I waited for Skylake but was disappointed,Is it worth waiting for Zen or will that be just a meme too?
>>51793464
it will be like second coming of jesus
>>51793464
I think Zen will be very good, but by today's standards. Dunno what Intel will have planned for the launch of Zen but I'm assuming they'll have something.
>>51793464
>but I was dissappointed
Get ready to be dissappointed again.
AMD have not taped out Zen yet. Expect an early 2017 launch or very late 2016.
AMD claims ~40% improved IPC so it sounds like zen chips will be on par with sandy vagina maybe even haslel performance.
The only reason these Chips will be better than intel ones will be price. This is the biggest reason I'm waiting for zen to build a desktop PC.
I think the waiting is worth it, especially if you want haswell-e i7 performance on the cheap like me.
>>51793464
You won't even need a video card anymore
>>51794087
Do you ever post anything that isn't pants on head retarded?
>>51794117
>claims you won't need a video card anymore if you use zen chips
>calls me a retard
>>51793556
https://www.youtube.com/watch?v=wbA-FbeNtqw
>tfw zen will be cloud compute for project bluebeam with a hidden architectural design
>>51794087
>AMD claims ~40% improved IPC so it sounds like zen chips will be on par with sandy vagina maybe even haslel performance.
AMD claimed the Fury X was faster than the 980ti across the board.
>>51794185
For once I agree with a tripfag.
>>51794569
It actually is. Problem is drivers have been shit.
>>51794656
They made that claim using practical benchmarks (see pic), so their driver shittiness was taken into account.
Of course they cherry picked completely unrealistic benchmarks like running games at 4K with medium settings and no AF, and enabling Mantle in supported games.
Watch that 40% IPC improvement be in something highly specific nobody cares about.
>>51793464
It will be a meme too. The good thing is that the amd chip will most likely not have the integrated hardware botnet like the intel does.
>>51794751
I don't think you understand the gravity of the CPU problem AMD has. They HAVE to make zen have 40% better IPC than excavator at least. There is no room for anything less.
AMD will be in financial trouble in Zen doesn't deliver since the few hopeful customers it has will flock over to intel. Zen is the make-or-break product for AMD. If Zen is shit like you imply it will be then AMD will have to file for bankruptcy or split into a CPU/GPU division and eventually just stop making CPUs for good.
>>51793464
I think it's worth it. I would definitely get a 5820k right now, but I think it's better to wait for Broadwell-E's 8 core behemoth that isn't the highest end. So I'll wait for Zen as well.
2016 is pretty nice: 28nm GPUs can finally fuck off, more NVMe SSDs, xPoint with Intel optane SSDs as well, new enthusiast CPUs, freesync probably overtaking g-sync, etc. All would be swell.
>>51794185
I'm looking forward to a zen equivalent to the a8 7650k.
If they maintain current pricing and improve the igpu to maybe a 260x or equivalent I'll be happy.
>>51793464
>AMD
>trying to fool everyone into thinking it's nvidia by turning the same shade of green
Never change, amdumbs.
>>51795015
It will all depend on what kind of memory they use. DDR4 is better than DDR3 but only if clocked high enough (ie 3,200 mhz). HBM memory is the best kind of memory but it's still pretty expensive.
They can clearly fit R7-260X graphic performance on an APU, they already did that with 28nm lithography on the PS4. The problem is the memory for the iGPU. The xbone solved the problem with quad channel DDR3 memory and the PS4 solved it with good old GDDR5. Obviously GPUs require a ton of memory bandwidth as the PS4 outperforms the xbone by almost 40% in terms of graphical performance.
I hope they let you add HBM memory modules to the zen APU motherboard because dual-channel DDR4 RAM obviously won't deliver enough bandwidth to any APU with more than 512 gpu shaders.
>>51795232
I think for now they'll just stick with ddr4.
It'll be cool if you could use HBM as shared system and vram though.
>>51795463
AMD better let us use 3.2 GHz DDR4 at least.
>>51793464
Well if it is good enough to blow intel out of the water then oh well. I already pulled the trigger on haswell About time AMD got better anyways. I may of picked a bad time to build a new pc -the card i haven't picked yet.
>>51793464
>Amd
>Good product
Kek
>will it be a meme too
You are trying too hard to 4chan OP
>>51793464
considering we just found out zen hasn't actually taped out yet and won't for a few more months, this means you'll be waiting at least ~18 months from now before anything even materializes w/r/t zen. better off just getting skymeme because we won't see anything new from intel or amd until mid 2017 at the earliest.
