PCPer Live! Radeon RX 480 Live Stream with Raja Koduri!

Subject: General Tech, Graphics Cards | June 29, 2016 - 12:03 PM |
Tagged: video, rx 480, raja koduri, radeon, Polaris, live, giveaway, amd

When it comes to GPU releases, we at PC Perspective take things up a level in the kind of content we produce as well as the amount of information we provide to the community. Part of that commitment is our drive to bring in the very best people from around the industry to talk directly to the consumers, providing interesting and honest views on where their technology is going. 

With the pending release of the Radeon RX 480 based on AMD's latest Polaris architecture on Wednesday, June 29th, I am excited to announce that Raja Koduri, SVP and Chief Architect of the Radeon Technologies Group will be joining us in studio to talk about the RX 480 and AMD's plans moving forward.

View Full Size

The AMD Radeon RX 480 Graphics Card

There is much to discuss. AMD and the RTG have promised that the RX 480 will be a revolutionary product, improving on performance per watt and performance per dollar in a way that no other AMD architecture has done. And the drive to include dramatically more gamers in the rising world of VR gaming will be an impressive feat as well, if they can pull it off. Topics like architectural improvements, asynchronous compute, multi-GPU and more are on the docket. You definitely won't want to miss it.

View Full Size

Radeon RX 480 Live Stream with Raja Koduri and Ryan Shrout

10:30am PT / 1:30pm ET - June 29th

PC Perspective Live! Page

Need a reminder? Join our live stream notification list!

The event will take place Wednesday, June 29th at 1:30pm ET / 10:30am PT at http://www.pcper.com/live. There you’ll be able to catch the live video stream as well as use our chat room to interact with the audience, asking questions for me and Raja to answer live. 

View Full Size

Raja is one of the more open and honest people in this highly competitive landscape and every time we have had the ability to do an interview he has provided insightful, and sometimes very new, information. 

As a price for hosting AMD in the offices, we demanded a sacrifice: in the form of hardware to giveaway to our viewers! We'll have at least two Radeon RX 480s to giveaway during the live stream but I am pushing  to get a bump in that count; we'll see if I am persuasive enough. All you have to do to win on the 29th is watch the live stream!

View Full Size

Some Radeon RX 480s will be up for grabs!!

If you have questions, please leave them in the comments below and we'll look through them just before the start of the live stream. Of course you'll be able to tweet us questions @pcper and we'll be keeping an eye on the IRC chat as well for more inquiries. What do you want to know and hear from Raja or me?

So join us! Set your calendar for this coming Wednesday at 1:30pm ET / 10:30am PT and be here at PC Perspective to catch it. If you are a forgetful type of person, sign up for the PC Perspective Live notification list that we use exclusively to notify users of upcoming live streaming events including these types of specials and our regular live podcast. I promise, no spam will be had!


June 27, 2016 | 03:36 AM - Posted by Jann5s

Questions:
what is RTG's stance on 3way+ Xfire and "enthusiast keys"?
How did the marketing guys let you launch the card without LEDs or RGB?
Are there any real density differences between 14nm GloFo and 16nm TSMC, or are the two nearly identical?
Related, are you expecting lots of frequency headroom on 14nm?
Can you comment on why current gen seem very sensitive to overvolting, i.e. large power increases with little extra clocks, is it due to tunnelling?

June 24, 2016 | 06:11 PM - Posted by Anonymous (not verified)

Will the giveaway be for US and Canada only?

June 24, 2016 | 06:12 PM - Posted by WhyCry (not verified)

Now, that's the interview I want to see.

June 24, 2016 | 06:59 PM - Posted by D1RTYD1Z619

I hope he doesn't flake and cancel

June 25, 2016 | 06:07 AM - Posted by Master Chen (not verified)

There's literally zero reason to do so.
RX 480 turned out to be a truly great product and they secured Scorpio and Neo contracts once again, and AMD's stock is on the largest rise for the last 5 years, so why the hell should they be cancelling interviews and being ashamed of themselves?

June 24, 2016 | 06:22 PM - Posted by John H (not verified)

What is the length of the RX 480 PCB? will we see coolers that short from OEMs? (i.e. Coolermaster Elite 110 only fits 8.3" x double width graphics cards).

Is there a faster bin of the Polaris 10 die expected in the near future? (RX 490 or 480X?).

Do we get a new ruby demo?

June 24, 2016 | 06:46 PM - Posted by Klyde (not verified)

The RX 480 is the fully enabled polaris 10, RX 470 will be a cut down version of it. Whatever is above that(RX 490?) will be Vega.

June 29, 2016 | 07:25 PM - Posted by Cade E Doebele (not verified)

You can always bin for faster clockspeeds. It's just not always reasonable to do so. You could also do faster RAM.

