Quite. I'll be moving to Zen 3xxx or TR.
-
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.
New Intel vulnerability: SPOILER
- Thread starter DragonQ
- Start date
Quite. I'll be moving to Zen 3xxx or TR.
Glad I don't do anything particularly with servers any more (other than locked down personal projects) - back in the day I use to host quite a lot of game servers with remote user access (though only via webmin/usermin).
It is keeping my brother in business though (contract IT) with all the system migrations, etc.
Soldato
- Joined
- 18 May 2010
- Posts
- 23,266
- Location
- London
There is no Red Hat advisor on this (yet) so may not be all that serious.
Soldato
- Joined
- 16 Feb 2010
- Posts
- 5,223
- Location
- North East England
Won’t somebody think of the children! All those stories of mass attacks !!!!! Oh wait.
You're such an apologist. They have confirmed that it is a problem with Intel chips, they have yet to confirm if it is a problem with AMD chips." so it may affect Ryzen also " looks at title lol . almost daily mail .
They're literally stating a fact as it's currently known. If it's found to be a problem with Ryzen, then there'll articles stating that a vulnerability has been found wjth Ryzen.
It's a complete non-issue, stop being an apologist.
It's still cutting corners. It's why vulnerabilities were found in the first place, and why the fixes involve performance regressions.
It's a little bit like the VAG emissions scandal. It's impossible that they didn't know about it.
Don't agree with the progression of your logic there - vulnerabilities found or performance regression isn't confirmation of intentional cutting corners.
There may be some degree of cost/corner cutting in respect to how they've held back progress overall - almost any hardware like this gets broken over time even Zen in years to come will probably see dozens of ways to exploit it eventually being found - same as security features, etc. on many 16 bit CPUs have been compromised now.
It is an area of technology that has gone through many evolutions and improvements iteratively over time it is unlikely the full extent was understood even by Intel in earlier implementations albeit the last few years various professionals have voiced concerns especially in respect to weaknesses in speculative functionality.
You'd have a point if there weren't numerous security issues, each one bringing a regression in performance when they're fixed.
It's like when nVidia were equipping their cards with weak amounts of VRAM with poor memory bandwidth.
They knew exactly what the implications would be, even medium term. I'm not even saying it's wrong that Intel cut corners, because these security vulnerabilities seem to only be a problem with enterprise, and pose minimal risk to people using their PCs to play games. Cutting corners in a certain context to maximise performance in a certain region isn't necessarily a bad thing to do, I just simply don't believe they didn't have any understanding of the consequences given the amount of vulnerabilities that have been discovered.
A lot of these security issues are variations of the same problem. Your logic seems to be of the smoke = fire variant however.
It makes me more concerned for Intel's management engine than ever as if a remote intrusion vector is found it is pretty much game over for any system exposed to the internet using many of the Intel CPUs of the last decade.
The complexities and evolution of speculative execution ( http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.119.2934&rep=rep1&type=pdf ) makes it less likely that the problem is mainly down to cutting corners - AMD has benefited a bit being slightly behind the curve as they've been able to learn from the mistakes of others when it comes to safe practises both execution stability and security wise.
Last edited:
Rroff you are right, these are simply variations of the same problem.
All CPU architectures will have the same issue in one way or another and to varying degree's (I don't care if AMD, ARM, IBM say other wise), the basic base design is the same as there is a finite way a basic CPU can be designed.
All silicon has vulnerabilities, in today's design's you cannot cover 100% in your verification.
All CPU architectures will have the same issue in one way or another and to varying degree's (I don't care if AMD, ARM, IBM say other wise), the basic base design is the same as there is a finite way a basic CPU can be designed.
All silicon has vulnerabilities, in today's design's you cannot cover 100% in your verification.
The way Intel have or more to the point haven't handled these problems is shocking. Intel have been aware of it's problems for a very long time.
It is fairly shocking. My skylake laptop never got spectre and meltdown patches so I replaced it for another Intel machine that did. The server estate I had budget last year but am holding out for Rome. Literally the second I can buy a gen 11 hp server rocking Rome I'll be done with Intel. Lots of the database stuff we do is slower than last year on the same estate, I didn't sign up for regression to this degree.
Patching a BIOS with newer microcode is a lot cheaper than buying a whole new laptop! My Ivy Bridge laptop is all up to date.
In the work I do it's important that the manufacturer releases updates to their bios when it comes to security. Why should I spend my time injecting microde? The manufacturer didn't release a supported bios so I changed machine. Could I have done it myself? Probably... would it have been manufacturer supported? Unlikely.
Besides messing about with semi-supported microcode with often potentially dubious 3rd party involvements or reliance on 3rd parties editing/distributing them is a recipe for disaster in a business context.
Would never get past compliance anywhere I've worked that runs to a professional standard.
For me it was cheaper and easier combined with the other issues you point out and what it's used for it was a no brainer. I'm now 8th gen rather than 6th and honestly there is little difference, if anything an unpatched skylake 6700hq is probably a tad faster than the patched 8705g. The skylake machine still gets used when friends come over for gaming as it's still a very capable machine.
Permabanned
- Joined
- 11 Jan 2019
- Posts
- 3,214
- Location
- bedlam
meltdown is the worst one so far, spectre realistically cannot be exploited, meltdown itself is difficult enough that I have kept mitigation disabled on my main rig. Not looked into this new one yet so no comment on that. But what I will say is whenever something is patched its my view a cost vs benefit assessment should always be carried out, sadly it seems to have become the trend where people are just told to patch every single "possible not necessarily live" exploit out there regardless of the actual risk and cost of doing so.