New CacheOut Speculative Execution Vulnerability Hits Intel Processors
Intel is not spared when it comes to the number of vulnerabilities that keep hitting their processors. The latest one is CacheOut, a new speculative execution attack that is capable of leaking data from Intel CPUs across many security boundaries. All processors up-to-the recent Coffee lake refresh are effected.
Despite Intel's attempts to address previous generations of speculative execution attacks, CPUs are still vulnerable, allowing attackers to exploit these vulnerabilities to leak sensitive data. Unlike previous MDS issues, the researchers show in their work how an attacker can exploit the CPU's caching mechanisms to select what data to leak, as opposed to waiting for the data to be available.
They then demonstrate that CacheOut can violate nearly every hardware-based security domain, leaking data from the OS kernel, co-resident virtual machines, and even SGX enclaves. CacheOut can bypass software fixes. Making it possible to extract data from both the kernel of the OS and from virtual machines, and also from something that Intel calls 'software guard extensions' (SGX) that normally is stored securely.
Researchers from the University of Michigan and the University of Adelaide have found this new bug, and posted a paper on it. Read the paper. It seems that once again only Intel processors are affected including Core, Xeon and Atom models. AMD is save from this vulnerability.
Senior Member
Posts: 6070
Joined: 2011-01-02
Yes, Intel can't design CPU with security in mind.
It's not about {insert exploit name here} is nearly impossible to reproduce in real-life situation. It's about Intel not paying attention to it.
"Under a certain condition your car brakes might not work, but it's nothing to worry about, since you drive your car as any normal person"
Senior Member
Posts: 11808
Joined: 2012-07-20
sure, you have a strong bottleneck, its a worst sensation than a simple fps flutuation caused by gpu
Btw, youre confused about what vsync and freesync does
I am not confused about those technologies. That's unless you can specifically say what you disagree with and then correct it.
Because I am perfectly aware of all underlying timing functionalities of each technology.
And you may be surprised by fact that 100Hz Free/G-sync screen is incapable to display two consecutive frames in shorter interval than 10ms from each other. (Which creates another minor timing issue if you have average 100fps, but frametimes fluctuate. As frame has to wait till it can be shown for 1ms in situation where two consecutive frames came at 9ms interval.)
Senior Member
Posts: 642
Joined: 2017-08-16
No, 2700x are completelly destroyed even the 3900x hardly you can see beating the 8700k
https://www.techpowerup.com/review/amd-ryzen-9-3900x/15.html
Are you comparing 2700X to the 3900X?
Different architecture, different tier lol. At least if you would compare 2700X to 3700X...and you are showing gaming benchmark, rly mate?
Senior Member
Posts: 13406
Joined: 2013-01-17
Trivial social media hype.
What I don`t understand is why browser can even execute low level CPU instructions (needed for such attacks) executing java-script? How? Why a script language can even emit low level CPU instructions?
Senior Member
Posts: 3580
Joined: 2010-01-16
Does it matter?
Imagine a hacker successfully injected malicious code (exploiting this or any other cache memory vulnerability) into some web page. And some user visited this page allowing hacker to monitor the cache memory reads and writes in real time (until user closed a browser). Do you think it will be easy for a hacker to understand what he sees? Is it even possible to view cache memory operations in real time? Hacker should actually store all the info (big amount - so the channel should be fast) to analyse afterwards, hoping that this dump of cache memory operations contains something useful.
Then no, it doesn't matter. None of it does. Do you see the point behind this exploit hysteria?