>present Ryzen >demo with a Titan x >'present&

no one even has the card yet "rough estimate" bait more.

>>58343474
I see you left out the part where the card was throttling OP, but don't worry, I've got your back:

>Though that performance might not sound so impressive, it's worth noting that all of the demo system's vents (including the graphics card's exhaust) were taped up, and it's quite likely the chip was sweating to death in its own waste heat.

>>58343542
>Though that performance might not sound so impressive, it's worth noting that all of the demo system's vents (including the graphics card's exhaust) were taped up, and it's quite likely the chip was sweating to death in its own waste heat. By my rough estimate, that puts the early Vega card inside somewhere between the performance of a GTX 1070 and a GTX 1080 in Doom. If that's an indication of where consumer Vega hardware will end up, we could have a competitive year to look forward to from AMD

>>58343688
between 1070 and 1080 in the most AMD friendly game out right now is actually terrible performance, will probably trade blows with the 1060 ti (when its released)

>>58343474
I honestly expecting Vega to trade blows with the GTX 1080 up till 1440p and then pull ahead at 4k.

>>58344763

thats little vega

big vega will fuck the 1080 up and be very competitive with the titan xp

>>58343474
Are you fucking retarded? The point of those benchmarks is that people can make comparison with already existing hardware, it would be fucking stupid to demo ryzen+vega in the same system because then people don't know if the improvements are due to Ryzen or due to Vega.

>>58344179
Well shilled

>>58344179
How's that 1080ti looking ... oh wait

>>58344780
And that's what, 4 years away?
Vega will match the last nvidia flagships and the current nvidia midrange, just like now, and will come out too late to dominate the market any more than a 480 dominates the brand new 1080p market now

>>58344780
No. AMD demoed one Vega GPU, and it was the flagship. They explicitly fucking state that its flagship Vega.
https://www.youtube.com/watch?v=Ku5OMWYVKSs

AMD themselves know by now they cant compete with Nvidia

Oh cool so they're gonna be potentially competitive with nvidia's 2016 lineup... In 2017.

>>58344940
Says flagship, but doesn't indicate whether that is Vega 10 or 11.

>>58345742
Flagship has one meaning.

>>58345812
Yeah, top end for release. That doesn't say Vega 11. Vega 11 could well still be under development.

>>58344940

That is vega ten, look at Raj holding Vega 11 chip, see the two 8gb HBM modules? that is vega 11, vega 10 was shown at CES because its the only one they have working right now

We won't see both cards till May, count on it

>>58345839
>That is vega ten, look at Raj holding Vega 11 chip, see the two 8gb HBM modules? that is vega 11, vega 10 was shown at CES because its the only one they have working right now

What should Vega 10 look like in comparison?

>>58345825
cool, so we can expect to get Vega 11 by the time the 1180 is out?

>>58345856
He doesn't know since hes making all of that up.
AMD made pretty clear statements. They showed off flagship Vega. Flagship has one singular meaning, it is the top end. The Vega GPU shown to the press, the one used in the Doom demo, that is the top of the line.

This one retarded anon has been posting his rampant speculation as if its fact for weeks.

AMD, just wait™

>>58343545
i still wanna know, why the hell would they gimp it like that

>>58344854
>The point of those benchmarks is that people can make comparison with already existing hardware,

Exactly this.

>>58346501
Two reasons.
1 - no sneak peaks at Vega through the vent holes
2 - to keep Nvidia guessing at how powerful it is

File: 1483476950793.jpg (118KB, 600x697px) Image search: [Google]

118KB, 600x697px

>Dips to 30s and tearing in their own video

You can't make this shit up. At least they dared to actually show frames unlike that fucking shill Zen stream where they showed some insignificant CPU benchmark without any results or score.

>>58346759
AMD showed Ryzen performance in Blender, and in Handbrake. Both of those are real world workloads. They even released the files they used so the public could test them themselves.

Why are you shitposters all universally so fucking stupid?

Where are those benchmarks showing a GTX 1080 can do doom at 4k ultra 70fps?

