NVIDIA Founders Edition Cards - Yea or Nay? @ [H]

schizo said:
Are you suggesting third-parties like Gigabyte, EVGA, and Asus have substantially more quality control issues than reference cards? I hadn't heard that before. Is there any data to support it?
Click to expand...

Well for one AiB customs will get silent revision/component changes during their lifetime. The most noticeable (since it's easy to tell) is that they will switch memory suppliers, this is why you will see review models say use Samsung GDDR5 but might later on buy one and get Hynix or Elpida. Or they might change the shroud fan to a cheaper material that mostly looks the same.

Going by the article then they will not be doing this at all for this line?

As an side this issue also happens with motherboards as well. I wish the industry would be called out much more for this practice.
 
Last edited:
Interesting. I like the fact you can get it early, but doesn't it seem like they just wanted to beat AMD to the punch as in releasing a new card first? It is almost like they made it up right before the announcement....Don't get me wrong this is a paper launch just like the Nano and Fury Duo pro, So the fact that they didn't even explain what a Founder Edition card was a huge mistake.

I like Nvidia reference design as well, but don't you think this might make AIB's a lil mad since Nvidia gets all the money first?

I would still wait for a MSI Lightning if I had a choice. Get 2 8pins on there!
 
My feeling is to get the maximum ROI on a card, I buy one and hold onto it until it doesn't do what I need it to do. I don't need the latests and greatest, I just want it to work. I don't care for the cooler, I'd rather see a longer warranty period.
 
limitedaccess said:
Well for one AiB customs will get silent revision/component changes during their lifetime. The most noticeable (since it's easy to tell) is that they will switch memory suppliers, this is why you will see review models say use Samsung GDDR5 but might later on buy one and get Hynix or Elpida. Or they might change the shroud fan to a cheaper material that mostly looks the same.

Going by the article then they will not be doing this at all for this line?

As an side this issue also happens with motherboards as well. I wish the industry would be called out much more for this practice.
Click to expand...


You also have to remember that AIB cards haven't always compatible with each other in the past. Buy a 1080 card today and decide to purchase a second one in a couple of years. There has always been a chance that the cards won't work at all in SLi due to different bios versions or other differences. Having a standardized line allows for the system builders to sell guaranteed upgrades to end users at a later date. This has to have been ticking off Nvidia for years. There is a reason that the Titan X was never allowed an AIB redesign. How many people complain that those cards were incompatible with each other?

1. The reference line ensures future compatibility for a SLi upgrade next year.

2. The reference line ensures that you aren't mismatching memory from different vendors that could possibly cause small issues later on.

3. Buying Nvidia reference ensures that the components will be within tolerance. Sometimes companies switch out parts to save a buck. If anyone can ensure consistency of the quality of the components in their product line it's going to be Nvidia. They don't have a dog in the race as they really just want to sell chips to AIB partners. Thus their only concern will be making sure that the quality of their reference design is maintained.

4. Buying reference means that your cards will looks exactly the same next year. Always hated trying to get a 2nd card and it was EOL for a new card that could OC 5MHz more. Then my PC ends up looking like it's wearing mismatched shoes. Tacky!

5. If system builders like MainGear stock up on cards, maybe you can purchase one from their stock for SLi if you decide to EOL upgrade to SLi instead of getting the GTX 2160. Err GTX 4K. Who came up with the name GTX 1080? :(

6. I've never had an issue with a reference Nvidia card that I purchased for myself or sold via retail. Likewise I've had zero issues with AIB cards other than NOT BEING ABLE TO PURCHASE A SECOND ONE THAT LOOKS THE SAME AS MY LAUNCH CARD AT A LATER DATE. Whomever changes the look of launch cards for AIB sales needs to be beat with an ugly stick.

7. Reference cards work well in SFF PCs. Who wants to dump heat into the PC?

8. Reference cards work well in cases where the motherboard is rotated 90 degrees like my Silverstone.

9. Reference cards are always used by water block manufacturers for maximum compatibility. No need to beg Koolance or EKWB for a water block to fit the 3rd revision of your aftermarket card that moved one screw 2 mm to fit a triple fan cooler that you're trying to toss in the trash.

