Advanced Micro Devices' Ryzen CPUs are freezing PCs running specific code, but AMD says it has a fix
Advanced Micro Devices, Inc. (NASDAQ:AMD) was expecting to be celebrating right about now. The beginning of its campaign to grab back marketshare from Intel Corporation (NASDAQ:INTC) is underway. The new Ryzen 7 CPU was released with considerable fanfare at the start of March, the Ryzen 5 is coming in a matter of weeks, and its new Naples CPUs for data centers have been announced.
Unfortunately, there’s a snag and it’s a pretty serious one. Those new AMD Ryzen CPUs are causing PCs to lock up. Advanced Micro Devices has confirmed the issue and is working on a fix.
Already fixed.
>FMA3
Wow it's literally nothing.
Some Gigabyte motherboards got a BIOS update for it yesterday.
>>59541216
Nothing has ever failed harder than this thread.
AYYMD IS FINISHED & BANKRUPT
>THEY FELL FOR THE RYPOO MEME
Oh lord you need driver updates so your CPU doesn't crash.
>>59541472
I know right?
https://arstechnica.com/gadgets/2016/01/intel-skylake-bug-causes-pcs-to-freeze-during-complex-workloads/
http://www.anandtech.com/show/8376/intel-disables-tsx-instructions-erratum-found-in-haswell-haswelleep-broadwelly
In to the filter, dumb tripshit
>>59541673
Not even remotely similar, the Intel bug was a rare computational error and didn't cause the CPU to permanently halt.
>>59541762
That sounds exactly like this bug actually.
>>59541673
DELET
>>59541673
BTFO
AMD SUCKS HD graphics is better than Cryzen
>>59541673
hahaha
>>59541673
OY VEY SHUT IT DOWN
>>59541673
kek this triptard just gets btfo day in day out
>>59541472
same as skylake..
>>59541673
TRIPFAG BTFO'D
bye bye tripfag
>>59541762
>cognitive dissonance
>>59541673
S A V A G E
>>59541673
it's illegal for you to visit these sites
>>59541216
>but AMD says it has a fix
holy shit....
fool me once, shame on you...
fool me 174,388 times...
>>59544109
Meanwhile Intel never did this.
>oh wait, Skylake locking Prime95 (actual software, not a fringe special test code).
But this shows Intel bugs get fixed no problem!
>oh wait, TSX had to be disabled from shipping Haswell chips without compensation/recall
>oh wait, there were those stability bugs with Skylake and Broadwell with TSX too
>Haswell chipset USB 3.0 sleep issue that was not fixed and hardware sold with it
>Sandy Bridge chipset recall because SATA ports die over time
>Atom C2000 servers recall because chip's clock generator dies over time
>>59544739
Is there an atom c2000 recall now? Last I head they were setting billions of money aside and saying nothing was wrong whilst hard ware companies were extending warranty.
>>59541762
Prime 95 is real-world software.
The code that trips Ryzen is not, it is just a loop od FMA isntruction that is unusable in practice (no load/store of values, just calculations lopp).
>>59544797
The vendors will replace stuff, or at least some of them. The issue is real, and unless they will go cynical with "hey it's out of warranty now", they will have to recall the boards (the SoC is soldered).
I don't know if it is possible to do some on-PCB hack to save an already existing affected board, but such would still need a recall of the motherboard.
Intel and the vendors prefer to be hush-hush about it but between the lines it is basically a recall situation.
>>59541216
> muh infallible Intel
This AVX bug is literally nothing, easily fixable in microcode.
The old TLB bug was 1000x worse.
Educate yourself on FDIV and Haswell TSX if you want to see some nasty ones.
>>59544999
The better question might be - what serious (freezing/crashing) bug actually happened with AMD CPUs/chipsets, ever since the TLB snafu in 2007/2008?
I recall Llano has an erratum in the divider, but that stuff must be unlikely to trigger because it was found way later and not even talked about much. So probably not a problem for users.
>>59541673
>>59541216
>but AMD says it has a fix
>but AMD says it has a fix
>but AMD says it has a fix
>but AMD says it has a fix
>but AMD says it has a fix
Thread shouldn't ever have gotten past 1 reply.