>>58346791

Because anime

File: yu.jpg (117KB, 900x1300px) Image search: [Google]

117KB, 900x1300px

>mfw poojeet shills already pull a rx480
>b-but it wasn't meant to compete with 1080

>>58346605
not only that, but if there was a fuck up, they have someone else to blame,

nvidia crashes with the amd cpu, its a VERY public 'look, nvidias fantastic drivers' moment, or if the cpu can't feed the gpu 'look, even our competition can't feed the gpu enough data to max it out'

its win win either way, using your cpu or your gpu at this point is just unnecessary risk. once it launches you use it and stand by it to the death but till them, let others shoulder the burden.

File: hatfan.gif (69KB, 675x475px) Image search: [Google]

69KB, 675x475px

>>58343474
>system's vents (including the graphics card's exhaust) were taped up

Now I understand what did they mean by MAKE SOME NOISE.

File: AMD ZEN.webm (1MB, 720x1280px) Image search: [Google]

1MB, 720x1280px

reminder

>>58347141
Fuck you, I can't stop laugh tho

>>58346501
so their fangays can make excuses for poo in the loo engineering

File: 1473692108454.jpg (71KB, 552x661px) Image search: [Google]

71KB, 552x661px

>>58347141
delete this !!!

>>58347141
lol

File: amd3.jpg (66KB, 568x612px) Image search: [Google]

66KB, 568x612px

>>58347141
dumb intelshill

>>58347141
People are crazy

File: 1425622646527.gif (3MB, 307x267px) Image search: [Google]

3MB, 307x267px

>>58343474
>Techreport
>not intel/njewdia shills

>>58347531
>AYYMD IS YEARS AHEAD OF THE COMPETITION
>EVERYONE IS A SHILL
>>>/x/

>>58346791
>They even released the files they used so the public could test them themselves
Against what? Nobody has the cpu how can you compare it?

>>58349692
...Because we saw the Ryzen be benchmarked, so we know it's time? Which you can then compare to your own time. What a novel idea.

>>58349721
We only saw their benchmark at their own presentation of a still work in progress product that might be released Q1 of 2017 maybe if we get lucky.
Let me know when actual people who don't work for AMD marketing team get their hands on it to actually test the hardware.

File: page63.png (236KB, 830x1049px) Image search: [Google]

236KB, 830x1049px

>>58343474
>'present' Vega
>demo with a 4790k

Now I understand, Ryzen suck on gayming

>>58349803
They're looking pretty good on the power consumption chart, a few tweaks and they can surely beat that FX processor up top

>>58343474
>b..but ryzen + vega can run 4k Doom @ 60+fps

Dont forget that the Ryzen logo is a poo smear

>>58343474
So....AMD is going to price the Vega card around $500, which will make Nvidia drop their 1080 price but leave the somewhat competitive 1070 relatively unchanged, yet also open a price war for the cut down Vega to fight the 1070.

And then I can only assume driver improvements will close gaps, making the Vega cards a situation not unlike the 290/290X versus the 780ti/Titan

>>58350500
Depends on how much cost they can cut.

Disable half the I/O easy enough reduce stacks to 4-high to improve yields and reduce costs. Gives you a comfy 8gb across a 2048-bit bus for a solid 512gb of bandwidth which is higher than the TitanXP.

Bin the chips a little bit for yields sake and as long as saturation is solved and they can pushed that core clock high enough you easilly got a 500$ card that smashes the 1080 in 4K which is huge since the 1080 is not a 4K 60fps ultra card even with its 2000+mhz overclock and it costs us 650$

What I really want to see is a 50% bigger die with over 6000 shaders and the full 4096-bit bus enabled on 16gb of HBM2.

I would sell a kidney to have that raging inferno of a grahics card.

>>58349756
CanardPC did a few weeks ago, shitpost more desperate shill

>Can't even run maxed doom in 4k at 60 fps which is super fucking heavily optimized for AMD in Vulkan.
>Better than 1080
>Better than 1070

Fucking lel.

>>58346791
>Blender
>Handbrake
>Real world performance

I will give you Blender more people are using that and Cinema4D because of their ease of use and on top of that AMD does support Blender being an open source 3D tool.