There could be a 480x with the same specs, except higher clocks and maybe more RAM (say 6gb & 12gb options).

Although you're probably right, the 480x being vega like the 490 is very likely.

June 24, 2016 | 07:02 PM - Posted by Anonymous (not verified)

"What is the length of the RX 480 PCB?"

This is just an estimate, but I believe it is pretty accurate.

According to this schematic I found (pic),the PCI-E x16 connector is 115mm long. I then took the picture of the PCB side of the RX 480 from LegitReviews (link) and worked a little bit of MS Paint magic.

Based on my measurements and conversions (pic), I believe the PCB is 193mm or 7.5". This puts it just 2mm or .08" longer than the Fury X PCB, and taking into account the less-than-precise method of measurement could easily have it at the exact 191mm of the Fury X PCB.

"Do we get a new ruby demo?"

Dear AMD and Nvidia: Please bring us fully-modern versions of Ruby and Dawn, respectively. Bonus idea: include a benchmark mode.

June 24, 2016 | 07:18 PM - Posted by Anonymous (not verified)

And having looked a little more closely at said schematic, and the measurement, I think I drew the red line incorrectly - it should be a bit longer, extending at least to the end of the "foot" that engages the PCIE slot clip. I'm going to re-measure it because I'm like that, but I'm guessing that adjustment ought to bring the 480's PCB length down a little bit, probably closer to 7".

June 24, 2016 | 07:40 PM - Posted by Anonymous (not verified)

Yep, right at 7" (pic). (Using instead Page 22 of a PCI-SIG presentation (pdf) for a known length to use as a reference.)

June 29, 2016 | 11:29 AM - Posted by John H (not verified)

You are awesome! thanks for this 'measurement' :)

June 24, 2016 | 06:24 PM - Posted by Diego Ignacio Falcone (not verified)

Yes! this will be a great interview ;) Raja is the Man!! the giveaway is open worldwide??

June 24, 2016 | 06:29 PM - Posted by PCPerFan (not verified)

I like this guy.

June 24, 2016 | 06:30 PM - Posted by 16stone

Nice Scoop!

June 24, 2016 | 06:33 PM - Posted by Prodeous

Guess this question is probably not for him, but what about pricing of the previous gen cards?

For me i'm more interested in OpenCL support, and CUDA integration as mentioned by few AMD posts

June 24, 2016 | 06:37 PM - Posted by Anonymous (not verified)

Prepare uranuses for pile of VR crap from dr.Bombay.

June 27, 2016 | 07:39 PM - Posted by Anonymous (not verified)

It's totally possible to irrationally hate everything about a company without expressing it using thinly-veiled racism.

June 24, 2016 | 06:47 PM - Posted by Bassie (not verified)

I would like to know what AMD's plans are for the 490. Will it be the lower end Vega card? Is it going to come out before the Vega cards come?

One more question. AMD talks a lot about HDR. I am sure that HDR is going to be great but so far not a single company has announced that it has planned to produce a HDR monitor. Just television screens (not useful if it doesn't contain FreeSync or if the refresh rate is too low). Does AMD know something we don't?

June 25, 2016 | 10:51 AM - Posted by Mageoftheyear (not verified)

I'd like to hear a bit more on 10-bit HDR monitor availability too.

Apparently the operation of HDR and FreeSync are mutually exclusive but I've heard no statement on why that is. I'd like to know from AMD if it's a bandwidth restriction on DisplayPort 1.4 / HDMI 2.0, display related, software related or related to the GPUs themselves.

For instance, if simultaneous operation of FreeSync and HDR are not possible with Polaris cards, will Vega be any different?

I'm psyched that this interview is happening, I really enjoyed the last one with Richard Huddy.

June 25, 2016 | 11:15 PM - Posted by Ty (not verified)

Wait, where was it said that HDR and freesync would be mutually exclusive?

June 26, 2016 | 06:57 PM - Posted by Mageoftheyear (not verified)

https://np.reddit.com/r/Amd/comments/4jwzpt/what_was_the_518_webinar_abo...

Apparently, during the AMD webinar on the 18th of May 2016.

To clarify my own question for you, I was asking about simultaneous operation of both technologies, not if only one or the other could be used in a monitor (i.e. choose between modes, not choose between a FreeSync monitor or a 10-bit HDR monitor).

While we're on the topic (@pcper) I'd also like to know if we will be getting Polaris powered laptops with FreeSync and HDR support - even if we will have to swap between those features.
For example, a single laptop where FreeSync is more important to me on CS:GO but 10-bit HDR is more important to me on (hopeful example:) Pillars of Eternity.

I would also hope that these panels used in the laptops would support the Low Framerate Compensation mode of FreeSync.