10. Let's admit it. That big hunk of metal cooler does look nice and should ensure resale value. People want aesthetically pleasing cards for SLi.

Personally if I were getting a GTX 1080 I would get the Founders Edition for maximum upgrade compatibility. Plus "Nvidia Founders Edition" does have a nice ring to it. The only other card I would consider would be an Asus Strix card as the bad boy version usually has some voltage tweaks.
 
I think everyone is missing what is going on here.. This is not really about Pascal or nVidia, it is due to the limited availability of GDDR5x memory for the next few months. The 1080 will be very hard to get until maybe September. This means that nVidia will be controlling distribution of 1080 cards until then as Founders Edition cards. I'm sure that AIB partners will get enough 1080 GPU chips to do the necessary design testing, but until they can get GDDR5x memory, they will only be selling reference cards at Founders Edition prices. What about the 1070? Much different story, expect to see AIB designs which push GDDR5 speeds to the limit. They will still have less memory bandwidth than the 1080, but they will be much better than anything you can buy today.

What about AMD? Same story, but more honest. AMD will be introducing Polaris cards soon. They are aimed below the new nVidia cards, and to a great extent towards laptops. There may be a Polaris 10 card with GDDR5x at launch but I expect it to show up later. Even later Vega 10 cards will come out with HBM2 memory and aimed way above the 1080, but about the same time nVidia will have Pascal 100 consumer cards. The real news here is that FinFETs, GDDR5x, and HBM2 will result in a significant leap in performance at all price levels, but the memory is not quite here yet. Expect quantity shipments of GDDR5x in the fall, and quantity shipments of HBM2 on about the same time schedule. (But early on. HBM2 will get sucked into high performance computing, GDDR5x won't.)
 
NVIDIA's CEO Jen-Hsun Huang spent a lot of time talking about craftsmanship. He discussed the advancements that had gone into the cooling solution as always, but he spent time talking about power components and circuitry down on the PCB. The GTX 1080 Founders Edition this time also has a vapor chamber cooling system. To my recollection, I have not heard him discuss these types of qualities in the "reference card" in my 18 years of listening to him..... All of these cards built by NVIDIA are put together with exacting standards in terms of PCB quality, power component and other circuitry quality, cooler quality, sound profile, and packaging.

When I read something like this it's basically saying AIBs aren't capable of assembling a proper GPU. Especially if the CEO is talking about exacting standards as a selling point. If certain specs (PCB quality, circuitry, cooling, power components) are to be met then why is there two price structures? Shouldn't all 1080s cards be made to at least the FE specifications and then improved on from there? He just openly talked about needing proper quality of all the components for the base card at a $700 price point. Having never spoken about his before he's basically admitting that any 1080 less than $700 uses inferior components because it isn't made to NVIDIAS exacting standards for a baseline card. As a consumer that would make me question what kind of components AIBs skimped on over the FE before making a purchase. Now I'd want to know which components were different, things they don't normally tell you on the box or spec sheet. Such a terrible idea for NIVIDIA to do this. This would piss me off if I were an AIB partner.
 
Any nVidia fan would gladly pay the extra $100 for SE bragging rights, keeps out the riff raff and will look baller in our hardware sig.
 
schizo said:
Are you suggesting third-parties like Gigabyte, EVGA, and Asus have substantially more quality control issues than reference cards? I hadn't heard that before. Is there any data to support it?
Click to expand...

I'm actually wondering about that myself.

I imagine it depends on the price point and the specific third-party in question. I generally buy cards like MSI Lightning because they have Military Class components (although there wasn't a 670 Lightning card so I had to settle for the Power Edition). I generally go with MSI for my motherboard as well. Very good capacitors and several VRM stages, they go out of their way to talk about the components they use. I pay attention to stuff like this because I worry about bad capacitors, MOSFETs, and voltage spikes harming my card, so I want the best power circuitry I can get. That extra stability and circuitry can be important if you don't want component failure over time. All I know is that way back in the past, when I made my first video card purchase, I got the cheapest PNY Verto I could get, and the thing started crashing in games in less than two years. But I've had this MSI Power Edition for over three years now, and never had an issue with it. I also haven't had issues with any other higher-end MSI cards I've purchased.

