[ / / / / / / / / / / / / / ] [ dir / ausneets / ck / f / fur / kpop / pdfs / strek / thestorm ][Options][ watchlist ]

/tech/ - Technology

You can now write text to your AI-generated image at https://aiproto.com It is currently free to use for Proto members.
Name
Email
Subject
Comment *
File
Select/drop/paste files here
* = required field[▶ Show post options & limits]
Confused? See the FAQ.
Expand all images

File (hide): 869777528291cef⋯.jpg (216.44 KB, 1366x2048, 683:1024, DGNJvv2VwAAM0EP.jpg) (h) (u)

File (hide): 597638f8c6c7f40⋯.jpg (84.57 KB, 400x300, 4:3, 1448952872633.jpg) (h) (u)

[–]

 No.847967>>847972 >>847988 >>848004 >>848050 >>848103 >>848297 >>848349 >>848396 >>848419 >>848465 >>848498 >>848993 [Watch Thread][Show All Posts]

httpz://googleprojectzero.blogspot.ca/2018/01/reading-privileged-memory-with-side.html

httpz://meltdownattack.com/

<Death of x86

<Death of botnet

<Wincux eternally cucked by x86

<Rebirth of IA-32/64 market and Atom (pre 2013) gets massive overhaul

<Year of the *nix Desktop finally

<Year of the Linux Gaming

<Year of the Open Hardware

<Year of the FOSS VN

<Year of Best Animu

<IBM POWER, et alia

2018 best year

 No.847972

>>847967 (OP)

><Death of x86

><Death of botnet

Not unless normies stop buying it.


 No.847988>>848626

>>847967 (OP)

Good thing, but the first benchmarks shows 2-3% decrease in performance in the W10.


 No.847992>>848004

>Death of x86

>Death of botnet

i think it's only going to get worse. intel's answer to this is going to be to shove even more botnet in their chips.


 No.848004>>848080

>>847967 (OP)

><Death of x86

Hate to break this to you but no one is going to want to dump the past 30+ years of legacy x86 code over this.

>>847992

>i think it's only going to get worse. intel's answer to this is going to be to shove even more botnet in their chips.

This


 No.848042>>848045 >>848050

Normies stopped using x86 ages ago gramps


 No.848045>>848046 >>850608

>>848042

>He doesn't know he's running an x86-64 system


 No.848046>>848050 >>848200

>>848045

Normies use smartphones/tablets not stoneage tier computers


 No.848050>>848051 >>848054 >>848080

>>848046

>>848042

There are 1.5 billion x86_64 computers in use right now, and the PC market growth is being driven by more powerful laptops. The "Post-PC" meme was investor PR because it was not expected that the PC market was going to experience anymore significant growth after the start of the 2010s. It never had anything to do with PCs "dying" a Smartphone is never going to replace a PC. There are more Smartphones on the planet because Smartphones are cheaper and more expendable, and a single person may upgrade to a new Smartphone every year while keeping the same PCs for roughly 5 years at a time (A full hardware cycle for PCs in the industry is generally considered closer to 10 years in many cases)

>>847967 (OP)

And nobody is going to switch to an architecture that has virtually no security auditing in comparison to x86. Please enter reality and stop LARPing


 No.848051>>848053

>>848050

>wangblows is secure because it has a lot of security auditing


 No.848053>>848054 >>848282

>>848051

Windows is proprietary so we do not actually know how much security auditing it receives compared to open source OS' like Linux, so your already non-sequitur already makes no sense. Get bent LARPers


 No.848054

File (hide): dcb4f272861878c⋯.jpg (21.91 KB, 500x499, 500:499, 1417894937877.jpg) (h) (u)

>>848050

>>848053

>LARPER calls other people LARPERs

i love this board


 No.848080>>848083 >>848084 >>848123 >>856538

>>848004

>what ARM is doing right now

x86 is tumbling down the same path as Java and Flash.

There's literally no advantage in using x86 other than you have more performance over ARM at the cost of crunching more TDP.

>no one is going to want to dump the past 30+ years of legacy x86 code over this

too much SAP? sanjay?

I bet you're masochist enough to like those little incermental performance intel does which only holds us all back from breakthroughs.

Google will wreck the shitty market now that they'd been making stuff that's cross compilable on multiple architectures with portage. This enables a good fucking game among architectures.

>>848050

>Smartphone is never going to replace a PC

A $50 android tv box comes with 3GiB DDR4, supports h.265|VP9 at 1080p60 comes with a cool qwerty and gyro remote. All that with passive cooling while PCs or even laptops in year 2005 can't do the same amount of "non-retro" entertainment.

Stop thinking x86_64 userbase care about architectures. Majority of people want cheap stuff that just works anywhere and don't have the time to build their own PC gaming console. Even a 4K smartphone can be way cheaper than a 4K compatible PC. That gaymer monitor would even cost as much as that 4K phone that can work interoperably with smart tv's and other phones.

>There are more Smartphones on the planet because Smartphones are cheaper and more expendable, and a single person may upgrade to a new Smartphone every year while keeping the same PCs for roughly 5 years at a time (A full hardware cycle for PCs in the industry is generally considered closer to 10 years in many cases)

Not only that they're expendable but also efficient. You can use and bring them anywhere and anytime unlike on PCs where the only time you can use is if you do desk work or during break, go home, or is some hikki or homebased prog.


 No.848083>>848088 >>848108

>>848080

>Not only that they're expendable but also efficient.

8 Single Precision Instructions Per Cycle is not "Efficient". ARM would have identical efficiency problems as x86 if they made ARM as heavily pipelined

>Stop thinking x86_64 userbase care about architectures.

Users do care about legacy compatibility, part interchangeability, and software upgradability. Which Smartphones lack


 No.848084>>848108

>>848080

>There's literally no advantage in using x86 other than you have more performance over ARM at the cost of crunching more TDP.

Indeed, now if only they would make a architecture more effeicient then (((ARM))) and not backdoored everywhere.

>jewgle

>using gentoo/portage

OHHHH that makes so much more sense now. They must be the ones putting zeromq into ffmpeg.

