[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vr / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / asp / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / qst / sci / soc / sp / tg / toy / trv / tv / vp / wsg / wsr / x] [Settings] [Search] [Home]
Board
Settings Home
/g/ - Technology



Thread archived.
You cannot reply anymore.




File: based amd.png (628 KB, 780x656)
628 KB
628 KB PNG
>buy amd
>threadripper 1950x
>processor doesn't support messaged signal interrupts
>it's an AMD issue
>isolated only to AMD processors
>if you use GPU passthrough you can fix it
>only then though
>the real fix would be to buy a processor made by a professional company, like intel
>it never happened with 1703
>only with 1803
>the stack wasn't fully implemented until 1803
>microsoft shouldn't have to fix amd’s fuck ups
>MSR relays
>MSR_NHM_TURBO_RATIO_LIMIT, MSR_PLATFORM_INFO, MSR_PKG_POWER_LIMIT can only be patched through KVM
>tfw amd cuts corners everywhere
>if you want a professional system that supports feature stacks without breaking them, then use intel
>it's the MSR's, which amd doesn't support, but advertises to the system as supporting them anyways
>spectre
>yes, amd was invulnerable, but amd doesn't have the branch prediction that was exploited at all
>so it's not necessarily that they did it right, they just outright lack the functionality
>however, from a practical standpoint, that does mean that they're secure against it
>back to amd’s msr relays fuckups
>if the cpu advertises that it supports something, windows assumes that it supports it, as crazy as that sounds
>tfw amd likes to lie to the kernel and pretend to support things
>IOMMU is broken as fuck on amd
>broken for over a year, enabled by default, can’t install windows due to it, wtf amd
>NPT was broken for over 10 years
>this is what happens when you cut too may corners
BASED AMD
>>
Whomst'd've are quoting?
>>
>>68010640
>thinking green text is only used for quoting
back to >>>/reddit/
>>
Thank you for the cool story, bro.
>>
>>68010601
> Windows spaghetting over NUMA
> Running Windows with a serious workstation CPU
> running Windows at all
>>
>>68010650
>not getting the meme

wew lad
>>
>buy amd
only stock at $2 and shorted at $21
>>
>>68012303
Except it's not a meme. Only /jp/ knows how to greentext.
>>
>>68012264
The issue is is that they advertise that is has that functionality and lies to the kernel that it has that functionality, when it in fact does not.
>Microsoft shouldn't have to fix AMD's fuck ups
>if the cpu advertises that it supports something, windows assumes that it supports it, as crazy as that sounds
>tfw AMD likes to lie to the kernel and pretend to support things
AMD's implementation does not work because they have shit code. It is not the fault of the OS to work around shit code. If they advertise that their CPU does something, and they tell the kernel they do those things, but they actually don't, it's not the fault of the OS when it runs into issues.
>>
>>68012264
>windows
>implying
>showing absolute ignorance about the issue at hand
https://community.amd.com/thread/219560

>>68012507
/jp/ died in 2012. Now it's newfag central.
>>
>>68012264
>ayyyymd
>linux support
Pick one.
>>
File: 1-6152.jpg (288 KB, 600x3400)
288 KB
288 KB JPG
>>68010601
>https://community.amd.com/thread/219560

i dont understand what are you talking about

but shouldnt YOU searched before buying, does the cpu supports the thing you want?

calling bullshit
>>
>>68012965
>company advertises their product as being great at virtualization because moar coars
>except said functionality is broken
guest you're left with 7zip benchmarks and MT cinebench
>>
File: 1535022775794.jpg (237 KB, 900x1152)
237 KB
237 KB JPG
>>68010601
>intel fucks up
>shills jump up to shit on amd
like clockwork
>>
>>68012673
Wendell called, he said to throw the Intel errata to your face.
https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/6th-gen-x-series-spec-update.pdf

Take your windows problems to microshit forums.
>>
File: What_a_story_Mark.jpg (9 KB, 480x360)
9 KB
9 KB JPG
>>68010601
Learn to use the fucking site first intern shill
>>
>>68013043
>fixed by BIOS updates 2 months ago
Back to the drawing board, Pajeeta.
>>
>>68010601
pasta from a previous thread.
your usual fake and gay shit.
>>
>>68015108
It's my pasta, dumbass. Check the timestamps.
>>
File: 1524196415771.png (185 KB, 600x600)
185 KB
185 KB PNG
>being this obvious as a shill
>>
>>68015308
>tfw pointing out AMD's flaws makes me a shill
How does Lisa Su's cock taste sliding down your throat? Got a little dribble of cum on your chin lad.
>>
>>68015321
>flaws
every week there's a new flaw or intentional backdoor discovered for intlel
>>
>>68015364
At least Intel still gives their consumers what they advertise, instead of advertising one thing and providing another.
>>
>>68015391
>still gives their consumers what they advertise
yeah sure
>patch micro code for swiss cheese cpu
>lose performance
>>
>>68015256
post the replies on your first pasta, there's a guy explaining your fake and gay claims.
>>
>>68010601
works for me, must be the NUMAlet OS you are using.
>>
>>68015415
Yeah and I refuted them. Also this was my first pasta. Not the other thread.
>>
>>68010601
messag-what? Why do you care? LOL egghead.
>>
>>68010601
>yes, amd was invulnerable, but amd doesn't have the branch prediction that was exploited at all
Branch prediction is a REQUIREMENT for speculative execution. AMD chips have branch prediction, just of superior quality to Intel's :^)
>>
>>68010601
>buy amd
Stopped reading right there
>>
>>68010601
>>processor doesn't support messaged signal interrupts
>>it's an AMD issue
I'm in fact using TWO 1080Ti's in SLI right now, both are PCIe devices.
Have a (You).
>>
>>68015639
Nvidia and Intel don't have these problems, yet AMD does.
>>
>>68012965
go to last post
>https://www.reddit.com/r/Amd/comments/7gp1z7/threadripper_kvm_gpu_passthru_testers_needed/
might help?
>>
>>68012507
/jp/ is now property of [s4s]
>>
>>68012767
>Lying on the internet
>>
>>68015430
well, op?
>>
>>68015652
You have no idea what is even being discussed, do you?
If AMD didn't support what OP claimed, you couldn't even run PCIe GPUs with it.





Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.