>>59541762
I'd drop that trip now if I were you.
it's literally the same stupid retard making all the anti-AMD threads I bet he spends all day looking for Ryzen defects to post about
Didn't even know such a thing existed my 1700X running smooth as butter no lockups whatsoever
>>59541472
goodbye forever
if we see you again we know you're a retard
>>59544077
t: CNN
>>59545468
It only triggers in that special code that the author only wrote for testing years back. Doesn't happen in any real software.
>>59541673
>>59541762
delete yourself, retard
>>59541472
>>59541673
Eternal bump for this tripshit getting BTFO once again. Honestly just make a new trip at this point.
>>59541673
Daily reminder for all tripfags.
>>59541673
>When you see a failure so bad that you do not hold back
Hot damn, good job.
eternal bump
>>59541673
You sent that tripfaggot to the shadow realm
Let this thread live, intel fag got BTFO
>>59549353
what freezing? I haven't had any
current uptime is 9 days
i feel really bad for tripfags.
i don't think they know what they are doing .
eternal rub it in
>>59545379
Fuck you, I want his retarded as yo stay tripped as he has never once posted something worth reading.
>>59541762
Get a grip clown, I know you're not actually this retarded on purpose.
intel has had this very similar problem since like haswell
>>59541762
>>59551675
Fight! fight! fight!
>>59551689
Fuck off, I'm not gonna fight my internet friend. But I have to honestly admit what he just said was so dumb I can't defend him.
>>59551737
:^(
>>59551737
>>59551675
And just who in the jolly shit are you?
>>59551882
a mason, that's all
>>59541243
mine wasn't one.
>>59541673
intel shills BTFO
>>59541673
Does the guy with the Chinese name think he's Asian or something?
>>59552370
Dunno, but his posts aren't worth shit
https://warosu.org/g/?task=search2&search_tripcode=!BE/4wes0mE
As expected of a tripshit
>>59541472
Seriously beawesome you WERE a nice tripfag. Now you're just plain stupid spouting nonsense
>>59552370
It translates to "I really love you, Sawako", but it is written in an exaggerated way even most married people are ashamed of doing, you will mostly hear it from weeaboo cancer.
>>59552289
It's beta and they removed the links after a while. Better wait for a stable release of the update, since buggy BIOS can be tricky.
>>59541216
>First new x86 architecture since the pentium 3
>expecting it to be a smooth launch.
early adopters get what they get
>>59555129
>First new x86 architecture since the pentium 3
?
>>59552289
>enhance capability with DOS operation
kek, windows 7 is "too old", but DOS is a-ok
>>59555934
He's retarded. Intel still continues to rape P6 core though.
>>59541216
I have a AMD Radeon graphics card which may be not supported by Ubuntu 16.04.
What should I do? Will upgrading to Ubuntu 16.10 or downgrading to 14.04 solve the issue? If yes, will it work as good as the graphics card works on Windows? Also will have to configure my computer or will the system work perfectly after upgrading or downgrading?
Am I damaging my computer by not using a OS supporting the graphics card?
Also my laptop takes like ~50 seconds to boot? Is that because of the incompatible graphics card?
Funny how the tripfag did not respond anymore. must be shitting himself.
>>59544739
Why compare it with Intel? I don't get it. AMD had to make a move and knew that everyone would look at them and they fucked it up. It doesn't really matter if your new gen of CPUs has problems if your last gen of CPUs is still highend. AMD's last gen of CPUs wasnt even recommendable for office pc builds because of all the downsides.
>>59556162
>Why compare it with Intel?
Hmm... maybe both companies main product are x86-64 CPUs? Really makes you think.
>>59555963
Yes because one of those user bases are forced to stay because of the applications they use.
The other is mindless sheep which you can just herd wherever you like.
>>59556174
But intel can afford to fuck up, AMD can't. It took them like what, 10 years? to finally release a CPU that can compete with Intels stuff and now it has bugs.
>>59556192
>But intel can afford to fuck up, AMD can't.
Oh no, AMD can allows themselves everything, as long as basic uarch is good enough. And Zen is surprisingly good and extremly efficient.
>>59556214
So are you telling me that I should wait for the next gen of ryzen cpus because those will actually be good? :^)
I think I've heard that for the last couple of years when it comes to AMD.