>A $50 android tv box comes with 3GiB DDR4, supports h.265|VP9 at 1080p60 comes with a cool qwerty and gyro remote.

It's also full on botnet and spying on you 24/7. Unless you are offering a alternative the user can control then you are pandering to the general web browsing crowd who doesn't care about architecture. I don't care about architecture either, but I want to control the hardware I purchase lest something like planned obsolesence with apple batteries happens.


 No.848088>>848090 >>848091 >>848108

>>848083

>Users do care about legacy compatibility,

No most users care about what Microsoft and Apple tell them to care about. They eat up the latest shit no matter how painful it is.

Samsung, HP, Asus and a few others are bringing Windows 10 laptops to the market.

CromeOS has always been there.

Now there are rumblings about Apple coming out with a laptop powered by their A11 chip.

x86 has jumped the shark.


 No.848090>>848094

>>848088

>bringing Windows 10

*brings ARM Windows 10 laptops


 No.848091>>848114

>>848088

Except nobody is buying Chromebooks, even Windows ARM laptops have been commercial failures. The average normalfag is retarded, but even they wouldn't want to spend their hard-earned cash on a literal Facebook browsning machine or a version of Windows that can't run their favorite games. the only delusional one is you and if you really want the success of alternative architectures to rely on herding normalfags then you need to get fucking bent you retarded LARPing piece of shit


 No.848094>>848108

>>848090

>brings ARM Windows 10 laptops

>ARM "Windows" laptop

Great, I'm gonna install Photoshit, MS Office and muh games on one of those.


 No.848103>>848109

>>847967 (OP)

The patch for Meltdown will fuck over enterprise users, but the generally population won't see a difference in performance. Pretty much every modern CPU architecture is vulnerable to specter.

Only thing I see happening in the short run, is AMD gaining some marketshare in the enterprise market. In the long run, it all depends what the big three do going forward. Whatever happens, we're all fucked anyway.


 No.848108

File (hide): a91cd3995598d6b⋯.png (10.87 KB, 653x367, 653:367, ms_loves_linux.png) (h) (u)

>>848083

>8 Single Precision Instructions Per Cycle is not "Efficient". ARM would have identical efficiency problems as x86 if they made ARM as heavily pipelined

Still werks just as efficient in mobility and being a RISC isn't all disadvantage, it's a completely different architecture. ARM and other arch just doesn't care about the desktop market yet because the phone and server market is very solid.

>Users do care about legacy compatibility, part interchangeability, and software upgradability. Which Smartphones lack

Smartphones come with warranty which is what pure FOSS will always lack unless it is red hat shipping some sort of expensive enterprise android but red hat is suspicious as their logo

After the warranty is expired, it's probably beyond repair and no one have enough pajeet support in android as windows. Smartphones evolving a lot is a good sign that the market enables performance competitions among disparate tech companies or OEMs.

>>848084

>It's also full on botnet and spying on you 24/7.

China is profiting in the other end of a stick but you get the hardware at a reasonable price. Open it up and search the numbers on the stickers and pcb. Or steal the source code on some Shenzen public /index or factory.

>>848088

This. pretty sure Windows will try to integrate into Linux now that they have systemd. What if Windows had some contract with intel to only shill and produce x86? They'll crash together, if true.

>>848094

Programs always follow the userbase. Give it time and even your OS will follow the userbase and provide those services in Linux.


 No.848109

>>848103

No sense bringing logic and reason into this. LARPers are getting a field day with this one and the next few weeks on here will be the same 2 or 3 retards shilling for PowerPC and ARM Laptops that want to lock you into an ecosystem harder then the average Smartphone OS because they have mental gymnastics to perform over why they need to hate an architecture for no reason


 No.848114>>848117

>>848091

>Windows ARM laptops have been commercial failures

Expensive SURFACE laptops have been a failure.

Once Samsung, HP, Asus and all the other shitslingers start pumping cheap ARM laptops with the Windows sticker on them in to Walmart the normies will lap them up.

>MUH GAMES

Candy Crush? Facebook games? Shit like that will work fine.

Actual gamers will stick to x86 and their home built machines. But their options will slowly diminish as the industry moves on.

Call me a larper all you want. I don't use windows or x86 so I couldn't give a single fuck what happens.


 No.848117>>848134 >>848198

>>848114

>Call me a larper all you want. I don't use windows or x86 so I couldn't give a single fuck what happens.

Fine, don't bring your delusions or cancer here though, actually too late, it's impossible to discuss any kind of technology here without you sacks of shit activating SHUT IT DOWN mode for /tech/ wrongthink

>Once Samsung, HP, Asus and all the other shitslingers start pumping cheap ARM laptops with the Windows sticker on them in to Walmart the normies will lap them up.

No they will absolutely not because they have no reasons too. Its too big a risk to take when, like you said, normalfags don't care about the architecture, and companies like HP are more than comfortable pushing out low-powered Intel Atom and Pentium machines for low price points because Normies actually expect their software to run.

>Candy Crush? Facebook games? Shit like that will work fine.

Grand Theft Auto V is the best selling video game of all time, it's 70GBs in size, and the majority of its sales have been on some x86 machine, wither thats consoles or on PC. the original versions accounted for only 30 million of the total 80+ million sold. It just goes to show your sheer disconnect from reality


 No.848123>>848128

>>848080

>x86 is tumbling down the same path as Java and Flash

You mean how companies are trying to get x86 emulation on ARM? I will believe it when I see it working flawlessly. When Apple switched from Power to x86 they tried it as well and lets just say it wasn't exactly smooth.

>I bet you're masochist enough to like those little incermental performance intel does which only holds us all back from breakthroughs.

I like them, not because I am a masochist but because I am a GPU programmer. I still enjoy Moore's Law performance gains.

Also its not entirely the fault of Intel but rather the fact that whenever they release a new architecture they have to make sure it supports legacy code.Nvidia (and AMD) doesn't have this issue to the same extent because of the way GPU drivers work.


 No.848128>>848130

>>848123

>You mean how companies are trying to get x86 emulation on ARM? I will believe it when I see it working flawlessly.

