Leaked AMD Fiji Card Images Show Small Form Factor, Water Cooler Integration

Subject: Graphics Cards | May 22, 2015 - 09:39 AM |
Tagged: wce, radeon, Fiji, amd, 390x

UPDATE (5/22/15): Johan Andersson tweeted out this photo this morning, with the line: "This new island is one seriously impressive and sweet GPU. wow & thanks @AMDRadeon ! They will be put to good use :)"  Looks like we can confirm that at least one of the parts AMD is releasing does have the design of the images we showed you before, though the water cooling implementation is missing or altered.

View Full Size


File this under "rumor" for sure, but a cool one none the less...

After yesterday's official tidbit of information surrounding AMD's upcoming flagship graphics card for enthusiasts and its use of HBM (high bandwidth memory), it appears we have another leak on our hands. The guys over at Chiphell have apparently acquired some stock footage of the new Fiji flagship card (whether or not it will be called the 390X has yet to be seen) and it looks...awesome.

View Full Size

In that post from yesterday I noted that with an HBM design AMD could in theory build an add-in card that is of a different form factor than anything we have previously seen for a high end part. Based on the image above, if this turns out to be the high end Fiji offering, it appears the PCB will indeed be quite small as it no longer requires memory surrounding the GPU itself. You can also see that it will in fact be water cooled though it looks like it has barb inlets rather than a pre-attached cooler in this image.

View Full Size

The second leaked image shows display outputs consisting of three full-size DisplayPort connections and a single HDMI port.

All of this could be faked of course, but if it is, the joker did a damn good job of compiling all the information into one design. If it's real, I think AMD might finally have a match for the look and styling of the high-end GeForce offerings.

What do you think: real or fake? Cool or meh? Let us know!

Source: Chiphell

Video News

May 7, 2015 | 09:51 AM - Posted by collie


May 7, 2015 | 10:03 AM - Posted by Anonymous (not verified)

Hmm no dvi/vga port? Slightly concerned about that, unless they provide a hdmi to vga AND hdmi to dvi adapter.

May 7, 2015 | 10:14 AM - Posted by Jabbadap (not verified)

Afaik vga was dropped already from hawaii cards(r9-290/x. And I mean really dropped, hawaii series has 2xDVI-D dual link digital only). So it have to be active adapter if you wanted to get vga out. Hdmi to dvi adapter is more possible though(digital to digital).

No fans on this? So I presume VRM:s are cooled within water loop(full cover block or not). Looking forward to see reviews from this.

May 7, 2015 | 11:12 AM - Posted by Anonymous (not verified)

Correct me if I'm mistaken, but there are no VRMs with HBM architecture? I thought it was on chip, hence the small form factor.

May 7, 2015 | 11:51 AM - Posted by Jabbadap (not verified)

heh I'm not talking about _vrams_, but _vrms_ aka voltage regulation modules, mostly some sort of fets...

...And hbm is vram too it just abbreviation of video random access memory, which hbm obviously is.

May 7, 2015 | 10:30 AM - Posted by SheepInACart (not verified)

HDMI and DIV-D have the same signal minus the fact that HDMI also carries extra info (inc trusted connection stuff) on other pins... so you can use a cheap (like >$5) cable or adapter.

May 7, 2015 | 12:11 PM - Posted by Anonymous (not verified)

Really?........vga?? What exactly are you worried about? That your brand new possibly $800 card will not work without a $5 adapter on your 12 year old monitor?......

May 7, 2015 | 03:39 PM - Posted by Bianchi4me (not verified)

Hey now, VGA compatibility is only the beginning for me. Then I have to split that VGA adapter down to an RCA plug, then take that down to speaker wire leads to go under the screw-down antenna terminals on my slide switch box that also supports my Atari 2600 and rabbit ears, then more leads to the UHF terminals on my black-and-white Sears Kenmore t.v. As you can see, being able to skip even one step is a godsend.