However, I'm not so sure whether cards priced at or below MSRP are any good... I don't buy those anymore after getting burned by PNY (literally, the fan on the thing melted). I'm actually worried about cards that only cost MSRP at launch (or that come from no-name brands) being dangerous or failing early, and I'm too impatient to deal with RMAs. I'd rather get something that's not likely to fail in the first place.

I mean, I really wanted to wait for the MSI Lightning edition of the 1080 because that's what I've come to trust, but the fact that there's a three-year warranty on the card is definitely making me consider it. Especially given that Kyle is saying that these designs are at least as reliable as those made by other OEMs (if not moreso). Not to mention that current information suggests these cards are likely to be so scarce that the FE will be all you can get for several months.

And I'm pretty tired of waiting out Maxwell to the point that I might buy one of these cards. I remember supply was so bad on the 680s at launch that people were paying far more than a $100 markup on eBay and constantly refreshing retailer pages hoping to catch the item in stock. That's actually why I have a 670... I snagged a 680 at retail by luck for a reasonable price, but then saw what they were selling for, and saw that the 670 was really close at higher clocks... and sold it NIB for around $200 over what I paid without opening it. It was so underpriced relative to what people were paying for it that I couldn't bear to keep it. Would have been like throwing free money away. I have to wonder if NVidia felt the same way and decided they wanted to be the ones to profit.
 
Astral Abyss said:
What happens if we want to go SLI and buy 2 Founder's Edition 1080s? Do we have to buy the custom Nvidia SLI bridge as well? Are AIBs going to include this new bridge?
Click to expand...

I was wondering about this too.

I'm guessing you'll have to buy it separately since it's a rigid bridge available in three different sizes.
 
^
If that Kyle quote from two posts above is accurate, then I strongly agree. If you don't care about higher quality components, then perhaps skipping the SE is right for you.
 
Singularity_Survivor said:
^
If that Kyle quote from two posts above is accurate, then I strongly agree. If you don't care about higher quality components, then perhaps skipping the SE is right for you.
Click to expand...
I absolutely love what ASUS, GBT, and MSI do with "extreme' edition video cards. These guys validate higher GHz binned parts with custom designs and we all know there is a huge market for those cards. Many of us are going to prefer those faster clocked cards. But the fact of the matter is that you have never heard one complaint about coil whine or premature failure from an NVIDIA built video card ever. On the other hand it is totally up to Asus, GBT, and MSI to put together video card builds that impressed the enthusiast community as those cards get no exposure other than online.
 
rinaldo00 said:
Isn't that because the AIBs push things to the limit while the reference card plays is fairly safe? ie max factory OCs on the GPU and VRAM
Click to expand...

Nope, or anyone OCing a ref card would run into these issues too...

It's just cost cutting manufacturers using cheap, crappy, barely spec chokes/coils plus lack of 'chinesium goop' applied, which damps or reduces resonance and vibration that we hear as coil whine. It's usually due to magnetostriction, which is where the field is densest around the core and thus creates the most stress on the coils and core. This is why my custom 1200VA hifi toroids have 1500VA rated cores.. Krell style.


From the [H] article and others, all I can see is that the ref card is $100 premium because nvidia commissioned it and charged that much. AIBs first round later on, will likely be similar spec/reference (and $100 cheaper), but who knows what components will be used and if you'll get coil whine/etc. That's something we'll only know when we have them in our hands and can check the components out. I bet they will try to cut corners though. Kyle makes an excellent point about reference cards not suffering coil whine.


Here's the thing - can we stop calling it a flounder edition, and just call it a damn reference card aka 'BBN'. Some used to call reference cards 'BBA' (Built By ATi) in far distant past.. this is the same but Nvidia. It's just confusing people and providing free marketing for them. So, suck it, nvidia marketing - it's an Nvidia built,1080 REFERENCE CARD. Reference is of course the terminology commonly used in this industry and I encourage everyone to continue to call a spade a spade.

Or what's next, AMD makes the 'Launch Edition' and we all start using bullshit marketing terms for reference cards, further confusing people who might be getting into PCs (there's quite a few these days, after the 'lastnext gen' con-sole debatable..).