But let us be honest with ourselves here nobody really uses handbrake or x product. We all know Nvidia can do so much with their cards even in real world use with Cuda. I render videos with cuda and I also render 3D in Cinema4D with a 760 and it still does a great job even on that front despite it being a few years older now. I will even be buying a 1060 because AMD has not proven that they have a plan.


AMD needs to start proving it can do what users use it for rather than pick and choose random shit to make it look good.

>>58343474
>they should demo there early pre production hardware silicon on top of more pre production silicon with pre-beta drivers

You're fucking retarded. The card was running pre-beta debugging drivers with little to
no Vega specific optimization on a early ES that is probably @ 1/3 the clocks that the production silicone will be running.
>>58349803
You do realise that was a crippled early ES @ 3.15ghz and the most recent steppings are 3.6ghz base and 4ghz turbo?

File: IQVrz.png (84KB, 299x288px) Image search: [Google]

84KB, 299x288px

>>58352100
A single half second stutter on early pre-production silicon running pre-alpha drivers with a CPU intensive debugging layer means less than nothing m8.

Why is /g/ so illiterate on large silicon production?

>>58352349
Cherry picking 10 seconds of gameplay on settings that aren't even maxed is truly enough to convince me! WTF I LOVE AMD NOW!

Oh wait we don't know literally anything.

Good to see there's no reason to upgrade from my HD 7950 yet.

>>58352314
Got a source on those steppings?

>>58352406
There is it's called Nvidia 1070. It's inexpensive and it's better than AMD.

File: 1456645624578.jpg (23KB, 381x250px) Image search: [Google]

23KB, 381x250px

>>58352381
>we don't know literally anything

I wouldnt say we don't know anything. We know some technical details about the arch but how it will perform several months from now is still up in the air. What we are seeing from the compute performance shown on Vega based Radeon Instinct slides we at least know it beats the new Titan in raw power. We also know that most of AMD's work this gen is being put into keeping the card bussy. There is no conceivable way that the final production silicon will be below 1080 levels unless the consumer version is heavily cut down due to yields.

>>58352314
>assumptions out of my ass: the post

>>58352443
(F4) stepping

>>58343474
>cutting of the quotes
Faggot

>So how's it play?
>High-level architectural discussions are one thing, but everybody wants to know what Vega silicon will look like in shipping products. AMD constantly demurred about actual implementation details of its Vega chips last month, but we've speculated that the Vega-powered Radeon Instinct MI25 would pack 4096 stream processors running at around 1500 MHz thanks to its 25 TFLOPS of FP16 power.

>AMD did have an early piece of Vega silicon running in its demo room for the press to play with. In the Argent D'Nur level of Doom's arcade mode, that chip was producing anywhere between 60 and 70 FPS at 4K on Ultra settings. Surprisingly, the demo attendant let me turn on Doom's revealing "nightmare" performance metrics, and I saw a maximum frame time of about 24.8 ms after large explosions. I also noted that the chip had 8GB of memory on board, though of course we couldn't say what type of memory was being used.

>Though that performance might not sound so impressive, it's worth noting that all of the demo system's vents (including the graphics card's exhaust) were taped up, and it's quite likely the chip was sweating to death in its own waste heat. By my rough estimate, that puts the early Vega card inside somewhere between the performance of a GTX 1070 and a GTX 1080 in Doom. If that's an indication of where consumer Vega hardware will end up, we could have a competitive year to look forward to from AMD. Until we learn more about Vega, though, we're left with only this tantalizing taste ahead of the card's first-half-of-2017 release.

File: 1269966652200.jpg (39KB, 400x456px) Image search: [Google]

39KB, 400x456px

>>58347141
My fucking sides

>>58352487
The card is not scheduled for release until q2. It first taped out about a year ago and they only had functional cards for a few months now. It is impossible for this to be anything more than very early silicon. Do you have any idea how complicated making a new architecture is?