May 22, 2015 | 10:58 AM - Posted by Eric (not verified)

I use a KVM and for those you are looking at a price jump from ~$100 for VGA to about $600 for DVI in the same number of ports. That is why I still need VGA.

May 22, 2015 | 06:47 PM - Posted by Anonymous (not verified)

Bro, do you even remote? How archaic are you? I work in a large company, and we dont even use a kvm switch on our server stacks since they're all giant VM farms. Be it physical or vm... you can remote into it.

May 7, 2015 | 12:12 PM - Posted by Anonymous (not verified)

that is so irrelevant it hurts my brain

May 28, 2015 | 03:13 PM - Posted by wiak (not verified)

my prediction

AMD Radeon Fiji WaterCooled Edition
~750-1000mhz core
4096 shaders
4GB HBM, ~550GB/s
3x DP + 1x HDMI 2.0

AMD Radeon Fiji
~750-1000mhz core
4096 shaders
4GB HBM, ~500GB/s
3x DP + 1x HDMI 2.0

AMD Radeon Fiji Pro
~7500-1000mhz core
3680 shaders
4GB HBM ~500GB/s
3x DP + 1x HDMI 2.0

May 7, 2015 | 10:07 AM - Posted by pollcats38

hope its real and needs cooling raidiator to go with it.

May 7, 2015 | 10:09 AM - Posted by cbower82

Super interesting !! Excited to see if AMD can hit something out of the park instead of a line drive to the left fielder

May 7, 2015 | 10:14 AM - Posted by Anonymous (not verified)

Would fit BEAUTIFULLY in my Corsair 250D. I'll just have to wait to upgrade....

May 7, 2015 | 10:28 AM - Posted by Irishgamer01



May 7, 2015 | 06:00 PM - Posted by fade2blac

It appears to be an artificial approximation. There is at least one glaring discrepancy between the two images. The width of the 'aluminum' section on which the RADEON logo is placed is about twice as wide on the new image.

May 7, 2015 | 10:31 AM - Posted by DarkArterialGore

It's so cute!!! If AMD can sort out their crossfire drivers 2 (if the price is good why not 3) of these could power my next build.

May 7, 2015 | 10:33 AM - Posted by SheepInACart (not verified)

Like most early card shots this is probably either fake, or just not what it claims to be. But I really hope its real, a small form factor makes heaps of sense if your water cooling from factory, and factory AIO cooled cards have been very popular with both the r9 295x2 and board partner cards like the EVGA gtx980 hybrid. So hopeful, but not really expecting.

May 7, 2015 | 11:26 AM - Posted by The_Icon (not verified)

AMD needs to get back in the game. Lets have the glorious discrete GPU battle of the glorious early 21st century.

May 7, 2015 | 11:51 AM - Posted by Daniel Nielsen (not verified)

Hope its real, looks really cool. Been wanting to switch my 290X out, makes no sense to get a GTX 980 with this around the corner. I just hope thermals and power are better than the 290X.

May 7, 2015 | 12:07 PM - Posted by Speely (not verified)

Damn that thing is sexy. This is SO going into my ITX build-in-progress now.

May 7, 2015 | 12:12 PM - Posted by JL (not verified)

Hmm....water cooling? I guess the gpu core even hotter than maxwell titan.

I am not complaining about the smaller card size though. In fact, prefer if both vendors use smaller card size.

May 7, 2015 | 04:05 PM - Posted by Anonymous (not verified)

You are not going to see smaller cards from Nvidia until they move to HBM next year.

May 22, 2015 | 10:46 AM - Posted by Anonymous (not verified)

Or, thanks to the HBM (what is much smaller than DDR5 and is maybe closer to the GPU chip), it has a much higher heat density, and air is no longer sufficient to deal with it, regardless of the venting design.

May 7, 2015 | 12:28 PM - Posted by Martin Trautvetter

If true, good riddance, AMD reference blower!

May 8, 2015 | 07:07 PM - Posted by arbiter

Well the ref blower likely still used as they will have an air cooled version as well so could see said blower yet.