>>51796269
>tfw you went Hasewell
i better pray it lasts haha.
>>51795232
>Ps4 outperforms XBone by 40% in terms of graphical performance
whose ass did you pull that number out of?
>>51796718
For once he's absolutely correct, simple shader math proves that.
>>51794185
Gtfo fucking cancer.
>>51796930
>Gtfo
Get the fuck out and kill yourself.
>>51796718
Unfortunately for you he's right.
>>51796930
Dam anon, who pissed in your cereal bowl?
>>51796718
PS4: 1152 gcn cores @ 800 MHz with 176GB/s GDDR5
Xbone: 768 gcn cores @ 800 Mhz with 68GB/s quad channel DDR3
Not only does the PS4 have more gcn cores but it has faster vram. The increased gcn core count alone gives the PS4 ~35% better graphical performance. In addition, the Xbone APU is bandwidth starved by the DDR3 RAM and performance will suffer because of it.
I don't own either but it's clear Sony made a really decent console this time. The PS4 might have 50% better graphical compared to the Xbone because microshit thought it was a good idea to use DDR3 RAM for a fucking GPU.
>>51797248
B-but anon, desktop apus use ddr3 ram.
>>51797078
The tripfag, that's who, honestly I think they should all just be banned.
>>51797331
And miss out on filtering them?, have you lost your mind?
>>51797308
Which is why they suck balls. The 512 gcn core APUs are like 5% better than the 384 gcn core APUs because of howw massively bandwidth starved they are.
>>51797408
True, because they post, and they think their voice is being heard.
AMD has had a bad track record for like 10 years now.
Zen will disappoint again.
>going to be the same AMD shit again
>mediocre single core performance
>good multicore
>intel will still be king
I'm trying to do my research about building a new desktop.
Should I not go for Skylake and stick with Haswell?
Or should I spring for the LGA 1151 socket, and expect it to be around for the next 5 years?
Money is an issue here, but I'm doing Autocad, Rhino, Revit renders, with some exports of fly through video.
>>51797497
> should I spring for the LGA 1151 socket
intel doesn't give a fuck about sockets.
Build the best rig you can afford for *today*
By the time you upgrade the socket will be dead.
1151 will likely only be around for skylake and kabylake and thats it.
>>51797497
Stick with Haswell man, skylel was a flop. If money isn't a problem then you most definitely want to go with Haswell-E processors and a compatible LGA 2011-v3 motherboard.
What's wrong with Skylake?
>>51797627
Why should he stick with an inferior platform?
>>51797942
Because there are no actual benefits going with skylake except like 3% better cpu performance which is negligible.
Also see pic related, left is skylel and right is haswell
http://www.pcper.com/news/Processors/Intel-Skylake-Processors-Can-Bend-Under-Pressure-Damage-CPU-and-LGA-Socket
>>51797939
see >>51798038
God forbid you buy an unlocked skylake processor anon. Almost all coolers fucking destroy the paper thin substrate on the fucking cpu.
>>51797939
Not a big deal like every other generation besides SandyBridge.
>>51798038
Holy shit
>>51798061
>almost all coolers
You mean like the ~2 edge cases reported so far ?
With one being becaused they used an electric screw driver to install the heatsink.
>>51798061
Nice lies, faggot
The only cases were any damage was done were prebuilts that got carelessly transported over large distances, CPU cooler manufacturers (Scythe) not sticking to Intel's guidelines (which haven't changed from Haswell to Skylake) and retards using machines to tighten the screws instead of their fucking hands, which most manufacturers warn about in their instructions.
>>51798171
Do you still wanna risk killing your CPU AND motherboard at the same time by using an unlocked skylake processor which fucking jewtel doesn't include a stock cooler for?
Skylake is a moldy turd intel shat out and you nigers will eat it up like good goyim.
Intel stooped being great after haswell.
>>51798216
Pass the salt please
>>51798180
not same anon but I wouldn't risk it though. I have an i5-2500K and I'll upgrade to haswell soon. Fuck skylake.
>>51798226
not him but did intel have to make the substrate thinner? Or is this just intel being a jew like always?
>>51798233
>SB to Haswell
I would hardly call that an upgrade.
>>51798216
>Do you still wanna risk killing your CPU AND motherboard at the same time by using an unlocked skylake processor which fucking jewtel doesn't include a stock cooler for?
dude, its a fringe issue that affected how many people ?