June 24, 2016 | 06:49 PM - Posted by Bassie (not verified)

Oh, one last question though I am sure someone else will ask it or have asked it. Why on hell does the 480 only use 36 out of 40 CU cores?
Yield issue? Future 480X? Apple?

June 28, 2016 | 01:57 AM - Posted by Kraaketaer

Sorry, what? Every single spec sheet (including all manner of leaks) pegs the Polaris chip at 36 CUs. The RX 480 is fully enabled.

June 24, 2016 | 07:20 PM - Posted by Anonymous (not verified)

When will the APU's built on an interposer package be ready for the laptop market, I'm looking forward to an interposer based APU with at least 4/8 Zen cores and a larger Polaris GPU die, and the whole APU on an interposer package having HBM2 memory. I know that AMD will be building some Server/Workstation APUs on an interposer package for 2017, but when will there be derived consumer APU on an interposer variants.

More needs to be done to demonstrate both Vulkan/DX12 explicit multi-adaptor and its use with your new GPUs, and that includes the potential usage for non gaming graphics software. Will there be more work between AMD and Khronos to create a direct OpenGL to Vulkan translation layer, so any OpenGL based graphics software like Blender, etc. can make use of Vulkan's explicit multi-adaptor via a Vulkan to OpenGL conversion layer that allows for the gradual porting of the openGL code over to using the Vulkan graphics API.

I believe that it's better to have the graphics API/OS and applications in charge of load balancing between the various GPU/other processors on a computing system than having that multi-GPU adapter done under complete GPU driver control. I'd rather keep the GPU driver as simple as possible and move all the multi-GPU adapter management into the graphics APIs/middle ware/OS and standardizing multi-GPU adapter management in the graphics API/OS software stack.

I see that AMD is working towards having some FPGA compute added onto the HBM stacks along with HBM memory dies on some HPC/server/workstation APU/GPU on an interposer variants, so will there ever be FPGAs available on any consumer gaming/graphics SKUs to add some new Graphics API features to the FPGAs to allow for better execution of any new graphics/compute features not implemented in ASIC hardware at the time of the ASIC's release.

Also with respect adding features to the HBM stacks, will there ever be an addition to the JEDEC HBM standard that will allow for a non volatile memory die to be added to the HBM die stacks. So with a NVM die right on the HBM stack there can be a fast in the HBM stack NVM to HBM DRAM memory store for paging/texture swap with a bulk memory XPoint/other bulk memory high density NVM die providing for some very fast access DRAM to NVM via TSVs/other wide in the HBM stack connection fabric. I'd really like to see the JEDEC standard extended to having a NVM standardized controller added to the bottom logic die of the HBM stack that can be made to work in concert with the DRAM die/s controller to work DRAM to NVM transfers in the background under the OS command/HBM firmware/HBM controller/s command, for some very fast HBM DRAM die to HBM NVM die Paging/swap or texture store/swap for Games/other uses.

June 25, 2016 | 01:12 AM - Posted by Paul EFT (not verified)

Shut up.

June 25, 2016 | 08:56 AM - Posted by Anonymous (not verified)

F-off peckerwood!

June 24, 2016 | 07:23 PM - Posted by Butthurt Beluga

You guys should smoke a cigar with Raja on stream.

http://www.overclock.net/image/id/12165378/width/900/height/900/flags/LL

June 24, 2016 | 07:25 PM - Posted by ALlen_One (not verified)

Can XFire be more reliable to help make 2 RX480's more attractive than 1 GTX 1080?

Can game devs rely on engines using DX12/Vulkan to get most of the multi-gpu performance gains without doing additional multi-gpu programming of their own?

June 25, 2016 | 02:12 AM - Posted by ALlen_One (not verified)

Sorry, "multi-gpu" = "explicit multi-gpu."

June 25, 2016 | 09:20 AM - Posted by Anonymous (not verified)

Yes the gaming engine SDK/tools/libraries will be available in the gaming engine development packages to automate the hard parts for the script kiddies! So that explicit multi-adapter can be given a software framework in the gaming engine that allows for easy management of explicit multi-adapter. There will be automated code designers that allow for easy coding for all the close to the metal DX12/Vulkan GPU load balancing functionality/other functionality.

Load balancing among the different processors CPUs, GPUs, whatever should have been the responsibility of the OS and the Graphics APIs/subsystems of any real OS in the first place, with the GPU's/CPUs/Other processor's drivers kept as simple as possible and under the management of the OS/Graphics APIs/subsystems. All GPUs/CPUs/whatever processors should appear to the OS as just another compute resource able to be utilized at any time without any of that graphics adapter switching from one or the other adapter foolishness.

The gaming engines should be calling the Graphics API/OS, and the graphics API's and OSs should be able to allow the gaming engine to call on whatever processor resources are plugged into the system/local bus, including any external GPU/CPU/other processors on the network, or thunderbolt/other technology connected devices.