Whenever anyone gets too close Intel lawyers start looming, we will see.

https://www.theregister.co.uk/2017/06/09/intel_sends_arm_a_shot_across_bow/


 No.848130

>>848128

It would never have gone anywhere anyways because emulators can't magically turn 8 SP IPC into 32 SP IPC


 No.848134>>848138 >>848141 >>848146 >>848423

File (hide): 6da17a1806918fb⋯.png (479.55 KB, 770x565, 154:113, Screen Shot 2018-01-04 at ….png) (h) (u)

File (hide): 43426fecec82ac5⋯.png (317.82 KB, 757x304, 757:304, Screen Shot 2018-01-04 at ….png) (h) (u)

>>848117

>No they will absolutely not because they have no reasons too. Its too big a risk to take when

http://windows-arm.com/index.php/hw-news/1096-windows-10-on-arm-laptops-coming-from-samsung-and-xiaomi

https://arstechnica.com/information-technology/2016/12/microsoft-pushes-for-a-more-modern-pc-with-vr-project-evo-and-arm-support/

https://www.theverge.com/2017/12/5/16734402/windows-10-arm-snapdragon-laptop-hp-asus-lenovo-hands-on

>Earlier today, Qualcomm and Microsoft announced the first Windows 10 devices with ARM-based processors will be arriving early next year. HP, Asus, and Lenovo will all be launching laptops and convertibles with Qualcomm’s Snapdragon 835 processor in them, with prices expected to range from $600 to $800.

More profit in ARM. And any chink fab can make an ARM CPU. Why be at the mercy of INTEL/AMD and pay their markup? Microsoft and Apple will tell normies what to buy. And they will buy it. They always do.

The whole industry is not games. Maybe someday you will grow up and realize that.

Keep your hand in the sand and stay mad.


 No.848138>>848148

>>848134

>Snapdragon 835

>600 to 800 dollars

I look forward to seeing absolutely nobody buying them then. Normalfags will see a less-portable Smartphone

Companies don't give a fuck who sources their chips. They want to close you into their ecosystem. And you're defending them because of your cognitive dissonance against a computer architecture

And you dare tell me to grow up, hilarious'!


 No.848141>>848148

>>848134

>Yes goy! Buy these superior ARM processors! You don't need platform standards or open bootloaders goy! It's different thus its good goy! Only 800 dollars for a Smartphone chip goy!

You better fucking hope these flop if you really don't like Jewery, fucking kike


 No.848146

File (hide): 70cdbecb893e94c⋯.gif (320.67 KB, 606x423, 202:141, 83ee5490524e49dfec47e436da….gif) (h) (u)

>>848134

>hehe take that Intel we have BIGGER BOTNETS now!


 No.848148>>848149 >>848152

>>848138

>And you're defending them because of your cognitive dissonance against a computer architecture

How am I defending them? They are coming.I have no control over it. You can keep denying ARM's growth and think x86 will be forever that's fine. I am the one full of cognitive dissonance sure.

>>848141

>You better fucking hope

Hopes and prayers lel. Ya keep HOPING that computers don't get more and more kiked. See how well that works out for ya goy. ARM gives them to much lock-in power,top to bottom botnet, and massive profit potential. I am sure they will totally pass that by.


 No.848149

>>848148

Became an investor please, I would love to see you go bankrupt from your complete lack of foresight


 No.848152>>848154

>>848148

All that will happen is that these ARM laptops will dissolve into another segment of the Smartphone market. There is a good reason x86-64 is the main architecture used for workstations, gaming PCs, and servers. I really don't think you understand how the industry works


 No.848154>>848156 >>848158 >>848211

>>848152

It's herd mentality actually. Very easy to understand. Jewgle says go this way and the entire industry goes. Now if only they would all give up legacy vidya and keep a x86 machine for the legacy. While everyone switches to something like RISC-V.


 No.848156>>848160 >>848168 >>856602

>>848154

It's not herd mentality. Engineers use x86 because of its instruction density, servers use x86 for legacy compatibility and performance, game consoles use x86 now because it eases burden on developers who use x86, along with gamers in general as well as workstation users, because of x86s superior performance

Devs want you to buy ARM because they want to lock you into their ecosystem and they feel like they can sell you on battery life. But you're out of your fucking mind if you think these toys will ever replace actual PCs


 No.848158>>848159 >>848165

>>848154

>Jewgle says go this way and the entire industry goes.

Holy fuck you are dumb.


 No.848159>>848165

>>848158

He has to feed his cognitive dissonance somehow because /tech/ told him an architecture is botnet


 No.848160>>848161 >>848163 >>848167

>>848156

SoCs replaced the x86 server market years ago. Do you still think the big servers still use xeon? They don't. They roll out multiple and scalable SoCs and implement MapReduce.

Even neural netwerks follow similar practice.


 No.848161>>848173

>>848160

>SoCs replaced the x86 server market years ago. Do you still think the big servers still use xeon? They don't. They roll out multiple and scalable SoCs and implement MapReduce.

I'm just going to go out on a limb here and say no, the largest server providers on the planet like Amazon and Microsoft are not rocking Snadragon 320 server racks


 No.848163>>848173

>>848160

Straight up delusional. This is the most Pajeet thing I've read today


 No.848165>>848166 >>848167

>>848158

>>848159

>what is android

Google is way beyond its market play. How can cucked companies even compete?

>will never adapt feminist 'empowerment' agenda

>intelligent choice for portage for migrating server architecture to power years ago

>helped cuck intel by finding the exploit

>pioneered material design that cucked windows modern/metro

>supreme botnet overlord


 No.848166>>848295

File (hide): 0b254025fead317⋯.png (574.54 KB, 1403x1408, 1403:1408, 0b254025fead317cb7eb309fa5….png) (h) (u)

>>848165

This.

>Not running a webserver and video editing environment on your Android phone while compiling the latest Android AOSP™ Kernel

How can Microcuck and Shitel ever compete?


 No.848167>>848173

>>848160

>SoCs replaced the x86 server market years ago.

Just stop please, you clearly have no fucking idea what you are talking about.

>>848165

>>what is android

A product that Google offered companies like Samsung as a way to enable them to compete with Apple in the smartphone market.