More people have fried their systems with a shitty PSU in the same time frame as this skylake thing.
>>51798038
HA funny i canceled my 4690k k for skylake and ordered...then i read about that and said nope went hasewell in the end.
>>51798261
>not him but did intel have to make the substrate thinner? Or is this just intel being a jew like always?
it is still rated for the same 50lbs force.
I honestly expect this 'investigation' to come up saying it was the users fault for an improper installation and/or moving around the case with a heavy heatsink on.
its like the toyota stuck foot pedal thing all over again
>oh shit I pressed the gas instead of the break!
>how can I get out of this !?
>I know, I will blame Toyota !
>>51798261
Does anyone here look like a fucking psychic?
>>51798271
Yeah, cpu-wise I won't get that much more improvement. However what peaks my interest is haswell i5s have 16 lanes of gen 3 pcie. I want to crossfire three 380Xs for 4k gaming. My current setup would bottleneck them heavily since I only have gen 2 pcie.
>>51798306
Intel didn't have to make the substrate thinner though. There was no fucking need to do it.
>>51798341
>random Anon pretending to know better than a bunch of educated and experienced engineers
You're just looking for a reason to shit on Intel
>>51798354
You sound like a paid shill desu. Intel can go fuck themselves. That thinner substrate is only used on laptops specifically because the coolers don't put a ton of pressure on them. They don't fucking belong on desktops.
>>51798341
again, it still has the same 50lbs force rating.
Or are you saying intel lied about the specs?
and that manufacterers are ALL colluding with intel in saying they can't replicate the results.
Because if one could demonstrate that damage can occur with less than 50 pounds of force Intel would be hugely liable
it seems far more likely given the small amount of reported cases on the issue that its due to user error.
>hurr I wasn't careful with my PC
>it damaged the mobo/cpu
>I better blame intel!
>>51798448
That skylake chip definitely broke under 50lb of force. Intel is about to get sued.
>>51798391
Your condition is called "being delusional".
You don't like what I'm posting so you project a hostile image onto me.
You don't seem to know what you're talking about on top of it all. Cooler weight isn't what bends the CPU substrate, it's the mounting mechanism. Even a CLC or the custom loop CPU block can bend it if tightened too much.
>>51798479
>Intel is about to get sued.
They don't give a shit. They've been sued for billions of dollars already.
>>51798479
>That skylake chip definitely broke under 50lb of force.
Source?
I hate to break it to you guys, but things aren't going to improve very much for the next five or ten years. You can wait for Zen or KB Lake, but they'll be incremental upgrades. Might as well buy quality and hunker down for a decade.
>>51798479
true or not i'm not going skylake for 5% increase. I'll stay hasewell till i see different.
>>51798479
>That skylake chip definitely broke under 50lb of force.
Says you and maybe 2 other people.
How many manufacturers have released statements and done internal testing now ?
All of them come to the same general conclusion
>Don't over tighten
>be careful when transporting your pc
http://www.pcgameshardware.de/Luftkuehlung-Hardware-217993/News/Skylakegate-Kuehler-zerstoeren-Sockel-1151-CPUs-1179237/2/#a6
>>51798534
http://hothardware.com/news/intel-skylake-cpus-are-warping-under-heat-sink-mounting-pressure
>shit website
The major websites are trying to avoid pissing off Intel, probably.
>>51798562
Considering how fast current processors are, 5% is no small improvement.
Is it worth it upgrading to if you only see 5% performance increase? No.
Is it worth upgrading to if you're coming from a generation before haswell or ivy bridge? Yes.
>>51798612
Likewise, some small shitty site has nothing to lose and everything to gain by spreading sensationalist FUD.
>>51798638
>can't refute the evidence
>tries to attack the source
Go back to reddit you feminist SJW faggot retard.
>>51798612
>ask about source on skylake bending under 50lbs force
>get a link to some shit website that just recycled text and images from other bigger websites
So where's the source of the original claim?
>>51798674
what evidence ?
The evidence so far points that SCYTHE coolers could be responsible for the warping, nothing to do with intel.
>And so it would appear this is an OEM 3rd party manufacturer issue, rather than a generalized issue with the processor(s).
But you people just can't accept that so instead you cry some stupid conspiracy theory.
>>51798703
original site is some german one.
linked to earlier
>>51797331
>not using anonimyze
Are you an retard? It's also available on clover
>>51798674
What evidence? They just recycled content from other sites and their last line of text is
>And so it would appear this is an OEM 3rd party manufacturer issue, rather than a generalized issue with the processor(s).