June 25, 2016 | 11:42 AM - Posted by Paul EFT (not verified)

Shut up.

June 25, 2016 | 03:26 PM - Posted by Anonymous (not verified)

F-OFF PECKERWOOD, your SMOM is calling you!

June 24, 2016 | 08:04 PM - Posted by zangza

AMD is looking up, I know we have seen the performance so far, but the performance will most definitely differ when the cards are released, hopefully for the better.

June 24, 2016 | 08:12 PM - Posted by prosp3ctus (not verified)

A bit if a tough question but how do you manage to constantly improve an architecture when you are stuck on the same node for a while

When the HD 7970 came out it had a 352mm die, the R9 280X based on the same chip came a while later, at the same die size obviously.

But then when AMD followed up with the Tonga based 380X, not only was the die slightly bigger, it didn't show any meaningful performance improvement over the (considerably older) 7970 chip, despite Tonga's improved features like tessellation and compression.

Sorry about that but as a tech and AMD enthusiast the 7970-380x stagnation just fascinates me.

Thx Raja.

June 24, 2016 | 08:23 PM - Posted by Murica702

AMD doing a good thing

June 24, 2016 | 08:54 PM - Posted by Genogerian (not verified)

I made this for Raja. Could you please deliver it to him?

http://i.imgur.com/TNgtNRe.png

June 24, 2016 | 08:58 PM - Posted by Kyneaz (not verified)

Awesome i won't miss it !

June 24, 2016 | 09:33 PM - Posted by marian4q

I am very happy to see more AMD guests here on PCPer!!!
We need more RED Tom Petersens.
I hope 480 will bring more balance on the GPU industry. As I do Zen will.

June 24, 2016 | 10:43 PM - Posted by arbiter

Big shock is AMD finally learning? Instead of sending a PR marketing guy that only knows what the PR sheet says they are SENDING someone with technical knowledge.

June 25, 2016 | 10:08 AM - Posted by Anonymous (not verified)

And still your bog standard gaming Git is not going to understand anything about technology, and will always fall back on simple minded fanboy like reactions to any real technology discussions that may have implications for both AMD and Nvidia improving their GPU/CPU products.

To the average gaming Joe who does not understand anything about gaming/computing technology this will be more about some green team fighting some red team and nothing more.

Look at some of Nvidia fanatics when AMD/AMD's memory partner first announced their HBM technology, well those fanatic Gits where all over HBM as unnecessary! Until that is AMD and SK Hynix made the effort to have HBM made into a JEDEC standard, and Nvidia jumped onto the HBM bandwagon. It's not hard to see HBM's advantage is in the super wide 1024 bit connection to each HBM stack that allows HBM to be clocked much slower than GDDR5 but still deliver more effective bandwidth at a much more power saving lower clock rate than any open standards DDR memory technology created before HBM. But there will still be some that will be stuck on higher clocks = better analogy, and will not be able to wrap their single brain cell around the fact that wider parallel connection fabrics clocked lower are better than narrow connection fabrics clocked at a much faster power wasting rate.

Most certainly AMD's RTG director is going to be better than any marketing monkey, but the average gaming Git is so below the median on the IQ level that it's going to be an eternal uphill battle trying to have any discussion that rises above that Green Team versus the Read Team mentality lowest common denominator that the marketing profession excels at manipulating, and the average marketing monkey is not at all that smart.

June 25, 2016 | 11:38 AM - Posted by Paul EFT (not verified)

Shut up.

June 25, 2016 | 03:32 PM - Posted by Anonymous (not verified)

F-Off mindless GIT, those webbed feet you got from your SMOM must be good for swimming at the toxic cement pond! You must be Spuckler's offspring with his SWIFE Brandine!

June 28, 2016 | 02:51 PM - Posted by Stefem (not verified)

Dude relax! :) is the same both ways AMD to NVIDIA fan and NVIDIA to AMD fan, didn't heard people that GTX 1080/1070 are just rebranded overclocked card? if you listen to those mindless people you will get crazy ;)

June 25, 2016 | 12:04 AM - Posted by nivek (not verified)

How has the recent surges in stock price affected the company internals? Has hiring / morale been boosted?

June 25, 2016 | 12:22 AM - Posted by skysaberx8x

Hi, this going to be awesome!

My question is:
Will we notice any differences between the 8GB vs 4GB variances in terms of perfomance?
Knowing there was very minimal difference in the Hawaii chip (290x)

June 25, 2016 | 10:58 AM - Posted by KVragec (not verified)

I do not think so. Yes i know i'm not mr Raja Koduri but 8gb of vram will only improve performance in games which can make use of it

June 27, 2016 | 05:03 PM - Posted by Dr. Poop (not verified)

Yes we've heard that with this new hbm2 tech that soon there will be 32gb gpus but at the moment it sounds like 8gb isn't fully utilized by many games. Is this going to change?