>>59556238
>So are you telling me that I should wait for the next gen of ryzen cpus because those will actually be good? :^)
Don't even fucking buy AMD. Buy Intel™, they have no bugs.
Oh wait, >>59541673
>>59556257
Just answer my question without using sarcasm. I have a 5 year old intel CPU and I'm planning on upgrading because my mainboard is starting to show its age.
>>59556286
Buy Intel. Just fucking buy it goyim.
>>59556298
goyim is plural for goy.
Also
>without using sarcasm
pls
>>59556307
I said you need to buy Intel.
>>59541216
Daily bumparoo
>>59556162
Anon, every CPU ships with 50-100 if not more errata - including Intels. Lots of them lock chip up, they are just rare (as is this one - reminder that no actual software triggers this bug), often so much that they are not even addressed.
Some CPUs don't even boot correctly if you disable microcode blob update from the motherboard BIOS.
Errata are fact of life in modern CPUs.
Interestingly though, nobody yet responded to >>59545090 which might show that AMD actually has had less bugs with actual impact in since 2008.
>>59556286
Not him, but the current Ryzen looks quite good already. I don't think it has the "v2 was much better than v1" syndrome like Phenom 1 had, or Bulldozer against Piledriver.
Actually, if Ryzen v2 was like 10-15% faster than the current one, that would actualyl be kinda scary. The arch has such a high IPC that I find it hard to believe next iteration can be more than 5% better per clock. I think what could mostly improve would be memory clocks (like, DDR4 2933 supported by default, 3200-3500 easier to get working).
Basically if you want multicore performance (and not just PC for gaymes), your choice is Ryzen octo or sexcore ot Broadwell-E 6-8core. Given the price of CPU and motherboards, the choice is clearly AMD.
>>59541673
New CPU architecturess have one average like 100 bugs that are fixed with microcode after release, even Intel ones. This doesn't mean anything at all.
>>59558181
what is this clip from
>>59559097
Even the old/"tick"/evolutionary ones. Most of them are actually left "No fix planned". And we are till alive.
>>59559107
Dirty Pair movie.
>>59541216
>Intel still having issues with 8th generations of architecture, still get a pass free card
>AMD have issues with brand new arch, slammed for it
>>59551737
Pic related
>>59553954
He has always been dumb as bricks, he used to believe AMD joined Khronos in 2013, along with a plethora of FUD and plain retardation whenever it shittalked AMD
>>59544999
What Am I looking at?
>>59560471
Trips
>>59560471
The first Pentium would give wrong FP divisions in certain occasions, that graph showcases the error, it's supposed to be flat
>>59541673
Daily remind that this thread is only being bumped to humiliate this trapfagging faggot.
>>59541673
Bumping this as a reminder, fuck tripfags
Immortalized
>>59541673
PISSPHONE BTFO AGAIN
>>59564534
No matter; we shall keep bumping this thread as a constant reminder to all tripshits.
>dumb tripfag BTFO
did he died?
>>59541216
goys BTFO
>>59541673
Bumparoo
>>59541673
Bump
>>59541673
lul tripfag
kek
>>59541673
HAHAHAHA TRIPSHIT BTFO WILL HE EVER RECOVER? HAHAHAHA
>>59541673
put me in the screencap
>>59541673
mom get the camera
Rubbing it in
The Right Way.
Bump before I watch a movie
You silly AMD fans will never understand, will you? I bet you also buy Lagdroid phones and Windoze laptops.
When I buy Intel-inside traditional devices, I'm making a clear statement to the world; I'm a normal, well-adjusted and well-to-do human being. Nobody will think twice about an Intel processor, because Intel is pretty much synonymous with processors. Instead they can focus on my highly successful career, my beautiful children (Christopher is starting QB for his senior team and Samantha is captain of the debate club), or my brand new BMW.
Now when someone sees an AMD sticker on your bargain-bin slapped-together hunk of junk, several things immediately pop into mind; cheap, shoddy, knock-off, weird. There's good reason for that. Intel is the original creator of the desktop processor, they're authentic. Thus AMD was, is and forever will be an Intel copy. Not only that, Intel is a much bigger and more successful company; if AMD made luxury products it would be great if their company was smaller than a masse manufacturing competitor, but that evidently isn't the case. I don't do underdog, because I never was and never intend to be the underdog, I intend to stay on top through it all. Intel is king.