Nowhere does it say that the CPU bent under 50lbs of pressure which would result in Intel getting sued
>tfw devils canyon for the next 5 years or more
Zen is gonna flop so fucking hard
any time I think about "NEW AND AWESOME" and "AMD" at the same time, I think about how big of a fucking disappointment the Fury X was.
>people shitting on Skylake for literally no reason other than it being new
Come on guys, this is pathetic
>>51798622
>upgrading a 3570k
Not possible, unless you're crossfiring
>Skylake haters STFU after getting BTFO
Kek
>>51794569
>AMD claimed the Fury X was faster than the 980ti across the board.
It is. In 4k.
>>51798800
I'm sure Zen will be an improvement over the 4 year old Vishera. It wont be better than Skylake in any meaningful way.
>>51797248
>I don't own either but it's clear Sony made a really decent console this time.
They both gambled and Sony ended up winning.
Xbone instead got a couple of other extras, like they increased the speed in the last minute, and it also has the fast eDRAM cache and a small flash drive for fast installing, so the HDD does not clog performance.
Also I recall something about GDDR5 being good for graphics but less so for computing, something about latency? I dunno.
But even with all that, yeah, the PS4 has clearly better graphics.
>>51794892
>Zen is the make-or-break product for AMD.
That does not mean that hey can actually get it to be good though.
And remember that they are still shipping all the chips for all the current consoles, AND they are building the ones for next gen consoles as well. So they ain't completely dead just yet.
>>51794087
Steamroller is 8% faster than Piledriver.
There are no desktop (non APU) steamroller chips.
Excavator is 8% faster than Steamroller.
There are no desktop Steamroller chips.
Zen is 40% faster than Steamroller.
Zen is 56% faster than Piledriver.
Sandy Bridge is 30% faster than Piledriver
Skylake is 20% faster than Sandy Bridge
Ergo, Zen is as fast as Skylake.
Please stop the "it's only between Ivy and Haswell" meme.
>>51799641
The 40% better ipc is not guaranteed. AMD might have been able to only get 30% IPC gains for all we know.
>>51799641
Zen has a 40% IPC uplift over Excavator, not Steamroller.
Zen is around 64% higher IPC than Piledriver.
Sandy Bridge can be 50% faster than Piledriver per clock, or higher in some cases.
http://www.anandtech.com/bench/product/697?vs=288
>>51799692
I mistyped, but the statement is correct otherwise.
>>51799706
>, but the statement is correct otherwise.
No its not. Zen won't match Skylake in serial performance per clock. It'll land somewhere around Ivy Bridge - Haswell in certain metrics, and I wouldn't expect them to have FPU performance anywhere near what intel is achieving.
People who do nothing but constantly hype up AMD products do the company more harm then any biased review ever could. Building up totally unreasonable expectations might as well be outright marketing sabotage.
>>51799692
>>51799706 again
I just averaged out a dozen single threaded tests between the 8350 and 2500k,
Adjusting for clock speed, what is by the benches a 26-30% lead with sandy becomes about 42% (4Ghz to 3.3)
>>51799742
I don't expect miracles and I'm not an AMD hype fanboy.
If Zen is 64% faster than Piledriver per clock (I assume relative performance in your oft posted picture means per clock), and Skylake is generally 20% faster than Sandy, then I am still correct.
Zen is equal to Skylake.
>>51799692
It doesn't matter how much faster you are than your predecessor, you need to be faster then your competitor.
>amd is 3/4 the speed at half the price
>amd margins/profits are almost non existent
Zen *will* be faster, but not enough to catch up to intel
Meanwhile intel will likely release a 6/8/10 core Broadwell-E in 2016
>>51799860
As I just showed, an 8 core Zen is as fast as an 8 core Skylake.
Shall we continue?
>>51799641
>There are no non APU desktop Steamroller chips
Athlon 8xx.
>>51799873
what are you talking about,
niether product you mentioned exists in the wild.
>>51799873
So amd says.
>>51800051
AMD made no such claims whatsoever.
The only real statement they made on performance is that their enterprise chips will be "competitive." They stressed that over and over. It was said about five times that they would be competitive.
Anyone with an IQ over 10 can read between the lines as to what that means for a server workload. They weren't claiming to take any performance crown, they were saying that they'd excel by throwing more threads around, and having superior socket to socket/rack to rack scaling. That is what SeaMicro's Freedom Fabric was all about, and AMD took that IP and adapted it. That is why they were so keen to promise "disruptive bandwidth."
