|
ComradeCosmobot posted:OpenCL and CUDA? oh god this means buttcoin is going up again
|
#
?
Jan 3, 2018 20:38
|
|
|
|
| # ? Apr 25, 2024 13:07 |
|
|
lol butt miners will just run with nopti
|
|
#
?
Jan 3, 2018 20:40
|
|
|
so what happens if there is some problem that affects like decades of old intel CPUs that is game over unless the system is patched
|
|
#
?
Jan 3, 2018 21:04
|
|
|
a resurgence of abacuses
|
|
#
?
Jan 3, 2018 21:18
|
|
|
Lol Intel's response: there's no risk or flaw and it's not only our products well okay then
|
|
#
?
Jan 3, 2018 21:22
|
|
|
like i imagine there must be a zillion old pentiums running in the control systems for a factory or a power grid or whatever, does everyone just throw up their hands and say oh well, or what
|
|
#
?
Jan 3, 2018 21:30
|
|
|
Oh also  "We do not believe this exploit will let anyone corrupt, modify or delete data" >You are in a user mode process. You can corrupt nothing, modify nothing, delete nothing, admit nothing, and are not even certain who you are. >read 
|
|
#
?
Jan 3, 2018 21:32
|
|
|
if they're old they're safe. 
|
|
#
?
Jan 3, 2018 21:32
|
|
|
flakeloaf posted:Oh also it’s true, it’ll just let the process read how to do those things!
|
|
#
?
Jan 3, 2018 21:35
|
|
|
our faulty fire extinguishers pose no risk of drowning, functional illiteracy or guinea worm
|
|
#
?
Jan 3, 2018 21:37
|
|
|
don't worry, we won't let anyone corrupt, modify, or delete your private bitcoin address *put giant QR code printout inside clear plastic sleeve, keeps face up on desk*
|
|
#
?
Jan 3, 2018 21:40
|
|
|
Linguica posted:like i imagine there must be a zillion old pentiums running in the control systems for a factory or a power grid or whatever, does everyone just throw up their hands and say oh well, or what Accept the risk, only run trusted code.
|
|
#
?
Jan 3, 2018 21:54
|
|
|
Linguica posted:so what happens if there is some problem that affects like decades of old intel CPUs that is game over unless the system is patched yes. if you didn’t pay for a long term support contract you’re not getting bug fixes for ancient hardware. that’s life.
|
|
#
?
Jan 3, 2018 21:56
|
|
|
The Management posted:where intel is going to get hosed is their data center customers. amazon, google, and Microsoft are not going to like getting significantly less performance from the chips that they bought, and they buy the highest end most expensive chips. they are going to demand some big discounts to make up for this. But Amazon, Google, and Microsoft don't have much choice as to CPU provider in the short to medium term, no?
|
|
#
?
Jan 3, 2018 22:01
|
|
|
they'll keep the chips and browbeat intel into a fat settlement cheque, and the public can continue to gently caress off
|
|
#
?
Jan 3, 2018 22:05
|
|
|
The Management posted:yes. if you didn’t pay for a long term support contract you’re not getting bug fixes for ancient hardware. that’s life.
|
|
#
?
Jan 3, 2018 22:10
|
|
|
Linguica posted:right but will there have to be a big Y2K type effort to fix old poo poo, except in this case the enemy is possible malicious actors and not just the transition to a particular year or do people just accept their poo poo might get hacked at any given time in the future lomarf people already accept this, haven't you seen the ccc skada talk?
|
|
#
?
Jan 3, 2018 22:14
|
|
|
The Management posted:yes. if you didn’t pay for a long term support contract you’re not getting bug fixes for ancient hardware. that’s life. also even if you did pay for a long term support contract because the vendor of your hardware decides this is outside the scope of what they support see: every piece of CNC equipment ever
|
|
#
?
Jan 3, 2018 22:14
|
|
|
Linguica posted:right but will there have to be a big Y2K type effort to fix old poo poo, except in this case the enemy is possible malicious actors and not just the transition to a particular year or do people just accept their poo poo might get hacked at any given time in the future Mate, security is so poo poo in the entire scada sector that this is the least of anyone's problems
|
|
#
?
Jan 3, 2018 22:16
|
|
|
flakeloaf posted:they'll keep the chips and browbeat intel into a fat settlement cheque, and the public can continue to gently caress off except they don't have all that much leverage to get such a settlement since they, as noted, don't have a lot of other options
|
|
#
?
Jan 3, 2018 22:23
|
|
|
Anyone know what subset of Intel chips have PCID? Ark doesn't list the feature and all I can find googling it are articles about the fuckup
|
|
#
?
Jan 3, 2018 22:26
|
|
|
the only way this will effect bideo james are online bideo james’ servers, and then not really since it depends on their scheme I suspect MMOs like Eve or WoW would be the hardest hit while FPS games will be mostly fine unless they’re doing something weird and unnecessary
|
|
#
?
Jan 3, 2018 22:26
|
|
|
Fiedler posted:But Amazon, Google, and Microsoft don't have much choice as to CPU provider in the short to medium term, no? not really but that doesn’t mean they don’t have leverage. refusing to pay multi billion dollar invoices for underperforming parts, refusing to accept new deliveries and leaving intel with billions in inventory on their hands, and threatening lawsuits over contract breach of chip performance guarantees are tools they can use. Intel is a supplier, they need to keep their customers happy, and when they obviously hosed up they’re going to have to eat that cost.
|
|
#
?