Basically, it's a reference card, built to Nvidia spec under their control, which will be available until the SKU is retired at a $100 MSRP premium, aimed at system integrators and mATX/modders/H2o types. End of story.
 
I have been reading here for over 10 years and I am just shocked that for the first time in that span people are raving over a reference card. I think that means Nvidia marketing worked.
 
rinaldo00 said:
I have been reading here for over 10 years and I am just shocked that for the first time in that span people are raving over a reference card. I think that means Nvidia marketing worked.
Click to expand...
Yes, but did you see the strikingly futuristic design of the blower housing? Holy shit, it's like the Alienware of reference cards. Clearly it's worth the premium.

Did someone say GPU Of The Year Edition?
 
  • Like
Reactions: N4CR
like this
Is Nvidia proving a point to AIB ? People are willing to pay more , so let them ?
 
limitedaccess said:
Well for one AiB customs will get silent revision/component changes during their lifetime. The most noticeable (since it's easy to tell) is that they will switch memory suppliers, this is why you will see review models say use Samsung GDDR5 but might later on buy one and get Hynix or Elpida. Or they might change the shroud fan to a cheaper material that mostly looks the same.

Going by the article then they will not be doing this at all for this line?

As an side this issue also happens with motherboards as well. I wish the industry would be called out much more for this practice.
Click to expand...

I agree. It is not just the visible components either (like memory). A quick example is: a few years ago after the launch of the HD7000 series, AMD disclosed to the public that its AiB custom manufacturers were causing some serious instability in their GPUs because the ceramic surface mount capacitors were below spec. You will also frequently find digital VRM designs on reference cards replaced with cheaper analog VRM designs. I also suspect that reference cards go through more rigorous validation, as situations resembling MSI's insertion of a capacitor in series on the Vref leg of a voltage regulator to boost GPU voltages beyond what Nvidia specified, would not happen on a reference card.

Also if you look at the aftermarket cooler designs, they are generation to generation, iterated designs that is much simpler to machine than reference designs (a good example is EVGA's sawed off heatpipes, which were cut because they did not fit the new card, having been previously designed for an older card). Reference designed squirrel-cage coolers have an integrated aluminum plate that makes contact with the VRM components, RAM chips and such. That makes the machining of these plates very costly compared to stamped, cut and soldered/glued together fins on a base plate. In addition, the fans used on a lot of the aftermarket designs are not very high quality and suffer from early failures. Some of these aftermarket coolers do not even feature fans with PWM fan control, and some do not even feature RPM sensing fans.
 
The moment you know you've had one sip of Kool-Aid too many.
 
Last edited:
So basically an AIB can use the same reference design and components as the founders edition card but use their own cooling solution and sell it for $100 less. I see that as a plus in my book! Early adopters always pay more for a new products. It's Economics 101: Supply vs Demand. I don't see how this is a big deal really. If the extra $100 bothers you then just wait a bit.
 
I can't make up my mind on whether to grab one or not. My EVGA Step-Up ends 6/11, so it may be a non issue since they told me 1080s may not be available for step-up even though they are for sale. I don't know if its worth spending the ~$100 to upgrade or pick up a 980ti when they drop in the $350 range. Im leaning towards the upgrade but i'm worried i'm getting sucked in to the hype/ooo its new and shiny.
 
Just slap a vapor chamber on GTX980 reference blower for $600 and call it a day. You don't need tooled triangles.

It just aggravating Nvidia pulls this stunt (out of no where) to raise prices as a reference addition is now a premium product.
 
...or maybe it's a bit of a win-win "gamble" by nV? Charge a bit more now for what appears to be higher quality because they're not sure if Radeon will come out with a similar card for cheaper (but a bit later on?). If they don't, then they're just fine as they have a premium product. If not, they made some extra cash before they drop the price down to what they originally intended? It'll only irk those that gotta have what appears to be the best right now - and likely half of them won't really care anyways because they wouldn't ever buy a Radeon card anyways? :) Regardless - it'll keep folks like Kelt Reeves happy with a great standard they can always count on.