Zen isn't going to compete with Skylake in performance per clock.
>tfw Piledriver
So zen is worthwhile upgrade to Zen
good.
Doesn't matter how good the zen arch is if the fabs are shit.
Considering that AMD is reliant on other people's fabs, all I can say is LOL.
>>51800243
>look at how dumb I am
Cute.
>>51800256
which says jack shit about yields
They can't even make enough high end GPUs
>>51800280
14nm parts have been in volume production since the end of 2014.
Cute.
>>51799918
APU with the GPU lased off.
>>51800293
Yeah and they fucking suck, just look at Samsung's 14nm vs TSMC's 16nm
Flocks of people returning samsung iphones to reroll the lottery for TSMC iphones.
>>51800293
The only ones with 14nm fabs worth a damn is Intel, and they sure aren't making AMD CPUs.
What does Skylake do better than Haswell?
Also Zen isn't coming until 2017.
http://seekingalpha.com/article/3741456-advanced-micro-devices-amd-at-raymond-james-technology-investors-conference-transcript?part=single
>>51800324
>Yeah and they fucking suck, just look at Samsung's 14nm vs TSMC's 16nm
I love seeing obvious under age children totally make fools of themselves.
http://www.tomshardware.com/news/iphone-6s-a9-samsung-vs-tsmc,30306.html
>>51800333
>I know literally nothing about this topic but I'll post anyway
>>51800348
Kumar made a statement about server chips, not Summit Ridge. Seeking Alpha also made a number of typos in their transcript.
AMD in a webcast reclarified that Summit Ridge launches in 2016.
Stick to Anandtech if you want to spin that FUD.
>>51799067
Crossfire also wrecks sli.
>>51800477
>a single game
>>51800371
>Based on the results of our testing, it's clear that both versions of Apple's A9 SoC deliver the same level of performance, but Samsung's 14nm FinFET process appears to offer slightly better power efficiency, extending battery life between 3.5-10.8 percent.
Samsung should be winning across the board, not trading blows with a supposedly larger and less efficient process.
Also AMD still has that lawsuit going on regarding the APU, and the FX, not the first time they overstated their specs.
>>51800523
>Samsung should be winning across the board, not trading blows with a supposedly larger and less efficient process.
No they shouldn't you bumbling tard. Apple is offering one product line with no distinction in SKU between chip vendors. They are not running each respective lot of chips to the best of their ability, they're running them so they offer the exact same nominal performance and battery life.
Why do people like you insist on pulling arguments straight from your ass?
>Also AMD still has that lawsuit going on regarding the APU, and the FX, not the first time they overstated their specs.
One scam artist filed suit against AMD on laughably flawed grounds. Its not going to trial. Tony Dicky is in fact a shitposter who frequents /g/.
>tfw still on 2010-2011 Phenom X6
>tfw want to upgrade
>tfw 2015 AMD chips are not that far ahead from Phenom
>tfw have to wait until 2017
fuck this desu
>>51800644
So go with haswell-e and get a 5820k
.A class action lawsuit by investors over Llano APUs is an actual lawsuit and the US court ruled that AMD must face the claims due to securities fraud.
>>51800664
The suit you're talking about was about an inventory write off, and the party who brought the complaint also has no idea what they're talking about. Representatives of a hedge fund alleged that this write off of APUs was responsible for AMD's stock tanking.
The Bulldozer family failing to inspire confidence and losing enterprise marketshare is what caused AMD's stock to take a hard downward turn.
Its not a good idea to make a factually incorrect statement the core of your lawsuit.
>>51798638
>>51798703
http://www.tomshardware.com/news/skylake-cpus-damaged-by-coolers,30690.html
http://www.scythe-eu.com/en/support.html
eh
>>51800299
>APU with the GPU lased off.
So pretty much an apu missing the half of what makes it an apu?
In other words, a cpu.
i believe in Jim Keller works
>>51793464
zen isn't even coming out until 2017.
>>51799641
zen doesn't even exist yet.
>>51800477
direct x 11
>Scooby Doo
>>51800962
Yes it does.
>>51800850
slap a CPU and GPU on single a long ass 4 slot card connected via PCIE4.0 powered via a 24pin connector and 16pin connector.