May 22, 2015 | 01:51 PM - Posted by JohnGR

I don't think so. The green press used the poor performance of the cooling system on the original Hawaii cards, to distract people's attention from the cards performance and paint a negative picture for the 290 series. I don't think AMD will introduce anything else than the water cooled card, leaving the air cooled cards to it's partners.

We could see a reverse in rolls here compared with the past, where the reference model was the cheap one and third party the more capable and also the more expensive versions of a card. With Fiji we could end up with an expensive reference water cooled card and plenty and (much) cheaper air cooled cards from AMD's partners.

May 7, 2015 | 12:28 PM - Posted by jerky33

I've been running Nvidia cards for years now, but would love to see AMD come out with something that can compete at the high end, I think it would really change the landscape of the market.

May 7, 2015 | 03:32 PM - Posted by Jmac (not verified)

Going the DisplayPort route make sense given that their FreeSync solution is dependent on the Adaptive-Sync feature set in DisplayPort.

I will definitely be considering a part (like this) from AMD when it comes time for my new Win 10/DirectX 12 build in the fall/winter.

May 22, 2015 | 10:30 AM - Posted by Anonymous (not verified)

Their "FreeSync" is actually VESA display port adaptive sync, "FreeSync" is just marketing, and the only free part is that it frees up a little bit more cash($$$) from the costs of the monitors that have the DP 1.2a standard built-in, as compared to those than come with G-sync. Make no mistake, all new technologies cost the monitor makers a good chunk of cash to adopt, and get certified to work with any standard, it's just that DP 1.2a is an open standard and will not have the usual crippling to the wallet markups that the proprietary system(G-sync) has.

May 7, 2015 | 03:42 PM - Posted by Mobile_Dom

Is there a benefit to going with full sized DisplayPort instead of Mini DisplayPort? Mini PD is way smaller than full sized, unlike normal HDMI and Mini HDMI.

it seems wasteful IMO to have 3 full sized DP ports when it seems like there could be 4 or more MiniDP ports instead

May 8, 2015 | 08:00 AM - Posted by Anonymous (not verified)

Chances are your display has full size ports so you would have to buy adapters to use the card which would be annoying.

May 7, 2015 | 03:55 PM - Posted by Anonymous (not verified)

Not sure what people expected. Once you put the memory in the GPU package, the only thing left on the board is the power supply components. The PCB routing should be trivial without the wide memory interface. Do we know whether it is 28 or 20 nm yet? At 28 nm, the water cooling would probably be needed because of the total heat output. At 20 nm, the water cooling might be needed because of thermal density issues rather than the total heat output. This may make dual cards a lot more reasonable size though; probably similar to current single gpu cards.

As far as this image is concerned, I would expect most manufactures to use a closed loop. There is some demand for custom loop parts though. I guess it could also be a closed loop card which is not fully assembled; if so, is the pump on the card, radiator, or separate? It is 2 slots thick so there probably would be room for a pump. Having the openings on the back seems suspicious though. It doesn't seem like it would have a fan in there, so why would there be vents on the back if it is fully water cooled? It could have air cooling for the VRMs, but I don't know how that would be set up with no blower intake on the front of the card.

May 7, 2015 | 04:15 PM - Posted by pdjblum

This is a quote from a news post on Tom'sHardware:

"Yesterday, during the presentation, AMD did not specifically announce the Radeon 300-series, but we assumed that's what they were talking about. In light of the new information, however, it is possible that not all of the 300-Series GPUs have HBM memory, or that this quarter's GPUs with HBM memory won't be part of the 300-Series at all. We recognize these possibilities because the specifications for the GPUs listed on AMD's website clearly state that they come with GDDR5 memory, and the model numbers range up to the Radeon R9 380."

Gather this one is the real deal, but sucks AMD is doing another round of re-branding with the 300 series they just announced.

