/w4v/ - Daily Wait for Vega Thread - On Tour At PDXLAN

File: Uj0c6ZDXe4otVKFPIMps0IrioCd-1KAoHZJVRPe1Fqc.jpg (76KB, 673x767px) Image search: [Google]

76KB, 673x767px

>>61486850

File: 1499056875663.jpg (83KB, 671x427px) Image search: [Google]

83KB, 671x427px

>>61486927
>kekcupid

>>61486850
Vega good yet?

File: 515.jpg (273KB, 1683x766px) Image search: [Google]

273KB, 1683x766px

What does this mean?

>>61488475
What I keep seeing is that a $230 GTX 460 has a larger die size than the $599 GTX 1080.

They lost a hell of a lot of money on the 460 or are robbing people blind with the 1080

>>61486927
>7'11", Overweight
>Speaks some Afrikaans, Dropped out of High School
>Smokes regularly, Does drugs often, Has kid(s) and wants more

>>61488556
Guess why nvidia's revenue skyrocketed with Pascal?

>>61488556
The 1080 was a more complex node than the 460. Complexity always results in higher manufacturing costs.

>>61486850
1070 performance for 1080 price and 290X2 power consumption cofirmed.

AMD is finished and bankrupt.

>>61490523
>Trades blows with a Quadro P6000 in geometry bound pro workloads
>1070 performance
?

>>61490620
Wait for confirmation of flawed hardware that can never be fixed.

>>61490642
So your claim is that the following new Vega uarch features are disabled in the Vega FE gaming drivers because the silicon is flawed and they CANNOT be activated?

So primitive shaders, tile based rasterization, advanced clock gating, AVFS, and HBCC are all broken in silicon?

>>61490789
Yup the hardware is borked gonna need new silicon.

>>61486850
Now that they confirmed it is just a 1080 that costs $50 less I think I will skip this gen. I was hoping they would do better given they had an extra 13 months on nvidia.

>>61491073
>I was hoping they would do better given they had an extra 13 months on nvidia.

>nvidia
>dedicated to building GPUs
>can spend billions on R&D

>AMD
>have to split money between CPUs and GPUs
>Zen is essential for the future of the company
>RTG gets starved of cash
>Wonder why shit like Vega can get pushed out of the door

>>61491073
Where have they confirmed this?

>>61491101
Poland demos. They alluded to the price being $200 cheaper for the system.

The differences were card, psu, and freesync vs nsync. So psu and freesync is $125-$150 cheaper on the vega build so $75-$50 cheaper for the gpu. That test, all other tests, leaks, etc, place performance at standard clocked 1080.

At best vega oc may edge over 1080 oc, but we will have to wait until they release to know for sure.

>>61491152
They will probrably do a similar demo in Portland so may get more specifics.

>>61491152
>Poland
Budapest is in Hungary.

>>61489127
2.6 times more expensive, really?

>>61491098
If your hardware is that uncomoetitice, why release at all?

>>61491226
Because there's profit to be made from it, dumbo.

>>61491152
They said $300 cheaper, and didn't say which parts were included in that calculation, so this post is entirely speculation. People still hope it's $100-$150 cheaper than 1080.

>>61491152
>So psu and freesync is $125-$150 cheaper
What, they used a cheaper PSU for Vega which is a >300W card compared to a 180W reference 1080? Or do you mean more expensive PSU + cheaper monitor still comes out $125-150 cheaper in total?

>>61491390
>cheaper PSU's are a good idea when dealing with novidya gpu's
https://www.youtube.com/watch?v=WAbl0fLY06U

>>61491435
Yes, you can spend less when your single graphics card doesn't come with a 375W TDP.

>>61491275
If the card actualy was cheaper than a 1080, they would just say the card is cheaper and that would be the selling point.

Them dragging in other components but being very mysterious about what models they actualy are and then saying its cheaper makes me believe it might be the same price or even more expensive than a 1080.

Which would make sense considering how big the die size is and it has HBM2

>>61491450
And then that retard comes back asking for advice on what cards he can get for his PSU because he is now stuck with it because he does not want to buy a new PSU with each GPU.

I bought an AX 850 for £124.24 in 2010 and it has seen 3 CPU's (Phenom II X4 955 to 5930k) and 6 GPU's (some SLI/crossfire) without any concerns.
I can't see it needing replacing any time soon, so overall it made much more sense to get that initially rather than buying a specific PSU for a specific setup. Buy once and buy right.
Won't see me complaining about one GPU costing £5 more a year in electricity or that I was too dumb to buy a decent PSU to support whatever comes next.

>>61491569
That's irrelevant, my question was in regards to the demoed Vega system supposedly using a cheaper PSU, which makes no sense for a card which could literally be a model with more than double TDP than a 1080 (375W vs 180W).