 No.848168>>848176 >>848177 >>848216

>>848156

>But you're out of your fucking mind if you think these toys will ever replace actual PCs

PCs? Like desktop PCs? Who the fuck still buys desktop PC other then gamers? That market is a shell of what it once was.

Most of the low end bigbox store desktops are just laptop-tier boards stuck in a big hollow box. Everything is moving to (((cloud))) computing so the thing on your desk in the office is just a thin client now.

Google and all the other big cloud players are all dipping their toes in to non-86 archs because power and cooling costs.

http://www.datacenterknowledge.com/hardware/red-hat-bets-data-centers-are-ready-arm-servers

https://www.bloomberg.com/news/articles/2017-03-08/microsoft-pledges-to-use-arm-server-chips-threatening-intel-s-dominance

x86 will forever control the server market if you think energy costs will never rise.


 No.848173>>848202

>>848161

>>848163

>thinks quantity is relevant when the most "relevant" web userbase and traffic goes to Google mainly.

>performance share

>pie chart definitely not VLC'd (tm) they say so I(you) says so too

Yeah, the server gods will definitely use highly scalable Xeon to make the strongest housefire.

>>848167

Google offered android free of charge and made internal competitions among OEMs to:

a c c e l e r a t e

No one's stopping the train except salty x86 faggots.


 No.848176>>848202 >>848206

>>848168

That was early last year, and it never happened. Because Intel Atom exists on the low-end and ARM experoences identical scaling issues on the high-end to x86

Please just stop


 No.848177>>848202

>>848168

Okay Pajeet I'll switch to an Allwinner for my next host


 No.848187>>848204

So now that you guys are actually angry with this x86 fukushima incident can you PLEASE start petition to open up ARM mali GPU? I don't think libv will be able to write entire gallium library on his own with recent paranoia attack and now is the chance to change ARM's mind on laptop market strategy


 No.848198>>848204 >>848206

>>848117

uhhh can't we just, y'know... recompile whatever software we need into arm? :thinking:

and i'm sure there isn't too much inline stuff


 No.848200>>848201 >>848202 >>848204 >>848409

>>848046

Most smartphones/tablets are x86 now too.


 No.848201>>848203

>>848200

really? Most? then where is arm making their money off of?


 No.848202

>>848176

>identical scaling issues

>cites nothing

>can't into performance/price - recurring cost

>believes in benchmarks, and pie graphs from viral blogs

>possibly another zenfone fan-gay from cuckchan

>>848177

>can't argue with >>848173 so I'll call names and mention some entirely obscure CPU

>>848200

>another zenfone fan-gay from cuckchan


 No.848203

>>848201

Mostly rasperg pis. ARM is dwindling pretty rapidly in every other market.


 No.848204>>848207

>>848198

lel, no. Wish it was that easy to switch between architectures. Why do you think companies have entire teams that are responsible for different platform versions.

>>848187

You do know some ARM versions are vulnerable to both Meltdown and all are vulnerable to Specter.

>>848200

There is no mainstream phone that uses x86.


 No.848206

>>848176

>Just spouting nonsense to make it look like I know some stuff

>scaling issues

Image scaling? lmao. You have to stop.

>>848198

You can but various engines and other legacy crap are dependent on the CPU instructions (games).

Google ports apps on both x86 and ARM although x86 doesn't usually have the best support.


 No.848207>>848387

>>848204

my pineboard is not affected by spectre as it does not employ branch prediction too deep into pipeline (A53).

Some processors that are affected by Meltdown are actually only leaking system registers that do not normally contain secret data. For boards that store cryptographic key are not affected by meltdown.

https://developer.arm.com/support/security-update

I'm still trying to write my own PoC just in case ARM is lying tho. Bit more tricky than my x86 dual core as there's no clear documentation on how deep in the cache does gcc's flush intrinsics actually clear


 No.848211>>848212

>>848154

>Jewgle says go this way and the entire industry goes.

Ah, yes, just like how everyone uses chromebooks nowadays


 No.848212

>>848211

Well at least it seems like they are good at consistently finding zero days that fundamentally change how some softwares are written


 No.848216

>>848168

>Cloudfare

Oh shit guys, none other than fucking Cloudfare, with their computationally intense workloads and massive industry dominance, is switching to ARM based systems. x86 is clearly done for now!

>MIcrosoft

>x86 will forever control the server market if you think energy costs will never rise.

Of course it won't, but you are a fool for thinking its going towards an ARM-centric paradigm.


 No.848274>>848280 >>848309 >>848347 >>848382 >>848619

>security fix results in 30% performance hit

>turns out this shit affects every architecture

>more security flaws found

>more security fixes

>more slow downs

>increasingly crippled cpu's

>common applications begin to slow down

>as the months go by the numerous security fixes have made firefox, facebook, and youtube totally unusable

>chip designers can't make a cpu without defects, they're too dumb

>even the fastest cpu's are castrated by the security fixes

>everyone finally sees the "computers" as the festering piles of shit that they are

PLEASE FUCKING HAPPEN


 No.848280>>848285

>>848274

Or

>people give up security all together

>microsoft doesn't patch security features, afaird they'll upset consumer base

>normalfags computers chocked full of botnet and spyware

>Linus spergs out about security, bans any security pacthes to the kernal

>linux becomes backdoored too

>major web services all collectively become fucked


 No.848282>>848283

>>848053

How is auditing those binary blobs coming? inb4 you larp as a trisquel user


 No.848283

>>848282

>not using gnu's deblobbed linux kernel


 No.848285

>>848280

>16-bit computing becomes mainstream again

>BBS and IRC become "the internet" for normalfags

>eventually Half-Life 3 is released, after speeds get faster

>it looks like Half-Life 1


 No.848295>>848298 >>850600

>>848166

>ancap

>trademarks

corporatism is not capitalism


 No.848297

>>847967 (OP)

>Year of the *nix Desktop finally

>Year of the Linux Gaming

It was in 2012 (steam on linux), stop wearing diaper you underage


 No.848298>>848306

>>848295

That's the joke.