It sends a bad image to the rest of the world when you do things like this. When you wear no-name clothing, when you buy off-brand food, when you listen to strange music, when you watch children's cartoons, when you have the wrong political opinions... when you buy AMD. They all send the message out that you're unkempt, poor, antisocial and possibly have a few screws lose. Your family will notice. Women will notice. Your future interviewer will notice.
Even if they'll both get the job done, you spend the extra on Intel-powered machines. Because even though both shirts will keep you covered the same, Armani is simply better that Walmart store-brand.
Then again, I understand that some of you simply lack aspirations and don't care or need to keep up appearances.
>>59567034
Ayyy, a new copypasta.
It's time.
>>59567094
>>59567105
>>59567116
>>59567127
>>59567141
Any specific reason you're dumping this stuff? You're never going to run into these unless you're bit bashing at near-metal low level, because the microcode update and kernel were written with this stuff in mind.
>>59564534
Hey I see my post
>>59566776
source?
This thread deserves a fucking sticky.
In the old days, it would have happened.
>>59541762
Someone gas this kike already
>>59541673
>>59567034
Is this /fa/ pasta? I like it.
Eternal bump
>>59567174
>>59567186
>>59567200
>>59567214
>>59567222
>>59567235
>>59567251
>>59567264
>>59567275
>>59567288
>>59567298
>>59567308
>>59567320
>>59567330
>>59567351
>>59567363
>>59567371
>>59567384
>>59567399
>>59567412
>>59567419
>>59567427
>>59567440
>>59567452
>>59567462
>>59567477
>>59567492
>>59567503
>>59567513
holy shit, I knew intel got bug, but this many
>>59541216
>literally have to force the lockup with specific instrucitons nowhere to be seen in normal day to day workflows
>"RYZEN IS KILL"
come on man, intel has the same shit on older procs and no one cared
>>59567167
Looks like prison school. No, its not porn despite getting staggeringly close.
>>59556121
Gnoo/Linucks
>>59541673
AHAHAAHAHAHAAHHAAAHAAHHH
>>59541673
bumping for more tripfailure shaming
>>59568114
To be fair that is really normal, tens of errata is the norm for current chips. Zen/Summit Ridge likely has similar list. AMD bugs have been less serious though, in the years since Phenom. >>59545090
Props for putting things into perspective, dumping anon.
>>59544999
>>59560471
>What Am I looking at?
Early Pentiums had a fast divider that could output 2 quotient bits/clock through a table-driven algorithm called SRT. Unfortunately a couple of table entries on the manufactured chips got left erroneously empty, making them utterly useless for numerical work even if the bug only got hit for one out of a billion operations or whatever.
kek, why is this thread still up?
>>59541673
TRIPKEK BTFO
>>59573598
apparently newfags aren't already used to aishiteru sawako embarrassing himself by posting the retarded, uninformed shit he always does.
>>59541216
/g/ - Intel Shills
>>59573743
or they're just happy that such a worthless tripshit was finally BTFO'd into oblivion
>>59574351
happens all the time, he'll still stick around because he's a cancer that isn't that easily removable.
Bump
T
F
O
>>59541216
>there’s a snag and it’s a pretty serious one. Those new AMD Ryzen CPUs are causing PCs to lock up
Why the fuck is this news? My video drivers on the shitty i5 desktops that are coming in from Dell require updates or they lock the computer up too.
>>59577911
Me on the left.
>>59541472
>>59541762
Just wondering how much does intel pay you?
>>59579360
Shills don't work that way I think.
If you read some really stupid brand-pushing crap on the internets, it's usually just some dumb-enough or too-emotional fanboy that goes vigilante because he identifies with his idol too much.
>>59582315
What am I looki-
>1.55V for those shit clocks
>>59541216
why does that man look so weird
>>59582315
nice anus board dumping 1.55 volts and nice unaccounted +20°C temperature offset
>>59575400
>2MB for a couple seconds of low-fps zooming in
>>59541472
You're the Lelouch of this board.
>>59541673
U FOCKN DID IT NOW M8
>>59582643
is that a man? thats not a man... is that a man?
>>59541216
Is that a man or a woman?