Jan 3, 2018 22:26
|
|
|
Preliminary benchmarks against the Epyc platform look pretty good compared to Xeon and the biggest issues seem to be software needing architecture optimizations. I think they could have some leverage
|
|
#
?
Jan 3, 2018 22:27
|
|
|
FAT32 SHAMER posted:FPS games will be mostly fine unless they’re doing something weird and unnecessary hahaha
|
|
#
?
Jan 3, 2018 22:29
|
|
|
FAT32 SHAMER posted:the only way this will effect bideo james are online bideo james’ servers, and then not really since it depends on their scheme Absolutely terrible news that Excel Online: The Game will be a little bit slower
|
|
#
?
Jan 3, 2018 22:29
|
|
|
Truga posted:hahaha I bet BattleEye will poo poo its pants over this
|
|
#
?
Jan 3, 2018 22:30
|
|
|
Truga posted:hahaha wouldn’t it only affect servers that are keeping track of items across massive maps/worlds? so it may affect something like PUBG but not so much something like CSGO or Overwatch I’m asking questions based on what I saw about databases but I’m literally retarded so I’d love to know if I’m understanding this wrong 
|
|
#
?
Jan 3, 2018 22:31
|
|
|
things are finally looking up for ol’
|
|
#
?
Jan 3, 2018 22:32
|
|
|
FAT32 SHAMER posted:the only way this will effect bideo james are online bideo james’ servers, and then not really since it depends on their scheme EVE is probably legit hosed give the nature of how that game is programmed, afaik
|
|
#
?
Jan 3, 2018 22:33
|
|
|
LethalGeek posted:EVE is probably legit hosed give the nature of how that game is programmed, afaik 
|
|
#
?
Jan 3, 2018 22:35
|
|
|
The Management posted:not really but that doesn’t mean they don’t have leverage. refusing to pay multi billion dollar invoices for underperforming parts, refusing to accept new deliveries and leaving intel with billions in inventory on their hands, and threatening lawsuits over contract breach of chip performance guarantees are tools they can use. Intel is a supplier, they need to keep their customers happy, and when they obviously hosed up they’re going to have to eat that cost. Yeah, I don't think so. Amazon and Microsoft don't have enough spare data center capacity to play a game of chicken with their CPU supplier.
|
|
#
?
Jan 3, 2018 22:36
|
|
|
FAT32 SHAMER posted:wouldn’t it only affect servers that are keeping track of items across massive maps/worlds? so it may affect something like PUBG but not so much something like CSGO or Overwatch yeah, that's what should happen. but most games have become gigantic piles of poo poo in the last 20 years, and apart from some notable exceptions run despite all the work done on them, not because of it. i'm not going to call games will turn to poo poo over this, obviously most won't, but i also won't be surprised in the least if some game that's basically quake 2 but 20 years later turns to poo poo over this patch. for example, the denuvo DRM has a driver that lives in the windows kernel these days, through which the majority of gameplay code most go, to try and prevent people from stealing a $40 piece of software. anti-cheat software like battleye as pointed up above could also get hit by this, and games like ARMA 3 already depend cpu or ram speed enough that for many people, playing on low or ultra settings gives the same framerate.
|
|
#
?
Jan 3, 2018 22:46
|
|
|
COACHS SPORT BAR posted:Anyone know what subset of Intel chips have PCID? Ark doesn't list the feature and all I can find googling it are articles about the fuckup goes back pretty far it looks like; my old cpu from like 7 years ago has it: http://openbenchmarking.org/system/1204224-SU-I73770KCO52/Intel%20Core%20i5%202500K/cpuinfo
|
|
#
?
Jan 3, 2018 23:16
|
|
|
COACHS SPORT BAR posted:Anyone know what subset of Intel chips have PCID? Ark doesn't list the feature and all I can find googling it are articles about the fuckup my skylake has it `grep pcid /proc/cpuinfo´ tells you pretty quick at least
|
|
#
?
Jan 3, 2018 23:18
|
|
|
afaik every haswell and later cpu has it, before that it may be a thing that's in some but not others?
|
|
#
?
Jan 3, 2018 23:22
|
|
|
deets: https://security.googleblog.com/2018/01/todays-cpu-vulnerability-what-you-need.htmlquote:These vulnerabilities affect many CPUs, including those from AMD, ARM, and Intel, as well as the devices and operating systems running them. e: https://meltdownattack.com/ (also thanks scaevolus and b0lt 4 the links) Avenging Dentist fucked around with this message at 23:28 on Jan 3, 2018 |
|
#
?
Jan 3, 2018 23:24
|
|
|
the fact that there's an associated security update for chrome is a bit concerning
|
|
#
?
Jan 3, 2018 23:31
|
|
|
Plorkyeran posted:the fact that there's an associated security update for chrome is a bit concerning from my very brief reading, that's probably for the spectre attack, which is different from the intel one (meltdown). spectre sounds like it's something you have to fix per-application the spectre paper won't load tho e: more stuff, including how spectre works https://googleprojectzero.blogspot.com/2018/01/reading-privileged-memory-with-side.html Avenging Dentist fucked around with this message at 23:41 on Jan 3, 2018 |
|
#
?
Jan 3, 2018 23:33
|
|
|
|
| # ? Apr 25, 2024 13:07 |
|
|
Any "background" section of a paper about an extremely bad exploit that includes the phrase "In 1967," should be extremely concerning
|
|
#
?
Jan 4, 2018 00:00
|
|