 No.848306>>850600

>>848298

I thought the joke was

>ancap


 No.848309>>848620

>>848274

I feel like I’m already seeing it. Could be confirmation bias but I was in a fairly large subreddit and it gave me a “too busy” error, also YouTube errored out on me a couple time today too. This was all on a corporate network with a ton of bandwidth.


 No.848347

>>848274

Or, alternatively, they do this and go back further and further and "find" new security problems with old chips, then slowing them down to "fix" the problems so you have to buy new chips that are 100% secure botnet.


 No.848349>>848375 >>856740

File (hide): c41162769005a76⋯.jpg (34.66 KB, 490x361, 490:361, A1000.jpg) (h) (u)

File (hide): cd586f24f919206⋯.jpg (80.38 KB, 656x597, 656:597, miho scared.jpg) (h) (u)

>>847967 (OP)

I-is M68k affected too?


 No.848375

>>848349

The ones used in Amigas should be fine.


 No.848382

>>848274

>security fix results in 30% performance hit

It's funny that /tech/ pretends to be more educated but still parrots clickbait headline numbers.


 No.848387

>>848207

>tfw Allwinner A20 (Cortex A7) are not affected

Sadly Cortex A17 are, which means the C201 is affected too, so no freedom respecting machine to migrate.


 No.848391>>848412

We should exploit these security holes to break DRM to liberate documents and sofware


 No.848396>>848403 >>848417 >>850605

I'd like to see different architectures flourish but I don't think it's possible mainstream until we have a transparent compatibility layer in an operating system, which might be difficult. I'm talking like, so you install some *nix on your RISC-V machine and invoke an x86 Windows binary and have it just work. While it would be nice and technically possible with things like qemu and wine, I don't think we're there yet and you simply cannot migrate from 1 hardware system to another without either !) porting everything or 2) maintaining comparability in some form that is transparent enough. In addition to this it ALL has to be ubiquitous which is just a pipedream at this point, like you'd also have to go the other way and say have a method of running RISC-V POSIX binaries binaries on x86 Windows and whatever else.

>PPC on ARM

>et al.

No matter what there cannot be change without significant improvement or a compatibility bridge. The best examples I always point to are video displays and their input options, remember the times when things like HDMI and displayport were new but everything still had component input, it's necessary otherwise people wouldn't migrate. The same is true for hardware platforms and the pairing of software, unfortunately there's just a lot of factors in software that are hard to account for in terms of compat.

>architecture / ASM

>OS ABI

>system API

>other shit

>>847967 (OP)

I want to see that girl's caboose


 No.848401>>848413

I'm trying https://gist.github.com/ErikAugust/724d4a969fb2c6ae1bbd7b2a9e3d4bb6 (the Spectre Example from the white paper) on a P8400 (Librebooted T400) and I get either get "Illegal instruction" or it fails to get the secret string.

Does this mean my lolis are safe?


 No.848403

File (hide): 48efc79d689861f⋯.jpeg (Spoiler Image, 235.29 KB, 1280x1919, 1280:1919, C6oIhkHVAAAxFIO.jpg_orig.jpeg) (h) (u)

File (hide): de0e748b58f71eb⋯.jpeg (Spoiler Image, 307.4 KB, 1366x2048, 683:1024, CthKm2GVYAAXWZi.jpg_orig.jpeg) (h) (u)

>>848396

Good enough.


 No.848409

>>848200

>>848200

Literally retarded.


 No.848412

>>848391

That was my thought as well. How can we use this to liberate stuff. Hack game consoles, phones, tv boxes to reclaim them as our own.

My other thought was to the JS whores of the Web: "I told you so".


 No.848413>>848414 >>848439

File (hide): 4eff218c37da7e3⋯.jpg (171.85 KB, 453x553, 453:553, 4eff218c37da7e3e4e619e3fc3….jpg) (h) (u)

>>848401

No. It's because your CPU doesn't support the rdtscp instruction, which that example uses for timing. As a quick hack you can alter the code to call rdtsc(); instead of rdtscp(&junk); in the two places that do it, and sometimes it'll work, but erratically. Try playing with the CACHE_HIT_THRESHOLD too to see how it affects it. Sadly this does not mean your system is necessarily more safe, it just probably needs to be targeted in a different manner.


 No.848414

>>848413

Oh cool that's markup. Those two functions have double-underscore prefixes.


 No.848417>>848420 >>848462

>>848396

quite possible, as most people these days only just use desktop replacement laptops for their daily driver while laptops are mostly inupgradable which makes sense to make use of SoC.

the devs will eventually leave x86 and adapt RISC.

Apple and armfriends will bring the fininishing blow to push (((pajeetwindows-jewtel))) shitty(literally) "portable market that never evolved" into its final death.


 No.848419

>>847967 (OP)

Oh, don't worry. They prepared ahead of time with systemd and other such bullshit. You think anything happens by mere coincidence? THINK AGAIN


 No.848420>>848462

>>848417

>Apple ships apple A11 Macs

>Appstore ships all software with ARM compat

>???

>2018


 No.848423>>848440

>>848134

>glossy screen

>no keypad

Literally shit computer. Worthless for gaming and worthless for working. What is this? A facebook(tm) machine?


 No.848424

I'm on LTSB 2016 and haven't got the patch yet.


 No.848439>>848443

>>848413

I did try the rdtsc that was suggested on the comments but I always get the ? guess even after modifying CACHE_HIT_THRESHOLD. None of my computers seem to support the rdtscp instruction, probably because they're 10+ years old.


 No.848440>>848443

>>848423

>>glossy screen

normies don't care. if the success of MBP is any indication, i'd say they actually prefer glossy.

>>no keypad

>he likes having the trackpad off-center


 No.848443>>848452

>>848439

Are you compiling with GCC and with -O0? It works some of the time on my X200T.

>>848440

>he likes a trackpad

I'm much more enamoured with the TrackPoint personally. Or a mouse when it's set up at home.


 No.848452>>848457 >>848477

>>848443

I did. I'm trying again and the first result is always 0xFF and the second best is usually 0xFE, with the rare 0xFD.


