http://forums.anandtech.com/showthread.php?t=2457025
>Zen was a clean sheet design that started few years ago. We are in the final figure of executing and the milestone that you want hear us talk about is Zen tapping out, which should be over the next several months. And then putting samples in the hands of our customer and then starting portfolio of revenue in 2017.
IT'S OVER AMD IS FINISHED & BANKRUPT
What a fucking shit show
But we get Bristol Ridge next year which is enough to blow out Intel.
>>51758142
Bristol Ridge ain't shit m8.
>>51758142
>Carrizo rebrand
>blow anything
TOPPEST LELKEK, AMDPOORFAGS
>>51758045
at amd we enjoy fucking up
I prefer they rush it, fuck up and are finally destroyed by Intel
Not like it matters, Intel has Bendlake refresh next year and cannonlake in Q3 2017 at best.
This means we will no longer see Zen before Q4 2017, at the earliest.
bye AMD, it was nice knowing you.
>>51758882
Why do you support Israel?
>>51759201
>starting revenue in 2017
>in 2017
>>51759230
They're the people who have been tortured through the entirety of human history, they deserve a break.
>>51759230
>Why do you support Israel?
when the choice is israel or isis (amd is funded by the same people funding isis) then the choice of supporting da joos is a no brainer.
>>51759230
because I don't give a fuck which country my processors come from as long as they deliver performance
>my sides when a /g/ is retarded enough to believe that a fabless piece of shit company with poo in loo engineers could do anything worth a shit.
I can wait for zen to come out as bulldozer mk II now with more slow edition.
>>51760328
They sure did a lot of stuff 'worth a shit' against someone significantly wealthier than them years ago.
But that's okay, not like you could possibly know, being 16 and all.
>>51760677
that was only before they sold their fabs and fired all their engineers though, now AMD slashes r&d budget and fires key employees (jim keller comes to mind) just to staunch their losses
>But the key is tapping out in the next several months, samples and customers for the validation of the product over the 2016 time frame and then the revenue ramp happening in 2016.
nothing to see here. OP just being retarded as usual
>>51762275
going from tapeout to initial product in less than 6 months would just result in another TLB bug or worse
the reality is that it will take at least a year from tapeout for zen to be viable
>>51762358
If they tape out in February it's reasonable to have a launch in late 2016 and the first full year of revenue being 2017.
This is not news. This is what they've said for months.
>>51762481
>If they tape out in February it's reasonable to have a launch in late 2016
it isn't, the only x86 architecture to ever tape out in less than 12 months was netburst
for reference, both bulldozer and barcelona took about 15 months from tapeout to initial product. we're going to be looking at q4 2017 and 2018 if AMD takes much longer
>>51762503
Are you retarded? They will tape out in Q1 2016 and be done at the latest by Q2 2017
>>51762086
>not knowing how Keller works
He goes in, makes a kickass chip, then leaves. He's done it with everyone he's ever worked for.
>>51762564
>make a kickass chip
>won't even beat last gen intel
>intel will still be ahead 2 gens+
>AMD competing with both Nvidia (same size) and Intel (10x their size)
>still competitive with both
>>51762556
>Are you retarded? They will tape out in Q1 2016 and be done at the latest by Q2 2017
still takes 6 months to produce enough chips to actually ship m8. this means either a paper launch in q3 or q4 and shipping in q1 2018, which is way off the mark needs to hit (they need to be available by q4 2016 or q1 2017, otherwise intel will just beat them to the punch with cannonlake and skylake-e)
>AMD claims 40% IPC gain
>hasn't even made a zen chip yet to test and bench properly
>>51762645
So they're offering ivy bridge 14nm processors in 2017/18?
What the fuck AMD???
Intel will still be easily in the lead in both price and performance. Will probably have 6 core i5s by then.
>>51762936
Intel won't do shit to performance unless they're forced to by AMD's competition.
>>51762645
>Zen is on schedule for availability in 2016 and a first full year of revenue ramp in 2017.
- Lisa Su less than 2 months ago during earnings call
>the revenue ramp happening in 2016
- CFO in OP
Either both the CEO and CFO of a large semiconductor company have no idea what they are doing or you are full of shit.
>>51763021
>Either both the CEO and CFO of a large semiconductor company have no idea what they are doing or you are full of shit.