>>58352487
>Deflection: the story of a shitpost

File: db8.gif (133KB, 600x356px) Image search: [Google]

133KB, 600x356px

>>58352526
Plot twist they used Nvidia 1060 to in their desktops and did not want anyone to notice.

>>58343474
>75fps on Ultra Nightmare @4K with Vulkan
>weaker than 1080 that only pulled 45fps with OpenGL

okay bud

File: 1483272193031.png (47KB, 212x233px) Image search: [Google]

47KB, 212x233px

>>58352568
>Plot twist
They used an AGP card and the results were just a video with staged actors

File: C1aelJ4UUAEG0za.jpg (115KB, 1200x900px) Image search: [Google]

115KB, 1200x900px

>>58352314
>>58352443
>>58352487
Pic related is the ES used at CES.
3.6GHz base clock, turbo up to 3.9
The ES string shows version F3, but that doesn't exactly correspond to stepping. Higher number indicates newer stepping though and some people assume that it's stepping B0.

As for F4, CanardPC hinted that it's already in the wild with 3.6 GHz base and 4 GHz turbo.
https://twitter.com/CPCHardware/status/817044837358780416

>>58345825
Are you dum?

This vega really ties the gaming market together. (c) The Dude.

>>58352591
oh look a retard

>>58352615
Fuck me, that looks great.

>>58345812
flagship was the fiji for amd not the fury x in particular...
vega is the flagship WE DONT KNOW if this is the little or the big vega..

also why people are so stupid amd clearly stated already that vega is going after volta which is due late on 2018 thats why they leapfrogged and is about to undercut the high end

Why is there so many people on this board that lick Intel's balls so hard?

I might use Intel now, but I would love to see AMD be competitive, all the better for me in the end.

>>58343474

last demo was Vega + Zen tho.

and even with FIJI drivers, it beat both the intel combo and the nvidia combo.

so having a Zen cpu + vega gpu will give the absolute best performance.

im eager to see what the 4core will clock like, and if they manage to beat the 6700k

Will Zen support GPU-passthrough?

>>58343474
youre retarded m8

all companies want to show off their products in their best light. AMD CPUs are heavily multi-threading oriented, and nvidia has great multi-threading performance on its drivers. AMD GPUs on the other hand have single threaded drivers, so they naturally run better on CPUs with great single threaded performance.

They're going to be playing catch up until the end of days

It's sad, really

>>58354311
If Intel keeps releasing shit with 0-1% improvements in performance, AMD might catch up in a few years.

>>58343474
>>present Ryzen
>>demo with a Titan x
>>'present' Vega
>>demo with a 4790k

They're trying to be inclusive with other brands to show you could mix and match if you want.

Maybe they could convince people who like Intel CPUs to buy a graphics card or convince people who like Nvidia GPUs to buy a CPU.

>>58354559
nah m8, see >>58354288

>>58354630
Then why did they demo zen with nvidia??!?

>>58346501
I'd like to think it's so they can push it further after ti release, but I doubt it.

>>58354689
zen is still a multi-threading heavy architecture, its been AMDs thing since they beat intel to the punch on dual core processors

>>58352247
>says AMD should make sure their cards do cuda well
>people use AMDs cards for Nvidia things

File: pepefinger2.jpg (42KB, 436x436px) Image search: [Google]

42KB, 436x436px

Ryzen is the last hope for amd. They cant spoil the benchmarks with using shit tier gpu.

>>58354519
They're only doing that because they have no competition. Intel undoubtedly has some really good stuff just sitting around to smack down AMD at the possibilty of them releasing anything better than Intel has out.

>>58355428
>Intel undoubtedly has some really good stuff just sitting around
If by stuff you mean lawyers, yes.
It's much more likely that they just didn't bother developing better stuff in the first place- because you need to explain otherwise why intel would spend money on developing something and then not bother making money from it.

>>58355445
They will make money from it. Who said it has to be today?

>>58352314
>You do realise that was a crippled early ES @ 3.15ghz and the most recent steppings are 3.6ghz base and 4ghz turbo?
still behind on gaiming vs software

>>58358018
Okay? Then get the quad core which will have higher clocks.

>>58358045
it will be behind on software now