 No.848457

>>848452

Interesting. What's your system? I found around 90-100 works best. This is on Debian Stretch, and with Libreboot (shouldn't matter).


 No.848462

>>848420

>>848417

This is highly likely since they've legitimately been doing that forever.

>DOS and Windows compat on Apple DOS and System 6-9 (maybe earlier too)

https://en.wikipedia.org/wiki/Compatibility_card

>System 9 and below compat on OS X PPC

https://en.wikipedia.org/wiki/List_of_macOS_components#Classic

>OS X PPC compat on x86

https://en.wikipedia.org/wiki/Rosetta_(software)

>executable format made for handling multiple platforms

https://en.wikipedia.org/wiki/Universal_binary

>they use LLVM which is already popular outside of Apple for aiding dynarec and porting to other platforms

I'm getting excited.


 No.848465>>848580 >>856740 >>856750

File (hide): f6a0b691ec30f0d⋯.png (56.97 KB, 1189x509, 1189:509, Screenshot from 2018-01-05….png) (h) (u)

>>847967 (OP)

>meltdownattack.com

I call bullshit.

From elsewhere:

>The meltdown exploit demonstrated that having the kernel mapping available in userspace can be risky. Modern processors prefetch data from all mappings to run as fast as possible. What data gets prefetched depends on the CPU implementation. When a running userspace program accesses a kernel mapping, it will take a fault and typically crash the program. The CPU however, may prefetch kernel data without causing any change to the running program. Prefetching is not usually a security risk because there are still permission checks on the addresses so userpace programs cannot access kernel data. What the meltdown researchers discovered was it was possible to measure how long data accesses took on prefetched data to gain information about the system. This is what’s referred to as a side-channel attack. The KPTI patches reworked how page tables are set up so that the kernel is no longer mapped in userspace. This means that userspace cannot prefetch any kernel data and thus the exploit is mitigated.

>Actually writing an attack to collect useful data from this exploit can take weeks or months to develop for a single system. Still, in the interests of security the KPTI patches were merged to close this hole. The side effect of not having the kernel page tables always mapped means that the TLBs must be flushed more often, causing a performance penalty. By default, Fedora has KPTI enabled to provide security for users. KPTI can be turned off by passing “nopti” on the command line if users don’t want the performance hit.

Note:

>Actually writing an attack to collect useful data from this exploit can take weeks or months to develop for a single system.

Meaning, this is literally nothing.

It really activates your almonds that bugs nowadays are given a catchy name and a logo. I call bullshit and I call problem-reaction-solution style fearmongering.


 No.848477

>>848452

Debian sid on a Librebooted T400 running Linux 4.14.0-2-amd64. I've tried setting all values between 60 and 300 to CACHE_HIT_THRESHOLD and I still get nothing.


 No.848498

>>847967 (OP)

this is a problem with side channels caused by meme tech like branch prediction and caching, it wasn't even remotely surprising.


 No.848580

File (hide): fbd599782d4a679⋯.jpg (89.46 KB, 1000x483, 1000:483, thumb-1119198-1000x483-2c8….jpg) (h) (u)

>>848465

Spectre a CUTE!


 No.848619

>>848274

More likely scenario:

>who cares about security fixes, my faceberg needs to run smoothly

>t. every normalfag ever


 No.848620

>>848309

>I was in a fairly large subreddit

You have to go back.


 No.848626>>848669 >>848670

>>847988

You mean 50% decrease in performance.


 No.848669>>848670 >>848789

>>848626

(((certain workloads))) are unaffected


 No.848670>>848789 >>856667

File (hide): fefaa7f13426859⋯.png (449 KB, 1017x667, 1017:667, 1515189795380.png) (h) (u)

>>848669

>>848626

I ran benchmarks on muh gaems after the Windows patch rolled out and nothing changed

Gee, its almost as if the majority of software isn't constantly read and writing to kernel memory! Intel can go bankrupt tomorrow for all I care but this bug simply isn't going to effect the majority of users. I think its safe to say that the year of the Threadripper server is here. Hell, with this level of parallelization from Threadripper I'm looking forward to the next supercomputer number 1 in the Top 500 to use AMD chips


 No.848685

>>848684

Feco if nutzi


 No.848686>>848693 >>848740 >>848793

>>848684

>calls other people shills

>while shilling for locked-down chink shit made by a tranny (ARM)


 No.848693>>848694

>>848686

Where does it say that? I guess knowing how to read is not needed to be the x86 white knight.


 No.848694>>848708 >>848987

>>848693

The C64 is the white man's PC.


 No.848707>>848783

>>848684

k

its at times like these I wish /tech/ had IDs


 No.848708>>848756 >>848954

>>848694

https://en.wikipedia.org/wiki/Jack_Tramiel

>Tramiel was born as Idek Trzmiel[4][5] (some sources also list Juda Trzmiel, Jacek Trzmiel, or Idek Tramielski) into a Jewish family, the son of Abram Josef Trzmiel and Rifka Bentkowska.[4]

(((C64)))


 No.848740>>848793

>>848686

>vouches for a process arch made by jews

good goy


 No.848756>>848771

>>848708

Richard Stallman's a Jew.


 No.848771

>>848756

Glad you see my point


 No.848783

>>848707

IDs are for fags and I intentionally change my ID every post on boards that have them.


 No.848789>>848840

>>848669

They're desperate, that's for sure.

>>848670

>I ran benchmarks

Sure you did, kid.


 No.848793

>>848740

>>848686

>8686

Checked your botnet double width dublos


 No.848812

Not to get too tin-foil hat here, but I think the new processors that have this "fixed" will involve an inaccesible shadow stack that holds the real stack being fed to the processor while we only have access to a virtual stack for debugging purposes.

http://archive.fo/RGPpS

>Intel Releases New Technology Specifications to Protect Against ROP attacks

I also think these new lines of processors will have very strong DRM and that all new media will start using it.

A technical paper on shadow stacks:

https://people.eecs.berkeley.edu/~daw/papers/shadow-asiaccs15.pdf

I can't upload to pdfs because I'm a torfag.


 No.848840

>>848789

>Sure you did, kid.

