Is AMD kill? Are Ryzen and Vega not enough to make it competitive

>>61161917
Zen has delivered a swift kick in the balls to Intel in their core market (servers). Vega may or may not even be competitive anymore depending on drivers. Future looks good for Zen, not so much for Vega.

>>61161917
Considering vega is a massive flop, their mobile APUs look like housefire shit.

As of this moment, AMD has a probability of about 75% to go FINISHED&BANKRUPT in the next 9 to 12 months.

Hope this helps OP.

>>61162038
By the time those APUs come out they might have actual working drivers, though. We don't know how well it scales at lower clocks.

>>61162050
Even if Vega is a flop, Zen is enough to keep them going.

is this the intel/nvidia shilling central?

File: amdsalecpuuserbench.jpg (417KB, 3455x1245px) Image search: [Google]

417KB, 3455x1245px

>>61162092
Sadly, no.

It boils down to EPYC specifically and how much it can rake in FAST.

Happy thought: Radeon division being 100% sold to Intel/Apple very soon.

>>61162138
I doubt they're going to sell off RTG.

>>61162239
If they don't fire King Poojet, they deserve to go bankrupt.

File: gfdfdf.png (269KB, 467x522px) Image search: [Google]

269KB, 467x522px

>>61162239
Either that or spending the next 4-8 years on the constant brink of bankruptcy.
Selling RTG would give them time to consolidate some gains on the CPU market and hope for the best.

>>61162275
They've already staved off the immediate threat of bankruptcy.

>>61162498
>They've already staved off the immediate threat of bankruptcy.
[citation needed]

File: 1498659836093.png (333KB, 402x750px) Image search: [Google]

333KB, 402x750px

>put Raja in his RTG containment zone
>he still manages to ruin Ryzen APUs

JUST FIRE HIM LISA

>>61162758
Reminder that both Rajeesh and Lisa are Diversity Hires and Amd deserves to go down in flames

>>61161917
>Is AMD kill?

Their stock went up almost 600% in the last 18 months, so they are doing pretty good I'd say.

not if they open source their PSP shit

>>61164072
Eat a dick faggot. Lisa was working on semiconductors at MIT while your daddy was dropping out of high school.

>>61164882
Encrypted RAM makes that irrelevant.

>>61164900
non sequitur
retarded argument too

What does that have to do with being a good CEO?
And are you sure she was the BEST and not just on par with a 100 others and worse than a few others and only chosen bc POC moron?

"Diversity hire" doesn't mean they put illiterate 90iq people there, it means they pick slightly worse or on par POCs/women solely for the color or gender.

>>61164911
PSP has ring 0 privilege. You can't even release x86 cores without it.

>>61164944
She turned around Texas Instruments, and then IBM. She is the most overqualified CEO in almost any industry you can think of.

So fuck off with your bull shit, she put in the fucking work, and made the sacrifices to be where she is at. Turning AMD in to a competitive company against two massive silicon giants gives her fucking deity ship in the modern computing game.

>>61161969
>not so much for Vega.
Zen doesn't matter though. It really doesn't. Not in the scope of AMD's master plan. The REAL GPU plan is Navi, set to launch in 2018 or 2019. Navi is essentially Zen applied to GPUs with amazing scalability.
Vega is just a testing ground of sorts for the ideas that will make up Navi.

>>61164950
Good luck breaking the encryption from the Core cache to the RAM.

>>61165007
*VEGA doesn't matter
Fucking mistyped.

>>61165007
>AMD's master plan
>just wait
I want Adoretard to leave

File: carly-fiorina-ceo-and-president-of-hewlettpackard-makes-a-point-a-picture-id51533704.jpg (488KB, 1024x873px) Image search: [Google]

488KB, 1024x873px

>>61164985
>women CEOs

>>61165021
You got BTFO this hard... my god.

>>61165007
Good luck with Navi being worth a shit based on the shitty Vega architecture.

>>61164985
someone has a boycrush+mommy fetish+yeallow fever all in one

I blame animu

>>61165015
>PSP Boot ROM runs *before* x86 core. Then non-bootrom PSP parts are stored (zlib-compressed) in the main flash. Bypass mechanism available via strap pin, but dummy and AMD signed bypass binaries needs to be always run. It is expected that newer CPUs will offload part of AGESA to the PSP, making memory init even part of the PSP

k

>>61165007
>Navi is essentially Zen applied to GPUs with amazing scalability.
RTG can't pull it off. They lack the resources, manpower and brainpower.