May 7, 2015 | 10:09 PM - Posted by Anonymous (not verified)

If the Wikipedia entries are correct, the 3xx series is OEM only. It is unclear why they exist at all. Wikipedia also list 4xx series as mostly rebranded parts except for the new Fuji parts (490/490x/495x2) with HBM. Only disappointing thing so far seems to be only 4 GB of memory for new HBM parts. It may not matter given how fast the memory will be and maybe changes with DX12. Although, I think you still have to enable higher AA modes to actually go over 4 GB. Of course, Wikipedia could be totally wrong.

May 7, 2015 | 10:09 PM - Posted by Anonymous (not verified)

If the Wikipedia entries are correct, the 3xx series is OEM only. It is unclear why they exist at all. Wikipedia also list 4xx series as mostly rebranded parts except for the new Fuji parts (490/490x/495x2) with HBM. Only disappointing thing so far seems to be only 4 GB of memory for new HBM parts. It may not matter given how fast the memory will be and maybe changes with DX12. Although, I think you still have to enable higher AA modes to actually go over 4 GB. Of course, Wikipedia could be totally wrong.

May 22, 2015 | 12:41 PM - Posted by Heavy (not verified)

im pretty sure everyone has said this that hbm is only going to be in the high end gpu maybe just the 390 or 490 whatever they'll name it because its new tech and its expensive.on the whole rebrand thing im pretty sure their not going to do it,i think they'll be using they same gpu that in the r9 285.

May 7, 2015 | 09:42 PM - Posted by SiliconDoc (not verified)

It's a freaking grey pegboard, you can hang yard rakes and trimmers and a few pliers and some rope on it.


If they go with this people will hate it.

May 7, 2015 | 09:59 PM - Posted by Anonymous (not verified)

Agree. Definately needs to be more shiny.

May 8, 2015 | 09:53 AM - Posted by obababoy

So just because you want some crappy alienware looking aerodynamic video card does not mean that everyone wants that. The slot a videocard goes in has straight edges, not some stupid faceplate fan shroud. Grow up mang.

May 8, 2015 | 10:47 AM - Posted by al bundy

Fear not. They are releasing the Cheese Edition with jewel-encrusted spinnaz and rainbow LEDs. It also has a perfume dispenser you usually only see in public restrooms.

May 22, 2015 | 05:43 PM - Posted by Heavy (not verified)

those picks are fakes its just an r9 285 saphire itx verszion with photo shop heatsink/watercooler in reality the card can be smaller then this just look at the sa[[hie verszion of the r9 285 http://www.kitguru.net/wp-content/uploads/2014/09/DSCF2883.jpg

May 22, 2015 | 05:45 PM - Posted by Heavy (not verified)

ops my bad wrong link http://www.overclockersclub.com/vimages/sapphire_itx_compact_r9_285_oc/1...

May 7, 2015 | 10:09 PM - Posted by Gunbuster

I'm going to say fake based on the fact it's not slathered in the AMD/ATI cheap plastic red coloring.

May 8, 2015 | 11:00 AM - Posted by praack

actually nice design, clean sleek..

May 8, 2015 | 10:15 PM - Posted by Anonymous (not verified)

Inconspicuous is nice when done right but this looks like it was thrown together without much thought. Incredibly unattractive style imo.
Surely their powerpoint design team could've done better.

May 9, 2015 | 07:02 AM - Posted by Alex Salmond (not verified)

If the top end cards are water cooled then what happens (rare, I presume) when someone wants to put more than two cards in their machine .. even two might be quite bothersome?

May 9, 2015 | 09:35 AM - Posted by collie

ideally you'd have them on the same loop, just like water cooling sli/crossfire as it is today.

May 9, 2015 | 12:22 PM - Posted by al bundy

Hopefully this marks a change and we move away from "Gamery" looking enthusiast parts from AMD.

May 9, 2015 | 07:52 PM - Posted by Jet Fusion

Do i need a personal Hoover Dam to power it?