>>61490966
yeah, the silicon breaks when playing games but not while doing profesional 3d rendering and compute

>>61491692
Don't worry anon, I'm sure RX Vega will be as fast in games in 2021 as a GTX 1080Ti, AMD is bound to get some drivers which make use of all its wondrous features for general purpose gaming too by then.

>>61491681
You are comparing a AIO Vega FE against a blower 1080 FE (which is 186W btw)? Granted neither seem to be very good at maintaining clock speeds.

>>61491197
Poland, Hungary, who cares? They're all gypsies.

What time will that portland lan start?

Would anyone even mind another Hawaii? Shit still's a beast.

>>61490620
You guys have a real misunderstanding of the differences in the pipeline between a gaming workload and a rendering workload
You can keep thinking it's going to be a 1080ti, but that doesn't matter when every single leak shows otherwise, you can keep screaming about "THEY TURNED OFF THE MAGIC SMOKE FEATURES WHEN THEY ARE ENABLED THEY WILL BE LIKE TWO 1080S" it's never gonna be true

Just take the loss this time around, and we'll all start waiting for Navi, you guys are hindus you're good at waiting

>>61494508
Hoo boy are you fucking retarded?

>>61494535
That was a great counter argument you presented
Really activated my smuckers

>>61494581
Oh, have a (You).

File: 1490986860095.jpg (52KB, 736x915px) Image search: [Google]

52KB, 736x915px

>They think that tweet is about vega

>>61494628
what tweet?

>everybody always talking about the high end
Why does nobody care about the budget cards?

>>61491226
Something needs to be shoved into APUs

>>61494856
Because the high end is where technology is driven, and the technology in high end cards one year is given to budget option cards two or three years later.
Yes, that's a two or three year wait, but that's the price you pay if you want to pay the lowest possible price for worthwhile upgrades.

>>61494967
that's saying that a rx560 performs like a r9 380 anon

>>61495003
Well not quite that soon then, but the point stands.
The 470 is pretty safely in the "budget" zone I'd say.
Easily the same power or better than the 7970Ghz edition.

>>61494508
>/v/ermin literally doesn't even know enough about GPU's to know what any of the new Vega uarch features even ARE, much less what they DO
>Decides to vomit out his utterly worthless opinions anyway
/v/iddiots should be gassed, I swear.

>>61495441
He said pipeline! That means he's smart!

>>61494856
Because they are available

@61495499
Sorry, you'll have to work a little harder if you want a (You).

File: 1482365057279.gif (3MB, 524x276px) Image search: [Google]

3MB, 524x276px

>>61495529
>@
Newfag detected.

>>61495499
>>61495441
Incredible counter argument
I'm sure your triangular v8 rasterization dual cylinder 4.5 liter compute engine will beat nvidia this time
What is your excuse going to be when vega inevitably shits out? Backhand tactics? Muh silicon jew?

File: 3080965.png (37KB, 711x533px) Image search: [Google]

37KB, 711x533px

>>61497058
I understand that you are very frustrated that no one will engage with your mindless shitposting, but unfortunately you haven't demonstrated enough of an understanding of how GPUs work to even be worth arguing with.

The minimum barrier to entry is to be able to explain the apparent massive contradiction between being able to trade blows with a Quadro P6000 in geometry bound pro workloads while apparently also performing like a 1080 in games.

Both should not be able to be true at the same time.

File: AMD-RX-Vega-GTX-1080-Ti-GTX-1080-GTX-1070-3DMark-Performance-Wccftech-1480x1215.jpg (148KB, 1480x1215px) Image search: [Google]

148KB, 1480x1215px

>>61497896
>both should not be able to be true at the same time
Quadros used to be good shit at gaming
So far you've shown no benchmarks with your magic 1080ti features

Here are some gaymen workloads though

>>61491275
It's actually $200 more expensive than 1080 since the Gsync Monitor was $500 more than the freesync one

>>61498293
You cannot even manage the minimum necessary level of coherence to actually get across whatever nonsense it is you are trying to say, and here you are trying to argue about things you don't understand in the slightest.

Please try again, but in English this time, autismo.

>>61498690
Not an argument
Burden of proof is on you for your magic gaymen features

>>61499363
Unfortunately I'm unwilling to take the time to hand hold you through understanding how GPUs work well enough to even have an argument about that Vega FE's numbers do or don't mean for Rx Vega, so you will have to remain frustrated.

>>61499495
I've already provided valid performance scores from RX Vega
You've provided nothing but FE rendering workloads

>>61495980
You are stupid....

>>61499513
>Valid
Why would they be valid? Literally every single new Vega uarch feature is still disabled in drivers. if you deactivate all AVX instructions on an Intel CPU and then benchmark it in AVX workloads, is that a valid benchmark?