Problem solved, resolved, absolved
Or ditch the MOAR core philoshphy and go for a unified single core nuclear reactor at 10GHz. No one fucking cares about developing applications that take full advantage of multiple cores anyway.
at this point AMD should start thinking up ways to make ARM processors that can support AMD 64, then they could market chromebooks to consumers.
>>51800850
still not the same as a pure CPU. Chip started as an APU, either couldn't get the right pieces on the GPU working or needed to bin a part for demand and as such, was lased in a way to disconnect the GPU from the rest of the chip.
There's still a GPU on die, it's an APU.
>>51801034
It's a cpu because the apu section is disabled.
An apu is nothing more than a gpu and a cpu with a few fancy features like HSA.
Can the 860k use HSA?
What's a safe voltage for i5-6600k? I managed to get 4.6ghz on 1.39v
>>51793464
Skylake is a 25-30% IPC improvement over Sandy Bridge. Zen is claimed to be a 70% IPC improvement over Bulldozer. This should put Zen noticeably ahead of any current Skylake chip.
>>51793922
>People actually fell for that bait
The server chips aren't taped out yet. Desktop chips are.
>>51802662
It depends on how lucky you were in the silicon lottery.
>>51802676
>The server chips aren't taped out yet. Desktop chips are.
saurce?
>>51802701
http://seekingalpha.com/article/3741456-advanced-micro-devices-amd-at-raymond-james-technology-investors-conference-transcript?page=4
Server chips 2017. Desktop chips late 2016. Same as Lisa Su said a couple weeks earlier at the earnings conference.
It's almost certain the desktop chips are taped out if they are aiming for late 2016.
>>51802662
whatever your HS&F can handle.
>>51802676
>Zen is claimed to be a 70% IPC improvement over Bulldozer
who is claiming this?
it's like a smear campaign by competition.
>claim drastically unrealistic performance increases and hope it goes viral
>then act as if it's a big flop because of it
>>51802895
Its closer to 80%. Sounds like you just never paid any attention to reviews, or you just don't know how to do math.
See: >>51799692
>>51802926
>Its closer to 80%. Sounds like you just never paid any attention to reviews, or you just don't know how to do math
zen isn't even close to release yet. stop shitposting faggot.
>See: >>51799692
for cherry picked applications :* )
>>51803018
>I'm a tech illiterate retard
We know how Bulldozer performed.
We know how Piledriver performed.
We know how Steamroller performed.
We know how Excavator performs.
We know that Zen has a 40% IPC uplift over Excavator.
A 40% IPC uplift over Excavator works out to about 80% higher than Bulldozer. The scores would be even higher if Steamroller and Excavator chips benched had L3.
Those figures didn't come from cherry picked benches either.
>>51802808
Yes, they said they would be talking about the taping out of Zen, not that Zen hadn't been taped out. They probably have some kind of documentary that they're producing about the making of Zen and why its so awesome that they'll release before the launch to build hype.
>>51803192
Keep in mind that the man speaking isn't a native English speaker, and he sounds like a moron 100% of the time. Public speaking and broken English just don't mix. Seeking Alpha did a really sloppy job on the transcript on top of it.
>>51803231
>revenue ramp happening in 2016
>portfolio of revenue in 2017
Those are pretty clear statements coming from the chief financial guy.
>>51803231
Also important to keep in mind that he's a financial guy, not an engineer, so for him the release of a marketing video about Zen taping out is more important than when it happened.
>>51794793
>>>51793464
I don't think you know what "meme" means.
>>51803292
Speaking of things "tapping" out instead of taping out however is not a clear statement. This is what people have been fixating on, whether or not Summit Ridge has already taped out. No one knows what Pajeet Durka Kumar was trying to say since he barely speaks passable English.
>>51803292
>revenue ramp happening in 2016
I interpret that as their new enthusiast products launching. 8 core Zen, and the second gen Fury cards. With the mainstream products following in 2017.
>>51793464
AMD has been a meme since the mid-2000s. Decent bang for the buck if you're poor or cheap, always inferior to Intel.
>>51798448
>Or are you saying intel lied about the specs?
Wouldn't be all that surprising.
>>51803325
The question was about server chips. Those might not be taped out yet for a 2017 release.
>>51803326
Arctic Island GPUs are scheduled for first half of 2016. Couple months after Pascal.
>>51803375
but anon, your life is a meme
>>51793464
It's going to be about Skylake performance I don't believe they can beat Intel. There literally no reason wait if you just want preformance and have money.
>>51803036
>Those figures didn't come from cherry picked benches either.