Or they're making false promises and covering their ass. It's a fact that NOBODY has gone from tapeout to product in the timeframe they're talking about.
>>51762758
There's quite a lot you can do in the kinds of chip simulators (can't recall the proper term, but it's basically a massively souped up FPGA) that companies like AMD and Intel have access to. They won't run anywhere near full speed, but they're more than adequate to make estimates on IPC counts as all they need to do is realistically simulate the behavior of the real chip.
I'm kind of worried for AMD at this point... Thought that the chips had already taped out and we'd se a launch in Q3 or 4 of next year, but if it's a 2017 launch it's going to look even more grim for AMD. Thou I guess them spinning off the GPU segment once again as it's own company makes even more sense as they could much more easily now sell it onto some capital investment firm or other tech giant interested in their GPU tech.
>>51762996
They just need another 2-3% better ipc per release and zen is doa.
>>51763307
There is more to a CPU than IPC
>>51763339
If zen is going to be a 2017/2018 release and isn't priced very competitively then it's doa.
I'm not paying i5 price for a zen processor as fast as a 3 generation old intel.
Damn, AMD. This is a shame!
>>51763054
I recall news from months ago that according to AMD engineers Zen had already been taped out prior to Keller leaving.
Are we assuming too much going off the word of a glorified bean counter?
>>51763390
50-60% more IPC than Piledriver would put Zen way ahead of any current Skylake chip.
>>51763743
>it's 50-60% now guys!!!
>>51763765
40% over excavator would be 50-60% over piledriver
>>51763765
40% more than Excavator.
Excavator is 10% more than Steamroller
Steamroller is 10% more than Piledriver
>>51763064
Independent GPU companies today are worth less than you'd think. The SOC business is the future, AMD is actually crazy-undervalued right now by a financial sector that simply doesn't understand technology.
>>51763868
>>51763813
>>51763743
We'll just have to wait and see when zen comes out.
Well all new consoles run on AMD, I doubt that they are hurting that badly.
>>51763868
These are rough figures based on 4 different benches.
Keep in mind that Steamroller and Excavator are only available in APUs lacking L3.
>>51759258
For real though, every time some dictator threw a hissy fit and decided to genocide people, they always gassed the jews. They're like the ketchup of genocide: if you don't attempt to kill all jews, you better have something even better planned, or else your genocide is going to be shit.
tapeouts take several cycles before they're finished--ie making sure the chips actually fucking work. from ~the slidez~ it seems that zen had its first tapeout october at the very latest so they're more or less on schedule for Q4 2016.
i agree AMD is fucked, but not as fucked as these people are saying.
>>51762758
>AMD claims 40% IPC gain
>Zen has hyperthreading
>hyperthreading gives 40% IPC gain best-case
>YFW Zen is just Steamroller with hyperthreading.
>>51766069
>tapeouts take several cycles before they're finished--ie making sure the chips actually fucking work. from ~the slidez~ it seems that zen had its first tapeout october at the very latest so they're more or less on schedule for Q4 2016.
i really hope this is the case and this pooinloo was just talking out of his ass, but if he wasn't this is really bad for AMD. they won't survive if zen is delayed any longer than it already is.
>>51766189
IPC uplift is independent from SMT.
>>51759322
But ISIS is a CIA-Mossad joint creation
>>51766305
Dude, I don't even goto /pol/, but I'm glad someone fucking knows what's up.
>>51766208
he's not really talking out of his ass, they will still continue to tape out for the next several months as he said. it's not a one-and-done thing unless you're nvidia and trying to push out fermi.
>>51766189
its 14 nm you tards my god go back to /x/
skylake is 14 nm . yea no 10 nm from intel and amd is catching up. 40% isnt even the limit for what it should be able to do. the extra 10% the year after on the road map zen+ is within the range you expect
considering amd has been making 28 and 32 nm chips. think of that every time you look at a old bench mark
>>51766189
>hyperthreading gives 40% IPC gain best-case
what other stats do you have hidden up your ass?
>>51766531
>the extra 10% the year after on the road map zen+ is within the range you expect
waiting evolved
>skylake is 14 nm . yea no 10 nm from intel and amd is catching up.
intel will be shipping 10nm chips by the time zen is out