May 11, 2015 | 06:14 PM - Posted by Anonymous (not verified)

I hope it does not require water cooling.

May 12, 2015 | 01:23 AM - Posted by collie

why not?

May 12, 2015 | 07:40 AM - Posted by Anonymous (not verified)

Looks about the same size as a lot of the 'short' 970s (which were dropped onto 760 PCBs).

May 22, 2015 | 10:52 AM - Posted by Rickard Eneqvist (not verified)

This is NOT the HBM card, if you follow the chiphell link they show the card with the cooler of and you clearly see traditional memory chip layout.

This is one of the rebranded older card fit into a very small form factor and given its only one 8 pin connector probably not one of the highend cards.

May 22, 2015 | 11:57 AM - Posted by Anonymous (not verified)

I saw that as well but I was wondering, the main complaint people have about the proposed spec is the lack of memory. Is there any technical reason they couldn't use traditional memory as something like a texture cache?

That would be pretty huge. Solves the one issue people seem to have.

May 22, 2015 | 01:27 PM - Posted by Josh Walrath

That would be a much more complex memory controller than what AMD has hinted at with HBM.  It is a good idea, but really tough to implement.

May 22, 2015 | 01:57 PM - Posted by JohnGR

Yes and they rejected it for now.

"- Could we see a mix of HBM and GDDR5, sort of like how a SSD and HDD would work?

- Mixed memory subsystems are to become a reality, but nothing yet, main goal is graphics cards."

May 22, 2015 | 02:57 PM - Posted by Anonymous (not verified)

Having any GDDR5 would defeat a major advantage of having HBM, which is offering high bandwidth at low clock rates, plus having a much simpler PCB design and not having to run the memory traces out of the interposer module. GDDR5 at high clock rates is a power eater. The extra latency from having to go off module for the added memory space is just not worth it, as proper staging and compression of textures from main memory into the HBM will probably be enough until HBM2's introduction. There is no beating 4 channels at 1024 bits wide per channel for HBM, and once the larger HMB stacks start arriving, texture compression will have to make do. With all that bandwidth a lot more extra processing can be done per frame and not have any adverse latency consequences that cause frame rates to suffer, that the older memory technology has to deal with. Expect HBM and the GPUs memory controller to be able to handle many more CPU requests for all processor cores/threads pushing data over more of the 16x PCIe's available bandwidth, as HBM's more lanes at higher available bandwidth should allow for better peak PCIe transfers. I hope that there is some benchmarks that can properly measure the PCIe bandwidth usage for these first HBM enabled SKUs, as I would love to see how much those 16x connections are going to be utilized compared to the non HBM cards.

May 22, 2015 | 03:55 PM - Posted by Anonymous (not verified)

If you are using the extra off package, PCB mounted memory as overflow, then it doesn't actually need to be that fast since you would not use it often. The heirarchy goes from small and fast out to big and slow. If they did run some external memory connections (unlikely, and probably unnecessary for a GPU), then it would probably be on a much narrower, slower, and lower power interface compared to the interface you would see on a non-HBM gpu. If we do see off-package memory with HBM equiped APUs and server chips, I wouldn't expect it to be connected on a 512-bit interface.

May 22, 2015 | 03:41 PM - Posted by Anonymous (not verified)

They will want to implement secondary memory for APUs or server CPUs such that the HBM would act like a cache. I am not sure how they would implement this on a gpu though. Setting it up like a cache may run into some complexities. I suppose they could have it on a completely separate memory controller off the gpu and have the driver manage it rather than a transparent CPU-style cache.

Anyway, capacity issues, if they really are an issue, will disappear with HBM 2. If HBM 2 takes too long, it may be worthwhile to add more HBM stacks to GPUs rather than attempting to add secondary memory. We will see controllers with secondary memory with HBM 2, if it is used for server chips and APUs.