The 40% for Zen is something we got from preview slides only. It could fall short in actual samples due to unforseen issues or it could be a cherry picked performance number itself. Every new arch that AMD introduced since the Athlon64 had a shit to of bugs at start and only became good after some respins, so I won't be surprised if Zen ends up falling short and only achieving that 40% at Zen+.
Still even if it is just better than Excavator and scales accordingly on FX chips, then we already have a good desktop chip in the Sandy Bridge range.
zen is a meme regardless of how good it turns out to be
AMD wants to be the enthusiast brand. Are we fucked with pricing then? I bet their 4 core CPU will be $300 and 8 core will be $500.
>>51803375
>AMD has been a meme since the mid-2000s. Decent bang for the buck if you're poor or cheap, always inferior to Intel.
I think you mean late 90s. Their K6 series offered such a great price/performance for your average desktop, that Intel was forced to launch the Celeron line.
And in the early to mid 00s, AMD was DESTROYING Intel with their desktop chips. Athlon, Athlon XP, Athlon64, X2, all were better than what Intel had.
>>51803485
That was because Intel was pissing about with NetBurst trying to get to 10GHz for marketing.
>>51797627
>skylel
>haslel
>>51803476
AMD can't really price their chips high even if they're better performers or Intel will continue to eat them alive because of user perceptions. AMD is going to have to beat Intel in price and performance.
>>51803596
Customer perception is a huge deal for AMD.
When I started streaming, I occasionally visited the Twitch reddit to ask questions (inb4 back to leddit). The fucking Elgato gaming community coordinator claimed once that you can use nothing less than an i7 just to use goddamn OBS for streaming a fucking console. Meanwhile, I was streaming AAA PC games at max settings, most of the time, using obs all day long with my 8350
People are that fucking dumb, and giving this advice to thousands of people. You can't make this shit up.
I went on to basically call him an idiot, but he never responded.
>>51803516
>That was because Intel was pissing about with NetBurst trying to get to 10GHz for marketing.
Not just that. The original Athlon legitimately beat the Pentium 3 too.
Zen will be between haswell/skylake and will have 4-6-8 cores with hyperthreading and they will be significantly cheaper than intel equivalents
thats a big fucking deal
>>51803707
Intel is fully aware of this. This is why they give out big discounts to retail workers to complete 'intel product eductation schemes'. So like prove you're a retail worker, go through some online training (involving questions like "What should you recommend to granny who wants to do both video watching AND web browsing at the same time? WRONG, it's the i7. The high performance i7 is perfect for all streaming, internet and tasks"
Shit like that. Probably about 30 minutes of this and you get massive discounts. Intel does this every so often.
>>51803732
>they will be significantly cheaper than intel
I don't think so
>>51803774
What makes you say that? Because if AMD's 8-core zen chip is more expensive than Intel's i7-5960X then I'll have no choice but to go with jewtel.
>>51803769
Goddamn... AMD really is the only company that isn't evil, aren't they?
>>51803845
I don't know about these things, but I've read somewhere that the new fab. process is more expensive
And that they won't be able to churn them out in high enough numbers for the processors to be cheap
>>51803769
not using xeon on 775
literaly 30 usd for i3
>>51803848
You can't be evil when you're behind. It's easy.
I very well remember the days when Google could do no harm.
>>51803884
But they weren't evil when they were ahead, either, by my memory. Google just got too damn big and corrupt.
>>51793464
>I waited for Skylake but was disappointed
Thinking of getting skylake.
What was so disappointing about it?
>>51803864
Nah, jewtel wants less than 500 bucks for its i7 6700K processor which is based on 14 nanometer lithography. AMD would have no trouble selling their 8 core chip for under 500 bucks.
>>51803969
Rumormill hyped it up as the biggest performance leap since Core2Duo.
Zen will revolutionize CPU's. AMD poised for a global comeback. The name itself "Zen" is already awesome. Wait for Radeon Zen. They will add a picture of a meditating monk in their products that symbolizes whoever uses AMD products will achieve enlgihtenment while using their computers.
>>51803969
Depends on where you're coming from.
A lot of vocal rich "kids" and by kids I mean adults that are easily influenced by hype on /g/ were thinking it was the next big thing. A GPU like jump in prefromance.
What they got was a typical Intel 10% improvement with new features.
If you're on a i5 2500k it's worth to upgrade. If you're below that it's 100% worth it no questions asked.