>>61165020
Look. Do I think vega will be competitive? Yes, but it won't be enough. AMD Knows this as well. Navi will tackle everything using a similar plan to what was used with Zen.
Base model GPU's? 2 CCXs. mid-reange GPUs? Just put 2 of the low end GPUs on one chip. High End GPUs? Just put 2 of the mid-range GPUs on one chip.
This makes everything cheaper and allows AMD to offer amazing performance at prices that nvidia can't compete with.
>>61165073
>RTG can't pull it off.
Yes they can.
https://www.youtube.com/watch?v=SMlSzts8fGQ

>>61165067
I blame your mommy issues for not bringing you chicken tendies for your gareed of women.

>>61165007
how do you Scale shit, Navi will be another pajeet shit

>>61165069
Well fuck... I was wrong.

Keep making a business case for them to make it open source.

>>61165097
>how do you Scale shit
Look at Ryzen to see how it's done.
Rysen 3, 5 and 7 are all the same chip, 2 4 core CCXs connected by AMD's genius infinity fabric with 80% yields, which is unheard of anywhere. Then look at the next step in the Ryzen line up, Threadripper. Threadripper is 2 Ryzen 7 cpus put on the same chip. Reports show Threadripper at $850 would cost AMD $120 to make and outperform Intel.
Then comes Epyc, AMD's Server platform, which is 2 Threadripper CPUs on one Chip.

THAT is how Navi will work.

>>61165087
Nothing here is inspiring any confidence that they can actually pull it off. If it were that easy, NVidia would've already done it and wouldn't be making 800mm^2 dies.

>>61165125
That Cpu, Gpu had ton of core to begin with

This is shit idea, pajeet should get hanged

File: AMD-Infinity-Fabric-Data-on-a-Chip.jpg (317KB, 1920x1080px) Image search: [Google]

317KB, 1920x1080px

>>61165133
>If it were that easy, NVidia would've already done it
No they wouldn't. For YEARS the trend in CPUs has been to make bigger (figuratively) and better chips.
Unfortunately, we've hit the limit of what this line of thinking can handle. Power efficiency is shit as are thermals. Look at Kaby Lake or Skylake-X. Intel and nvidia need to get onto developing what AMD has already spent years doing. The infinity fabric is going to change the industry. Not overnight, but it will change the industry.
Ryzen is a proof of concept. AMD CAN and WILL pull this off.

>>61165187
>literally just wait the post
amdtards never learn from history

>>61165187
NUMA had been done on CPUs years ago, has never been done on GPUs because the bandwidth/speed requirements are insane.

File: 1402274035390.gif (18KB, 252x281px) Image search: [Google]

18KB, 252x281px

>>61165187
Sure thing AMDdrone. I'll wait™ for it.

>>61165007
>Just wait!
Will AMDrones ever stop waiting?

>>61162050
amd paid off half their debt last year, the remaining debt is long term low interest debt. At the moment AMD is breaking even in terms of expenses, as ryzen is released to OEMs and Epyc begins to launch, its practically inevitable that they will return to profitability even if its only by a slim margin. AMD is also in the clear to take out more debt if need be. They are going nowhere for the next few years even in the worst case scenarios.

>>61165252
This. They thought they could change the programming paradigm to MULTITHREADED EVERYTHING and leave single-threaded performance behind with Bulldozer/Piledriver and all of the APUs they were based on. It screwed them (and end users) over for YEARS.

RTG is making the same wrong-headed bet with higher-precision hardware acceleration and HBM now.

>>61165042
?

>>61165345
Only the death can save them from all this waiting.

>>61165446
what if they are trying to pull nvidia trick and separate hardware by purpose?
same way quaddro and gtx are same chip but quadros are shit at gaming

it will lower power consumption significantly

>>61165966
>quadros are shit at gaming
P6000 beats Titan Xp at gaming

>>61166002
in dx12 games due to having hardware for it

not in anything else as far as I know

>>61162092

ryzen has cut our server budget by 40% and has cut our power usage by 20% intel just got their shit pushed in

>>61165345
will you ever stop shitposting?

>>61165268
which explains why both Vega and Navi are both using High Bandwidth Memory.

File: dicks.png (28KB, 717x337px) Image search: [Google]

28KB, 717x337px

i really hope that ryzen 3's will be cheap enough to kill intel's i5 and i3

File: US_gender_pay_gap,_by_sex,_race-ethnicity-2009.png (71KB, 800x450px) Image search: [Google]

71KB, 800x450px

>>61164944
>asian woman
>diversity hire
Asian women EARN more than black men.

>>61161917
Ryzen is mighty fine.

Vega, we shall see.
At the moment it does look like a turd though. It'll be interesting to see whether it's actually driver issues, bugs in the silicon or a plain garbage architecture.

>>61165268
Infinity Fabric scales to 500GB/s

>>61161917
Ryzen is great but their graphics cards are utter dog shit. I got a 580 and I've nothing but regret.

>>61168160
>580 and I've nothing but regret
Which rock did you lie under during the past 2 months?
You could easily have sold it for tons of money.