Also will xpoint tech play nicely with these cards? Or will motherboards change dramatically in next 5 years?

Does he know Mr. Macintosh?

June 25, 2016 | 01:03 AM - Posted by Anonymous (not verified)

AMD fanboys and shills rejoice!

June 27, 2016 | 07:43 PM - Posted by Anonymous (not verified)

http://i.imgur.com/q7Sq0CL.jpg

June 25, 2016 | 03:16 AM - Posted by Kevin (not verified)

I just like that they are doing a card this powerful at the 200 dollar price point.

I cannot Imagine what the 490 is going to be like!

June 25, 2016 | 03:23 AM - Posted by Xaris Mitakos (not verified)

What about us that we live in Europe and we cannot watch live the interview....will we be able to participate in the giveaway? Or is it only US and CANADA giveaway??? Unfortunately I live in Greece and cannot watch the live show....
Thank you in advance for a prompt reply

Best

June 25, 2016 | 03:53 AM - Posted by Arjun Anand (not verified)

Awesome! I will be looking forward to this :)

June 25, 2016 | 06:02 AM - Posted by Anonymous (not verified)

Hey Ryan, hey Raja!
Please include these questions on the livestream:
1) Radeon stance on high bandwidth multi GPU interconnect
2) What is the current adoption rate for GPUOpen initiative - are there any key partners in gaming industry to announce?
3) How do the next few years look in terms of transition to Heterogeneous System Architecture?
4) With the new manufacturing process what is the limiting factor for not having enthusiast desktop graphics with APU, is it the heat?

June 25, 2016 | 08:39 AM - Posted by Anonymous (not verified)

5) How does the AMD TrueAudio technology from 2013 compare to nVidia's announced sound ray-tracing simulation?

June 25, 2016 | 08:45 AM - Posted by Anonymous (not verified)

6) There have been images floating around the web, so I guess we'll see a new Overdrive interface in the Crimson drivers, is that specific to RX400 series or will we see that on the 28nm GPUs as well?

June 26, 2016 | 05:09 PM - Posted by Anonymous (not verified)

up-vote on this.

June 25, 2016 | 11:19 PM - Posted by Ty (not verified)

Another question he probably won't answer but...

That slide that shows NAVI with a "scalable gpu architecture"

...

Does that mean a bunch of small die gpus can be mashed together to create a single effective megazord like combination of gpu power that would act as a single ultra large die gpu as opposed to the current models in cf/sli where they are separate and have limitations?

June 27, 2016 | 02:14 PM - Posted by RushLimbaughisyourdaddy (not verified)

It very well could be that "scalability" on AMD's roadmap for Navi means multi-GPU on something like an interposer. A great video by ADored TV that goes over AMD's Roadmap which covers this idea:

https://www.youtube.com/watch?v=aSYBO1BrB1I

June 27, 2016 | 09:54 PM - Posted by Stefem (not verified)

Please ask this to someone who know and can answer (like Raja) not someone that act like a (fan)boy and call his very approximative work "analysis" and "benchmarking" c'mon! you read PCPer!

June 25, 2016 | 09:22 AM - Posted by qwekqwek (not verified)

Hi, if you can please also ask this question:

There are still some features in Catalyst that are not in Crimson yet, mostly on the Advanced Video settings. Do they have any plans on bringing these back in Crimson? (De-noise, Noise reduction, De-blocking, Deinterlace, etc).

The Crimson edition is focusing more on gaming and left out the good media enhancements, which is why I am still on Catalyst 15.11 on my builds except for that one PC which I game on.

June 25, 2016 | 10:19 AM - Posted by Anonymous (not verified)

Anxiously waiting to see if a 470 can be underclocked from 110W TDP down to 95W for use with http://www.hd-plex.com/HDPLEX-Passive-Video-Card-GPU-Heatsink-System-for... without losing too much performance. Would love to hear about this specifically, as well as more general information on passive GPUs in the future with performance/watt being a big emphasis in AMD's marketing.

June 25, 2016 | 11:15 AM - Posted by Anonymous (not verified)

RAJA! RAJA! RAJA!

Awesome!

June 25, 2016 | 12:56 PM - Posted by Mageoftheyear (not verified)

I know today is about Polaris but will discrete Vega GPUs be making their way into high-end laptops?

I'm hoping to one day buy a laptop powered by Zen & Polaris on the APU side (Raven Ridge) augmented by a discrete Vega GPU to take advantage of multi-adapter support.

P.S. ~ Ryan, very important: run a poll during the interview for those in favour or against Raja regrowing that glorious face-hugger of his. RTG just isn't the same without it. :3

June 26, 2016 | 08:18 PM - Posted by StephanS

Was Polaris / Vega designed mainly in China ?