But these rich hype /g/ kids had i7 4770k or i7 5820k or E3 Xenons or something.
So they all got buttmad when it basically just X99 going mainstream with DDR4 ram, M2 slots on the mainboards, USB 3.1 and C and so on...
Basically if you're a haswell or similar no need to upgrade, and ZEN is going to be around Haswell, Skylake rank so no need to upgrade then too.
Moores law in CPUs is snail speed now. A lot of people are still going by with 2008 PC like me. I had a E6700 with DDR2 and a 460 GTX like 4 months ago and I can couldn't the game I couldn't play on two hands.
I jumped to a i7 6700K at 385€ and it's a huge performance boost I even switched to W10 for the wow a new OS a effect. But honestly not much changed. It's no where near the Pentium 4 to Core 2 Duo jump or the Pentium MMX to AMD Athlon jump...
Basically /g/ can hype up CPUs as much as it wants but they're not that special to begin with a X4 860K or an old i5 2500k will play all games today just fine if your target is 1080p60 with say a 970 or R9 390
>>51796930
Instil gahno
>>51800850
Still has no L3 cache, which would affect performance.
Still an APU chip.
AMD products are advanced compared to Intel simply because of hthe name. "ADVANECD Micro Devices.
While Intel is just Integrated Electronics. Sounds Generic. lel.
An honest question.
AMD proved that it can make a CPU/GPU APU thats pretty okay with the PS4/Xbone.
So why isn't it releasing these types of APUs on the market to flood out intel and nvidia?
I doubt it's that hard to make new socket for APU with better cooling and sell these say for 500$ that basically combine some 8350 with a say R9 380x or something.
You could make PCs a lot more compact and steambox like etc...
>>51802676
>Skylake is a 25-30% IPC improvement over Sandy Bridge
With all CPUs clocked to 3Ghz, Skylake is a flat 25% improvement over Sandy bridge, with the niche case Dolphin emulation test skewing the results considerably.
If we remove the lowest and highest tests (3dPM - single threaded's 5.5% and Dolphin's 69.3%) this becomes just under 20%.
http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/9
Again I say this,
Sandy is about 30% over Piledriver, Skylake is 20% faster than sandy,
Zen being even 30% over Excavator (which without an L3 cache, is 17% faster than Piledriver) puts it only marginally behind skylake.
Zen == Skylake
>>51804168
Yes but the main issue is memory for the iGPU. Motherboards would have to include at least 1GB of GDDR5/HBM vRAM soldered on the board which would make them expensive. Dual-channel DDR3 is simply not enough for GPUs with more than 384 gcn cores.
In the end it's just cheaper if you make a 860K+ R7 250X desktop build.
>>51804168
They already sell them, also 8350+R9 380x would stomp any console.
>>51804260
Zen has double the cores of Skylake. There are no signs of Intel planning 6 or 8 core CPUs for the mainstream.
>>51804275
Here is the build I was talking about:
http://pcpartpicker.com/p/DGhj6h
Only ~$300.
>>51804168
>basically combine some 8350 with a say R9 380x
But that's not even close to what they've done with the PS4. They've combined eight netbook processor cores with a 7870 underclocked to the level of a 7850, so in other words a laptop r9 290m.
Excavator cores are much bigger, so you'd only get four of them. The GPU would be heavily bottlenecked by the speed of the system memory, so you'd have to limit it to, say, 512 stream processors, which is well into the realm of diminishing returns.
Yeah, I really wish AMDS would release a processor like that. It'd probably be amazing in a laptop with a variable TDP between 12 and 35 watts.
>>51804168
APUs for high end desktops don't make much sense yet because of cooling issues.
They are planning "PS4-power" Zen+Artic Islands APUs for 2017 and huge APUs for servers/HPC. 16 Zen cores on one APU
>>51804366
But then you realize Carrizo is useless when it's limited to a 15w TDP, which every OEM that uses it has done.
Nice shill attempt anyhow.
>>51804717
It's competitive with the ULV core i3s and it offers decent performance without sacrificing battery life.
Go look up benchmarks, you'll see that the 35 watt mode adds at MOST 40% extra performance so it's not worth it. No 35 watt CPU is.
>>51804075
>and I can couldn't the game I couldn't play on two hands
I know what you were trying to say but god damn if I didn't have to read this like 6 times.
>>51804075
>what they got was a typical Intel 10% improvement with new features
And by that, you mean a 2-3% improvement over Broadwell, which is underwhelming even by Intel standards.