It will be interesting to see how that will be implemented. It may be possible to put the secondary memory controller(s) off the stacked DRAM logic die rather than off the main chip which would have advantages. Given the sizes involved, the caching mechanism may need to look more like paging system memory out to disk rather than a traditional CPU type cache. If that is the case, then you would probably want to implement it as part of the virtual memory system, which would not be simple. You would need to add another level with some pages in HBM, some in slower memory on the PCB, and possibly some out on a storage device. Although given the sizes involved, going all the way out to disk may no longer be necessary if you have huge amounts of RAM. If we have memory modules with stacked memory chips, then the capacity could be huge. You probably could have a tiny but powerful system with one APU with HBM, one so-dimm sized memory module, and one m.2 ssd.

May 22, 2015 | 11:25 AM - Posted by heydan83

Men this is what I´ve been wating for... small, just one PCIe port when water cooled with a water block and powerful, if all of this is true, definitely my next GPU with a freesync monitor.

May 22, 2015 | 12:08 PM - Posted by collie

I wonder if they are gonna do something silly, rebrand all the old cards to the rx-3xx and name the flagship "Hercules" or "Perseus", like "A TITAN KILLER!!!!"

May 22, 2015 | 12:56 PM - Posted by funandjam

Actually that type of marketing is done in all kinds of markets, new features are introduced at the high end, and then trickle down to the lower tiers as time goes forward and newer models are released, and those models that were once highend are rebranded with newer model numbers as lower tier products, happens all the time.
Can't really say anything bad about AMD doing it, especially since this new memory tech is brand new and the manufacturing of it is not mature yet, so yields are lower and cost is higher.

May 22, 2015 | 03:54 PM - Posted by collie

well yea, and using a lower number is fine (still dont like it, understand it) but is seems the r9-270 is going to be renamed r9-370, and it;s still an 7xxx part (7870 i think, dont quote me)
So if it was the new 250 or 260 I would be ok, this is just a need for new/rebrand

May 22, 2015 | 07:14 PM - Posted by Heavy (not verified)

how come no one get this the newer cards are morelikely to have the same chip thats in the r9 285 mean great performance with less power

May 23, 2015 | 01:28 AM - Posted by Anonymous (not verified)

Correct me if I'm wrong, but isn't the r7-230 from the 5xxx gen?

May 25, 2015 | 10:40 PM - Posted by collie

Close. The R5-230 is a 6450

May 22, 2015 | 01:00 PM - Posted by heydan83

Actually I think that would be great, hehe...

May 22, 2015 | 01:30 PM - Posted by Josh Walrath

Zeus? Olympian? A reader mentioned "Uranus"...

May 23, 2015 | 04:37 AM - Posted by collie

We need to get over that old joke and rename that planet "Urectum"

May 22, 2015 | 01:58 PM - Posted by JohnGR

Maybe they could call it "Rambo"

May 22, 2015 | 12:46 PM - Posted by mohihehi (not verified)

I love the chrome finish.

May 22, 2015 | 01:19 PM - Posted by Master Chen (not verified)


May 22, 2015 | 04:06 PM - Posted by Anonymous (not verified)

I really want to see the board itself. It should be super simple with just the interposer package and the power delivery system. The traces should also be super simple: just PCI-e link and display port, both of which are relatively low pin count. You also have the power and ground from the VRMs, but with no memory interface, it should still be a super cheap board. The interposer isn't cheap though.

May 23, 2015 | 04:55 AM - Posted by collie

I have a feeling, I have nothing to base this on, but I have a feeling that the board will be INSAINLY simple. Just a guess tho, but I have a AM1 system and the MB has nothing too it, I have a feeling that this card, IF it's real, IF it looks like this, I think it's gona be a card of connectors and not much else.

May 23, 2015 | 07:56 PM - Posted by Anonymous (not verified)

just release the cards already AMD!!!

May 24, 2015 | 03:19 AM - Posted by anon (not verified)

deadlines only lead to beta's becoming products! idiot

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

This question is for testing whether you are a human visitor and to prevent automated spam submissions.