Just asking because Raja flew to china to congratulate the design team, correct ?

Also the India/China team seem to have skyrocketed in size in the past 5 years. Why is AMD firing people in north America and instead growing the team in Asia ?

Serious question, because I'm curious to see if the US is faced with a lower education system compared to where we where in the 50s.

It seem India / China is teaching computer science , but the US is not. (Sating this because in silicon valley I dont believe in 2015 CS was part of the curriculum of any <19 year old)

June 27, 2016 | 02:17 PM - Posted by RushLimbaughisyourdaddy (not verified)

A really good theory is that labor is just simply cheaper over there.

Depends on where you want to go to school at and what major you are going for. If your major is something like nursing or architecture, then of course you are not going to take Computer Science courses!

June 27, 2016 | 12:44 AM - Posted by Twvalk (not verified)

Ryan / Raja
I have a 380 currently, I want to get a 480 and do 480 primary duel cards setup with the 380. I'm running win10 / dx12. Is there enough support for this from drivers at launch of the 480? Will there be, if not at launch? When?
If not, when I get a 480 should I keep the 380 for second monitor, instead of duel card use as two separate cards?
Please let me know how support for previous gen cards will work with new gen cards.
Also will I get full benifit from the 480 using an older amd mb?
System:
Asus m5a99fx pro2.0
Fx8350
16gb ram
Saphire r9 380 4GB nitro
Ocz arc 100 120gb ssd
Wd 1tb hd

Thank you!

June 27, 2016 | 01:37 PM - Posted by mikesheadroomlive

Lisa Su detailed the Polaris lineup as ranging from $99 to $299, but we're told that the entire Polaris 10 family is the 460, 470 and 480. With the top-end 480 being $229, is there another product coming to fill that hole up to $299, or was there a miscommunication?

June 27, 2016 | 09:52 PM - Posted by normpearii (not verified)

Vega 10

June 27, 2016 | 05:51 PM - Posted by Dr. Poop (not verified)

Ryan you should grow out a competitive mustache

June 27, 2016 | 06:47 PM - Posted by Anonymous (not verified)

folks, there's no need to watch the stream unless Raja turns you on, we already got all the information we need:

- RX 480 will be available only in the reference variant for a couple of weeks after its launch
- the cooler is tra.sh that probably costs 20 cents to manufacture and makes the card throttle hard with DEFAULT clocks
- the chip is a dog that likes to overheat
- gaming performance is slightly under the GTX 970, that's right, a 2 year old mainstream card
- VR is a tragedy, GTX 970 shits on RX 480 in all things VR, which is sad for a card marketed as for its VR capabilities

Overall a disaster of a launch. This card won't sell for $239. Even GTX 1070's p/$ ratio is better than this pos.

June 27, 2016 | 07:47 PM - Posted by Anonymous (not verified)

Ladies and gentlemen, here we have a perfect example of a post that you should simply snicker at, and then scroll on.

June 27, 2016 | 09:53 PM - Posted by goosegrease

Any idea how long the livestream is going to be?

June 27, 2016 | 09:54 PM - Posted by normpearii (not verified)

Why was the R5/R7/R9 naming convention dropped? Seemed to work fine for 200 and 300 series.

June 27, 2016 | 10:39 PM - Posted by arbiter

trying to shead the rebranding image that seem to only come with that naming i beat. Pretty much every card with R5/R7/R9 on them were mostly all rebrands.

June 28, 2016 | 01:53 AM - Posted by Anonymous (not verified)

My suspicion (and only suspicion) is that they didn't need the added segmentation in the lineup. Assuming that they're not going to be doing X-series GPUs (480, 480X, 470, 470X, etc.) and assuming that it's because the entire lineup for this release and in the near future consists of just 3 GPUs, it seems silly to further segment your three cards. What are they going to do, R5-460, R7-470, R9-480? What would even be the point?

But they still really really like that X.

So, screw it. Make it the RX-series. They still get the "R" for Radeon, they still get to use that X they like so much, they have their 3 GPU models that don't really need to be segmented any more than they are - it just makes sense to drop it.

It does make me wonder about the precedent, though. They now have RX graphics cards. They've said in the past (I dunno if they've said it recently) that the non-APU Zen parts would be FX branded. Will the APU's drop their A10/A8/A6 segmentation and become AX? What if their DDR3/4 memory becomes MX branded and their Radeon SSDs become SX branded?

I'm not sure whether I think that would be a good idea or a terrible idea, really.

June 28, 2016 | 08:04 AM - Posted by Anonymous (not verified)

Ask:
Is this a fully enabled chip and if not, when is the fully enabled version out?

June 28, 2016 | 10:35 AM - Posted by YTech

Since the AMD Fury GPU cards and today with the RX 480, I haven't seen OEM partners doing much to improve or change the style and design of the cards.