I posted the results in the GG thread on /v/, it was GTA V

I didn't want to post them here because you retards would just cry propriatery botner something something


 No.848854>>848865

File (hide): 14fa9b424ce3c9c⋯.gif (441.56 KB, 441x270, 49:30, 1393781058125.gif) (h) (u)

>expect to see a thread full of people talking about possibility of disabling intel management engine and amd psp

>expect to see people talking about possibility of core boot or similar projects for modern systems

>instead get insane ARM spammers taking us on a magical mystery ride of retardation


 No.848865

>>848854

> doesn't even read the OP

Thread isn't about fixing x86. It's a hopeless cause, a dead-end closed-up architecture of botnet and endless bugs. Anyone that still wants to cling to it is a cuck tbh.


 No.848875>>849012

Semi unrelated note

Most normies outside of redditors don't seem to know about this yet.

To illustrate the point, talked to a gaming fag a while ago and he seems to think everything will be fine and Microsoft and Intel will get everything fixed and this won't be a problem.

When the reality sinks in or the news breaks and performance starts to suck should I pick up the hardware they'll be dumping?

I'm picky about what gets to go online so they'd amount to little more than being controllers, media centers, emulator stations.

But I'm in a college town and I see a lot of MSI, Alienware, Asus gaming laptops around.

Or should I wait for more development in Power, arm etc.

And pick up what comes out of there?


 No.848954

>>848708

(((Tramiel))) was only the guy holding the chains on the dev team. They're unpozzed.


 No.848985

File (hide): 59ba6f87e0a3d2c⋯.jpg (15.44 MB, 9999x9999, 1:1, intel_collage.jpg) (h) (u)


 No.848987

>>848694

>The VIC20/C64/Amiga is the white man's PC.

.. until kike tramiel destroyed it.


 No.848991

File (hide): 9577c1ee91957ef⋯.jpg (76.08 KB, 600x610, 60:61, gopnik award.jpg) (h) (u)

Any idea when buster is going to get the patch? Sid and stretch have it.


 No.848993

>>847967 (OP)

><Death of x86

><Rebirth of IA-32/64 market and Atom (pre 2013) gets massive overhaul

FYI IA-32 and IA-64 are totally different instruction sets. IA-32 is part of the x86 family, it's literally i386. Atom also is literally x86 and x86_64 as well. IA-64 is actually a unique cpu arch, but it's also made by intel, the people who put us in this situation.

RISCV plz


 No.849008>>849012 >>849013 >>849017 >>856743

File (hide): a69908dbbb95c46⋯.png (47.52 KB, 858x405, 286:135, pasted image 0.png) (h) (u)

Christ on a bike

Look what the Meltdown fix did to a server at Epic (devs of shit-tier freemium shooter) compared to 2 unpatched ones.

https://archive.fo/gDphe

epicgames.com/fortnite/forums/news/announcements/132642-epic-services-stability-update


 No.849012

>>848875

>To illustrate the point, talked to a gaming fag a while ago and he seems to think everything will be fine and Microsoft and Intel will get everything fixed and this won't be a problem

He seems to be far more intelligent than you are since games do not need to access kernel memory multiple times a second like servers do so as far as that goes there will be no difference whatsoever

Its only servers getting hit like >>849008


 No.849013>>849016

>>849008

Fuck, I guess they'll need to buy some new Intel CPUs to compensate.


 No.849016>>849019

>>849013

I really do hope server admins start showing Threadripper some love now as a result of this shit. Threadripper makes a hell of a lot more sense for servers Meltdown or not


 No.849017

>>849008

buy new intel-aviv cores, goyy!


 No.849019>>849023

>>849016

Epyc you mean.

AyyMD couldn't meet demand even before this shitshow.


 No.849023

>>849019

Hopefully they can start ramping up production now that they have a lot more back orders to fill no doubt


 No.850420>>850595

File (hide): 7944274919f5962⋯.jpg (30.79 KB, 616x360, 77:45, question.jpg) (h) (u)

Please forgive my lack of knowledge regarding the subject at hand, but are Qualcomm Krait CPUs known to be affected by these here specific computing hardware design flaws?


 No.850595

>>850420

Qualcomm has said some of their chips are vulnerable, but has not specified which.

Krait is based off of A15, which according to ARM is vulnerable to both variants of Spectre. It also appears to be vulnerable to a less severe variant of Meltdown which ARM does not appear interested in mitigating.


 No.850600>>856429

>>848295

>corporatism is not capitalism

but it's only possible because of capitalism

>>848306

ancap is meme-tier ideology

the internet was a mistake, bring it all down


 No.850605

>>848396

Or C finally accomplishes the whole point of its existence and is a portable language, all you have to do is switch a few compiler flags and it'll compile for the motorolla 6500 or x86-64, winblows, mac, linux, plan9, Intel ME and so on.


 No.850608

>>848045

I’d just like to interject for a moment. What you’re refering to as x86 is in fact x86-64, or as I’ve recently taken to calling it, x86 plus AMD64.


 No.856429>>856571

>>850600

>but it's only possible because of capitalism

Au contraire, it's only possible because of the state.


 No.856463

File (hide): f96eeb991300094⋯.jpg (32.09 KB, 318x405, 106:135, oh god.jpg) (h) (u)

Do these exploits affect DEC Alpha64 processors?


 No.856538

>>848080

>There's literally no advantage in using x86 other than you have more performance over ARM at the cost of crunching more TDP.

>I don't know what efficiency is

Look man computations per watt is what matters in the data centers. If one cpu drew 2 kilowatts that would be fine as long as it does more with those watts than a number of lower power CPUs with the same draw


 No.856545>>856557

File (hide): ca6af950aaf9d76⋯.webm (5.84 MB, 640x360, 16:9, Ubuntu_Causes_Girl_To_Dro….webm) (h) (u) [play once] [loop]

Mawaru Penguindrum came out seven years ago.

Give it up, nerds. Your /salt/-white meme systems are DoA.


 No.856557

>>856545

FUCK THAT VIDEO

FUCKING CRINGE


 No.856571>>856579 >>856617

>>856429

>statism meme

<who enforces your voluntary contract, pajeet?


