New place of work has a work around for av being shit.. They don't seem to run any.
New place of work has a work around for av being shit.. They don't seem to run any.
Originally Posted by lubica
FWIW that's because AMD's CPUs just don't do the thing that Intel's CPUs do that introduces the vulnerability. It's not that they've done it better, they just don't do it. Speculative execution of kernel-space references is one of the major reasons intel's single threaded performance has historically caned AMD's.
But yeah this is going to be a pretty big deal. 10-30% performance hits across the board for cpu bound applications. Yay.
Didn't Ryzen at least hit performance parity with Intel, before they then have to take this hit..? I want to see a lot of benchmarks redone.
And hoping that MS (& Apple) don't do what Intel tried to for Linux, and make all x86 CPUs suffer the slow path rather than detect if it's actually an Intel chip first.
1st, 2nd, 3rd, 4th, 5th, 6th, 7th, and 8th generation Intel® Core™ Processor Families
Intel® Xeon® Processor E3-1200 v5 and v6 Product Family
Intel® Xeon® Processor Scalable Family
Intel® Xeon® Processor W Family
Intel Atom® C3000 Processor Family
Apollo Lake Intel Atom® Processor E3900 series
Apollo Lake Intel® Pentium® Processors
Intel® Pentium® Processor G Series
Intel® Celeron® G, N, and J series Processors
Did Intel release a statement?
Or where did you get that from?
The Rapier is my love boat
~lowsec smallscale pvp 'n stuff~
GGWP
wtf lads
Viking, n.:
1. Daring Scandinavian seafarers, explorers, adventurers, entrepreneurs world-famous for their aggressive, nautical import business, highly leveraged takeovers and blue eyes.
2. Bloodthirsty sea pirates who ravaged northern Europe beginning in the 9th century.
Hagar's note: The first definition is much preferred; the second is used only by malcontents, the envious, and disgruntled owners of waterfront property.
LOLintel
2/10/17 Greatposthellpurge never forget
23/10/17 The Greatreposteninging ?
Intel says memory security issue extends beyond its own chips
https://www.engadget.com/2018/01/03/...c-to-its-cpus/
There's been 2 updates to that article since, 1st being AMD denying Intel's claims
2nd being an apparently early ending of the embargo over Meltdown and Spectre. https://meltdownattack.com/Update: AMD isn't having Intel's claims that the issue is hardware-independent. In its own statement, it asserted that architecture differences meant that there was "near zero risk" to AMD-made processors. That lines up with the initial report, which referenced communication from AMD suggesting that its processors weren't vulnerable. There's clearly a he-said-she-said dispute going on, and it may be a while before we get the full story. You can read the full statement below.
Glad my newest processor is for vidya only. Plz don’t be affected too bad. Plz don’t be affected too bad.
Sent from my iPhone using Tapatalk
https://twitter.com/i/moments/948681915485351938
tl/dr: all cpus out there are fucked, intel's is actually the less serious one since there's a software fix for it. On the other hand the second affects ALL cpus, and is impossible to fix without tossing modern cpu design in the bin.
edit: someone else's post #258 was deleted. This was mostly a response to that, which disagreed with I Leg's TL;DR, and I do too:
Read the Project Zero papers, linked on the very public & promoted ...attack.com website(s), rather than the non-technical summary blogs, or just some NY Times journo's twitter.
Spectre might be able to be fixed in microcode and/or OSs on loading executables (auto-insert judicious cache flushes & fences/memory barriers, yo), and for Zen it isn't tested/proven as being vulnerable, but AMD claimed it has a new, more complex speculative execution system (the concept that's being found to be flawed) so it could be at least as vulnerable as other vendors (spectre.pdf, page 12, section 8, 3rd para).
Some of the 'possible' variations in section 6 are rather theoretical-only atm. E.g. the branchless variation where an interrupt is what combines with the mistrained branch predictor to then mistakenly access and leak a register value used as a memory address (who's lookup is cached and thus the reg value can be deduced), in otherwise unbranching code that would normally have no speculation.
I don't have it 100% absorbed/correct, it's late and my MIPS background keeps screaming 'but we already had kernel-only address space & ASID since day1, and pipelines aren't ~200 instructions deep for RISC, and we've been working with & regretting manually managed caches & TLBs for 30 years already".
Seriously, fuck x86.
And fuck Intel trying to keep painting AMD with all these vulns too. Especially that CEO, on TV, after recently selling down to his minimum required shares.
tbh, if this turns out to be the death-knell of x86 then it's about time.
Viking, n.:
1. Daring Scandinavian seafarers, explorers, adventurers, entrepreneurs world-famous for their aggressive, nautical import business, highly leveraged takeovers and blue eyes.
2. Bloodthirsty sea pirates who ravaged northern Europe beginning in the 9th century.
Hagar's note: The first definition is much preferred; the second is used only by malcontents, the envious, and disgruntled owners of waterfront property.
Nobody at my work, not even ops, is talking about this. I'm disappointed. This is a big deal.
AWS prices gonna go up?
Originally Posted by Loire
Bookmarks