1- Are the partners restricted per their AMD agreement to modify the cards to their own IP needs?
2- Are the cards locked down by design where it's not possible to re-engineer them any further?

For example, look at the products by SAPPHIRE and XFX.
From the original design of the cards, they've added 3 fans and that's about it. But the liquid cooling models is generally the same, except maybe a better cooling fan.

If you look at the earlier cards, you see PCB redesigns, heatsink redesigns, and other neat features.

June 28, 2016 | 09:51 PM - Posted by Anonymous (not verified)

The Fury X and Nano were both specifically reference-only. As I understand it, the AIBs weren't allowed to make custom designs of those two cards.

The vanilla Fury, however, the AIBs were given free reign. Sapphire's Tri-X Fury used a reference Fury X PCB design, as did XFX's Fury cards, both the water and air cooled. But Sapphire's Nitro version of the Fury was a custom PCB; Asus' Fury Strix was a custom PCB; Gigabyte's Windforce Fury was a custom PCB.

Outside of the Fury X watercooler and the Nano air cooler, there hasn't really been an AMD reference-cooled card since the 200's, as far as I know. I remember seeing pictures of the 300 series reference cooler but I don't think I've ever seen a reference-cooled 300 for sale anywhere.

I'm not sure why Fury X and Nano were restricted by AMD to reference-only boards and coolers. I suppose it's possible that AMD considered them special, being outside their standard lineup pattern, and wanted that reference look and design associated with it.

I don't think there's any reason to believe there will only be reference RX 400 cards. Every other generation of cards has had custom PCBs, including the R9-300s. These will too, they'll just release a bit later than the reference cards, like past generations.

Fury X and Nano (and, I guess, the Pro Duo, if we want to include it) are statistical outliers, in this case.

June 29, 2016 | 10:55 AM - Posted by YTech

Thanks for the info :)
A little more info at the event would be nice too.

So far, the Fury and Nano appears to be AMD's best cards.
If the OEM partner could have approval to improve them to their needs, it could become a most valuable product. Then again, perhaps AMD has other plans for Vega that might be their new jewel.

June 28, 2016 | 10:53 AM - Posted by YTech

1- Roughly how many engineers have worked on the new RX cards?

2- How many of them enjoy gaming? How many of them enjoy VR?

3- Will there be a collectors edition card in the future that has all the engineers' name on the shroud or PCB?

Side Comment:

When you're aiming for space, data can get corrupted or lost, but most material will stay intact. Imagine picking up a lost AMD product in space 50 years or even a century from now ;)

Although odds aren't in our favour due to density and other phenomena, possibilities do exist and can become true. We do have satellites...

June 28, 2016 | 12:14 PM - Posted by Anon (not verified)

1) How long until custom designs are expected to be out?

2) Is the tessellation performance further improved from Tonga or did it stay on the same level?

June 28, 2016 | 12:24 PM - Posted by eternalozze (not verified)

is there going to be a under 75 watt part in this new RX line? all of the R series GPUs under 75 w that are available are just rehashes of models that came before.

I am asking this question because OEM's base systems are using like 240 w power supplies, IE my new Dell 3650 that i am using a hd 6570 on. It would be cool to put something modern in it for light gaming. The gtx 750 ti is the best bet right now but if AMD has a new under 75w part coming i would rather give AMD my money.

June 28, 2016 | 09:59 PM - Posted by Anonymous (not verified)

If it's AMD you're after, you'll be looking at the RX-460.

"Beyond the RX 480, AMD has also announced the RX 470 and RX 460. Little is known about the 470, but they mentioned that the 460 will have a <75W TDP. This is interesting because the PCIe bus provides 75W of power. This implies that it will not require any external power, and thus could be a cheap and powerful (in terms of esports titles) addition to an existing desktop."

Source

June 28, 2016 | 02:39 PM - Posted by Thedarklord

I am really loving all of these Tech industry people that PCPer is getting to come in and do reviews/chats/etc.

June 29, 2016 | 03:30 AM - Posted by NamelessTed

I am very interested to see if the RX 480 will have supply issues or not. Obviously the GTX 1080 and 1070 have been constantly sold out since their releases. This has caused retail prices to be significantly higher than the MSRP quoted by nVidia.

Why is everybody ok with increased prices because of availability? When Microsoft or Sony release a new console, and they say the price is $500 it doesn't matter if its sold out, retail locations still charge $500. The Wii was constantly sold out for months when it released but if a retailer had stock they were charging MSRP.

Has AMD ever considered trying to cap the prices on their GPUs or create some sort of incentive for retailers to charge MSRP? I think if a company says they are going to charge $200 or $380 for a product then that should be the price available to consumers.

