needed if patches affect performance as much as it seems.
Not going well for them lately! I hope ARM64 can get itself in the desktop space.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
I'm not sure how accurate that table is but yeah a redesign is needed if patches affect performance as much as it seems.
Not going well for them lately! I hope ARM64 can get itself in the desktop space.
as commented the tables is incomplete/inaccurate - missing Arm(Apple+Android phones,rpi,...) if nothing else.
Indeed, I am surpsrised (non commercial) folks are not postponing the patches untill they are proven - short memories ? what is the rush to install ?
cross posting from Cinema forum - this could potentially be a lucrative exploit
Associate
In Chrome 63 you can turn on the experimental feature at chrome://flags/#enable-site-per-process (as per Google support article) - Chrome 64 is due on the 23rd.
Caporegime
This quote from Linus Torvalds is too funny to let this place pass it by, sorry i can't link it it contains a sweary word that i edited
Last edited:
- Joined
- 30 Jul 2006
- Posts
- 12,129
Yes, so funny that it appeared a few days ago on this very thread.
I don't understand Linus's point - if he is saying that the numerous Intel/ARM cpu architects involved over many generations should have understoood and avoided the problem, then that seems a very arrogant point of view, if it was so obvious why did he not stand up.
[ARM64 is not immune]
For mitigating the problem in current architectures, Linus does not seem to propose an alternative solution,
we are where we are , and, as written in multiple posts
"Performance at the expense of security is a compromise made daily in every aspect of IT."
[ARM64 is not immune]
For mitigating the problem in current architectures, Linus does not seem to propose an alternative solution,
we are where we are , and, as written in multiple posts
"Performance at the expense of security is a compromise made daily in every aspect of IT."
Performance at the expense of security
Now that's a logo - but would it fit on a Intel box
Associate
The context is that Intel have repeatedly denied that there is a bug (the failure to check the permission bit when speculatively executing - this causes the Meltdown vulnerability) in their CPUs (see for example their FAQ).
While Arm's Cortex A75 (which is so new it doesn't seem to be in any shipping products yet) is subject to Meltdown, Arm hasn't made the same flat-out denial, though their FAQ is diplomatically woolly on the matter.
Soldato
Google claims they already have a mitigation.
It's not hardware, but some software magic that will lessen the impact of patches.
Associate
- Joined
- 27 Apr 2007
- Posts
- 965
You should read more on this as unfortunately it's not just an Intel issue although at the moment Intel are impacted much more than others.
Associate
Ugh the Verge article has gotten some things mixed up as they haven't understood this Google blog post they linked correctly.
KPTI is the mitigation in Linux for Meltdown; this causes a significant perf hit for VMs and I/O intensive loads. Windows has a similar mitigation for Meltdown with similar effects.
In the Google article on Retpoline which they linked to, it says near the beginning: "We wanted to share a binary modification technique that we have developed for protecting against “Branch target injection”, also referred to as “Spectre” ". In other words the Retpoline technique is only relevant to one of the variants of Spectre, and so is not a replacement for the OS level mitigations for Meltdown.
The Google blog post never claimed that Retpoline was a replacement; they in fact say that they deployed KPTI in addition. Google's comment on their observed performance impact of KPTI being fairly benign is simply to state that while certain synthetic benchmarks can show marked performance hits, whether you'll suffer a big hit or not comes down to your workloads, and for their own workloads they got off lightly. Basically - and which has been stated before several times - YMMV.
The Verge somehow thinks KPTI was a fix for branch target injection (Spectre) and would be replaced by Retpoline because Google found it was faster - that's incorrect.
Definitely not me but... On Windows 10 you cannot stop it? At least on Windows 7 it cannot be installed.
Last edited:
Associate
Meltdown is an intel issue (ok and cortex but they're hardly used), they're just trying to spin it as everyone. Spectre is a different issue which is much less severe but can in some circumstances affect some rivals processors. To be clear this is very much an intel issue.
I write this as their effort at spin is working well as many people in here are repeating the pr line of it being everyone which is effectively wrong.
Last edited:
Associate
Yeah, but this is OCUK. Intel fanboys everywhere around here.
Guessing Haswell-E won't get one?
Soldato
I wasn't expecting any BIOS update for my asrock z97 board, no new BIOS for 16 months, but then at the bottom I see that a beta BIOS was released less than a month ago, so I think I'm in with a good shot of getting one eventually. If they do, then it would certainly encourage me to go asrock again next time, the boards been great too.