>>61499595
*in gaming drivers, I should clarify

>>61499595
You have no proof of your special features being of any benefit or having any affect on performance
All you provide is speculation, I gave numbers

>>61499612
>proof
http://www.gamersnexus.net/guides/2990-vega-frontier-edition-undervolt-benchmarks-improve-performance

GN's article proves that Vega's Advanced Clock Gating isn't working at all, and that Adaptive Voltage & Frequency Scaling are in a failsafe mode where 1.2v is being used regardless of clocks. This is why GN was able to increase the performance of Vega FE by undervolting it to 1.09v and ~280w.

http://digiworthy.com/2017/06/30/amd-vega-frontier-tile-based-rasterizer/

Here's proof that Tile Based Rasterization is not working on the current Vega gaming drivers.

http://www.overclock.net/t/1633406/so-what-went-wrong-with-vega/120

HBCC and Primitive Shader are also disabled currently in gaming drivers.

Just to maybe get to to at least TRY and educate yourself a little bit, here is a nice article about what the implementation of TBR meant for Nvidia in going from Kepler to Maxwell:

http://www.anandtech.com/show/10536/nvidia-maxwell-tile-rasterization-analysis

>>61499882
Thanks for the links.

>>61499882
Now show the scores for when all of these apparent disabled features are enabled
If you can't, then stop speculating it will improve the performance by 7000 points

>>61499926
>Now show the scores for when all of these apparent disabled features are enabled

I will. In nine days. Until then no one can have a solid idea of how Rx Vega will perform as there are simply too many wildcards that could have too large an effect on the gaming performance depending on whether they are implemented and how effectively they are implemented at launch.

And anyone who wants to claim any different must, at minimum, explain the apparent massive contradiction between Vega's gaming performance and geometry bound pro application performance and why they believe Vega's new uarch features won't make much of a difference to gaming performance or power consumption.

>>61500026
Then stop implying 1080ti performance when every imperial metric points to otherwise

>>61500354
>Then stop implying 1080ti performance when every imperial metric points to otherwise
The only thing I have implied is that you don't understand the slightest thing about GPUs and your entire participation in this thread has consisted of nothing but utterly mindless shitposting.

I never made any claim that Vega would perform at a specific level in gaming, because there are still far too many unknowns, but that also means that all of the /v/iddiots screeching about how Vega is DOA or 1080 level are talking complete nonsense and should go back to their own board if they can't be bothered to actually fucking learn something about the subject they are shitposting about.

600W 2Ghz Vega on LN2 when?

>>61500418
Literally everyone knows it's not going to be 1080
It's between a 1080 and a 1080ti

>>61500511
>knows
Just go back to /v/. You clearly have no interest in making the slightest effort to learn anything, and your mindless shitposting will fit in much better there.

File: 2017-07-22-001004_767x366_scrot.png (92KB, 767x366px) Image search: [Google]

92KB, 767x366px

convince me not to click order

>>61500675
Do it
the 1080ti is better than Rx vega

Or wait for Navi

>>61500675
Do you even have a 4k display? or a fast enough CPU for 144hz gaming? If you don't have either free-sync is always cheaper, and you don't need a hybrid 1080ti in any case

>>61501116
whats a good cpu for 144Hz gaymen?

not them

>>61500675
Just buy a FTW3, you don't need water cooling on Pascal.

>>61501148
R7 or i7@5ghz probably Ryzen so you don't have to delid your CPU and run it on a 240mm AIO

>>61501148
A delidded, water-cooled, overclocked (at least 4.9) 7700k.

An overclocked R7 would also be fine (at least 3.8, but getting a chip that could hit 4.0 would be ideal).

>>61501165
Hmm I guess 1700 it is

>Just Wait

Fuck that. Managed to scoop up a 1080 in that recent newegg sale a couple days ago for 499. Literally got the last one by a matter of seconds. Vega rx has been looking like trash all along and there's nothing to suggest it will be much faster than a 1080 any way.

>>61501212
AMD defense force incoming

Are there going to be AMD cpus with integrated gpus this year? I want to upgrade but I don't need a graphics card

>>61501423
Yes

>>61501423
So just buy ryzen with any cheap $25 GPU?

>>61501212
the heck do you need new GPU in summer? there are no games to play.

File: 1483678747438.jpg (138KB, 1200x675px) Image search: [Google]

138KB, 1200x675px

>>61501212
>Good goy

>>61500468
1500W LN2

File: asus.jpg (399KB, 1350x843px) Image search: [Google]

399KB, 1350x843px

>>61501148
Buy a good board and at least 4133mhz ram with that i7

5ghz with ddr4 4500

>>61501212
>Trades blows with full gp102 in geometry bound pro workloads
>nothing to suggest it will be much faster than a 1080 any way.
?