If people want to buy a card and try to resell it for more that is their right. I feel like it is misleading to advertise a price with no expectation that a card will sell for that price on release.

June 29, 2016 | 06:24 AM - Posted by Dazzler (not verified)

1st:
In case RX 460 & RX 470 won't be released today.
Mr. Raja: Where are them?!?

2nd:
RTG's goal was VR for everyone under $200/$300, it seems that cheapest prices are only in USA. What's your opinion that if RX480 sells for over $300 outside USA, can you still consider that you have accomplished your goals.

June 29, 2016 | 07:01 AM - Posted by Gairfame2627 (not verified)

You've mentioned that the RX 480 WILL deliver $500 US performance. Since the R9 390x is below this, can we expect Fury (non x) performance? Also will this card sit at 5.5 or 5.83 Tflops?

June 29, 2016 | 12:05 PM - Posted by Cataclysm_ZA

My question for Raja:

I'm a writer for a South African tech/gaming website.

AMD invited some high-level journalists from websites across the globe to Macau to learn about the RX 480 and the Polaris architecture in general.

However, the amount of information you divulged to journalists attending that press conference, compared to the amount of info I learned about the card in the Tuesday briefing last week, is orders of magnitude higher.

Why did AMD choose to approach certain sites with more information about Polaris and give them a better understanding of the RX 480 and Polaris' capabilities, and give other sites less information?

Why choose this disjointed approach, especially if most of the sites will be reviewing the RX 480 during the launch window, and some will arrive at their conclusions and judgements of the card without knowing the full picture you showed to press at Macau?

I hope you can address this for me. I'm still looking forward to reviewing the RX 480 in the near future.

June 29, 2016 | 12:31 PM - Posted by qwekqwek

Will the RX 460 support native HEVC / h265 encode and decode?

June 29, 2016 | 12:40 PM - Posted by Earthworm Tim (not verified)

Are there any low profile enthusiast GPUs coming? I'm a big fan of low profile designs but feel left out by whats on offer.

June 29, 2016 | 12:44 PM - Posted by diggiddi

Are the crossfire multi gpu improvements going to trickle down to older gpu's, specifically R9 290/390?

June 29, 2016 | 12:56 PM - Posted by DK76 (not verified)

Twitch is down!!!

June 29, 2016 | 01:17 PM - Posted by KyonCoraeL

Question:
Earlier this year on initial announcement of Polaris, you had discussed the concept of HDR in games. There are standards in video such as HDR10 or Dolby Vision for Video. Is there such a standard for Video Games? If there is one, are there any existing AMD video cards and monitors which support it?

Thanks

June 29, 2016 | 01:26 PM - Posted by ppi (not verified)

Questions:
1. What is AMD's answer to nVidia's simultaneous multi-projection? Does not it invalidate your CF strategy in VR?

2. Why AMD cards have consistently lower geometry and texture throughput (especially fp16) vs. competing nVidia chips? (see e.g. http://techreport.com/review/30328/amd-radeon-rx-480-graphics-card-revie... ) Would not devoting more die area to these units benefit gaming performance more than SPs?

June 29, 2016 | 01:56 PM - Posted by slyons (not verified)

Could we see shorter length RX 480 cards? It looks like the reference cooler is quite a big longer than the PCB. Can we expect to see non-reference cards with shorter coolers that could more easily fit into a small form factor system?

June 29, 2016 | 02:40 PM - Posted by StephanS

So far no RX 470 or RX 480 4GB for sale in North America.

Also TH reported that the RX 480 BREAKS the pcie spec by drawing 83w vs the maximum of 75w.

Is AMD planning a board recall ?

Will AMD replace motherboard that breaks when used with a RX 480 ?

How about power delivery noise affecting on board audio ?

The GPU seem to have more to give. but the cooling & power delivery is limiting the card. (800mhz in furmark, even so the silicon can run at +1.4ghz )

So, will we see a 300w polaris 10 card with liquid cooling so it can run at 1.4+ghz without throttling ?

June 30, 2016 | 02:19 AM - Posted by D1RTYD1Z619

nevermind

June 30, 2016 | 03:31 PM - Posted by Anonymous (not verified)

Ryan, I'm very glad that you discussed pc vs console. This was a discussion much needed! The current generation of consoles is really good value for money and the next ones seem to be even better if they manage 1080p60 or incorporate freesync into the next consoles.

July 3, 2016 | 10:26 PM - Posted by Elbaz (not verified)

It is a great GPU specially price.

August 22, 2016 | 02:59 PM - Posted by BossTek (not verified)

Eh yo yo yo yo! Yeah yo! Yo AMD sucks yo because it not cost allot not gangsta yo yo! Yeah know what I sayin yo? Oh yo we at BossTek are better than anyone yo! We is because we say yo every other word. Yo.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.