 No.856579>>856585

>>856571

>he does not know that this is answered in detail in every ancap 101 book


 No.856585>>856588 >>856589

>>856579

<he cannot answer the question, even with his collection of books at hand


 No.856588

>>856585

>It is my job answer the same shitty question 10 thousand times


 No.856589>>856590 >>856595

>>856585

>It is my job answer the same shitty question 10 thousand times


 No.856590>>856594

>>856589

That doesn't seem like an argument to me, there's absolutely nothing here but pathetic evasion, as expected from ancap mongs


 No.856594

>>856590

not an argument


 No.856595>>856599

File (hide): 89ba55000adadfe⋯.jpg (44.24 KB, 600x490, 60:49, dumbancap.jpg) (h) (u)

>>856589

Here let me help:


 No.856599>>856604 >>856740

>>856595

Wrong ancap does not mean laws dont exist and that people dont enforce them. There are prisons, security forces, judges, etc.


 No.856602

>>848156

>servers use x86 for legacy compatibility and performance

>game consoles use x86 because it eases burden on developers who use x86

I guess you could say the industry is locked into the x86 ecosystem.


 No.856604>>856607 >>856608 >>857142

File (hide): 7a5af7dc578fe61⋯.jpg (127.88 KB, 1468x889, 1468:889, blackwater-logo.jpg) (h) (u)

>>856599

That's a state. That's literally a state. That it's privately owned means that it's a collection of authoritarian regimes. Suharto's Indonesia is such a place, one family owns prisons and judges and security forces.

You don't get to pretend it's not coercive force if it's owned by a handful of families instead of "the public". I also find it laughable that you don't see how our current "state" is just such a group of judges and prisons and security forces owned by a select group of families.


 No.856607

>>856604

Wrong.

The existing state is not owned by anyone. Kill off any of the top 10 families and everything would operate about the same.

If it was owned the state would operate for their profit and would operate entirely differently. The United States government spends trillions every year. The richest person has adding up their total assets only around 100 billion.

Private ownership of the United States or any government would be incredibly profitable for the owner. The fact that we dont have trillionaires is proof enough anarcho capitalism is not the current configuration.


 No.856608>>856611

>>856604

A person is free to do what they want with their own property. A business owner can fire all the employees. A property owner can evict those on it. A slave owner execute his workers, etc.

It is not initation of force to operate with your own property as you see fit.


 No.856611>>856612

>>856608

>It's not water if it's in MY bottle


 No.856612

>>856611

Its not theft to take from your own fridge, someone elses though..


 No.856613>>856614

>"We have permitted cynical political reactionaries and the spokesmen of large corporations to pre-empt these basic libertarian American ideals. We have permitted them not only to become the specious voice of these ideals such that individualism has been used to justify egotism; the pursuit of happiness to justify greed, and even our emphasis on local and regional autonomy has been used to justify parochialism, insularism, and exclusivity -- often against ethnic minorities and so-called deviant individuals. We have even permitted these reactionaries to stake out a claim to the word libertarian, a word, in fact, that was literally devised in the 1890s in France by Elisée Reclus as a substitute for the word anarchist, which the government had rendered an illegal expression for identifying one's views. The propertarians, in effect – acolytes of Ayn Rand, the earth mother of greed, egotism, and the virtues of property – have appropriated expressions and traditions that should have been expressed by radicals but were willfully neglected because of the lure of European and Asian traditions of socialism, socialisms that are now entering into decline in the very countries in which they originated." [Bookchin]


 No.856614>>856616

>>856613

Proudhon was the first AnCap


 No.856616>>856618

>>856614

>"Anarchists conceive of power as an essentially malignant evil that must be destroyed. Proudhon, for example, once stated that he would divide and subdivide power until it, in effect, ceased to exist ... a notion as absurd as the idea that gravity can be abolished ... The truly pertinent issue ... is not whether power will exist but whether it will rest in the hands of an elite or in the hands of the people ... Social revolutionaries ... must address the problem of how to give power a concrete institutional emancipatory form."

(ibid)


 No.856617

>>856571

>who enforces your voluntary contract, pajeet?

The neutral third party in the contract specified to enforce it, clearly.


 No.856618

>>856616

*some anarchists


 No.856664

File (hide): 50c234cf415db61⋯.jpg (71.12 KB, 650x650, 1:1, GoogleNowHiring.jpg) (h) (u)

The google cultists in this thread man, geez.


 No.856667

>>848670

You wont notice much with games retard. It's server applications that are effected.


 No.856740

>>848349

<tfw your machine is vulnerable but nobody remembers how

>>848465

<tfw when this has been around since at least 6 months and thats only in the white hats private club i cant speak for how long the chinese knew

where did you think they got the time to make a logo for them came from?

>>856599

<tfw you think you support something but have no idea what it really means

how is it alt-right people are so consistently dumb as fuck?


 No.856743

>>849008

well now i know why they wont release private servers for this game

no reason a game needs to use that much shit in the kernel

[yeah preformance hit is only on syscalls so idk wtf garbage theyre coding at epic]


 No.856750

>>848465

go be a sheep somewhere else


 No.857142>>857144

>>856604

except only the private and secret families have control over everything. authoritarian state is cuckoldry at its finest.

>but le state

>muh child slave

>muh

nu uh. ancap seeks recognition and transparency in order to actualize the NAP in the first place.

>what is anarchism

>what is uber and other free entry growing markets

>what is internet

people don;t e


 No.857144>>857146

>>857142

*don't even know they're using ancapist services.

ancap seeks to replace this world's occult cabal ruling.


 No.857146

>>857144

... just reclaiming humanity once and for all. if you guys still don't agree with ancap then you all deserve a dystopian life but it won't happen. it's like those killrooms where secret families torture poor souls for several days. you are all fodder to the kabalists.




[Return][Go to top][Catalog][Screencap][Nerve Center][Cancer][Update] ( Scroll to new posts) ( Auto) 5
169 replies | 28 images | Page ?
[Post a Reply]
[ / / / / / / / / / / / / / ] [ dir / ausneets / ck / f / fur / kpop / pdfs / strek / thestorm ][ watchlist ]