Good thing it looks like we'll really know in the very near future. Want some good competition and a great card!
 
cageymaru: Those are all totally valid reasons to buy a reference card that I don't care about. If there was some sort of pervasive problem with quality control, that would be one thing. Changing memory suppliers, changing looks... I don't care. As an end-user I certainly don't care that Maingear or Falcon NW prefers reference cards for consistency.

I'm a consumer. I plan to buy one card. I want it to offer the best performance, and the least noise, at the value sweet spot. I don't see any scenario where the reference card will be that.

(But that's just my opinion, until the reviews are out.)
 
My take on it: The Founders Edition of cards are built and sold by Nvidia, who also sells raw GPUs to 3rd party vendors like EVGA to create their own board designs and coolers. In order to not compete with Nvidia's own vendors in the desktop video card market, the Nvidia-created Founders Edition (unique name (not just a number) is used to distinguish it from other OEM designs) is priced higher. This is a normal non-compete practice done by manufacturers to remove competition with resellers/vendors. This allows 3rd party vendors to create their own "reference" design boards for a cheaper price and not worry about direct competition from Nvidia.

For the GTX 690, which was only built by Nvidia and not allowed to be copied or modified by any 3rd party vendor like EVGA, only reference cards were sold no matter which vendor you bought it from. It was only sold for a limited time period and the price was same no matter where you bought it (not including fake markup done by retailers due to supply/demand issues - the MSRP was always $999). I bought mine from EVGA for $999 and am still using it today.

The Founders Edition appears to be the next step beyond that - a direct manufacturer's card that will be available for the entire production lifespan of the GPU but at a higher price. It satisfies all of the checkboxes from vendors and resellers and integrators: Non-compete price, stable supply of cards, static construction design for the lifetime of the card and a warranty. No bells and whistles, just reference with stability and longevity that can be leveraged and relied upon.
 
crusty_juggler said:
Yes, but did you see the strikingly futuristic design of the blower housing? Holy shit, it's like the Alienware of reference cards. Clearly it's worth the premium.

Did someone say GPU Of The Year Edition?
Click to expand...

I like that the reference design (Founder's Edition) is understated elegance - basically a northward revision of the older reference designs. Yes - there ARE hints of the original high-end (original TITAN) reference design; in fact, it's there on purpose. Still, it's NOT identical. It's a bit edgier without being gaudy or - worse yet - tasteless. It can actually ADD elegance to an otherwise-plain interior. (Note that I said "elegance" - not "bling".)

Also, this will make up for a shortfall of GTX970 reference availability - GTX1070 FE will, in fact, be available from day one (yes; there will be a premium); however, was that not the case with GTX970 reference? Lastly, what was the premium being asked for those GTX970 reference-cooler versions compared to standard-AIB? Compare that to the difference between AIB SRP and GTX1070 FE - are you, in fact, getting screwed?
 
No, reference cards have never been sold at a premium before. Kyle said that NV basically broke even on them, selling at base MSRP.

So far there's no reason to think the extra $100 buys you anything other than the card a month early. The only reason to get a founders edition is if you can't wait a month or you're running in a system with very poor ventilation and need a blower. Maybe if you plan to add an aftermarket water block, but the first round of AIB models are usually reference cards with custom coolers, so really not even then.
 
BreezeDM said:
Why is Nvidia involved in selling directly to consumers? That seems like more of a PITA than it's worth for them. I knew they support and sell Professional cards directly, but those are marked up. Even with the marked up founder edition, I wouldn't bother if I was Nvidia.
Click to expand...

Every GPU manufacturer has had a direct-sales channel in addition to AIB partners except for the very floor (including ATI Technologies back prior to the acquisition by AMD - in fact, ATI took it further, and had a retail storefront in their HQ building in Canada). Even firearms manufacturers do the same thing.

Further, you yourself just stated why - specialty product. It doesn't JUST apply to Quadro; it applies just as much to TITAN and successors. You CAN still buy them through resellers - how many Quadros are sold direct as opposed to either Newegg OR Amazon - and that is just to Canada?

Lastly, some customers want the extra hand-holding for reasons of comfort - and are, in fact, willing to pay extra for it. (That's not unique to nVidia, either - in fact, it didn't even START there. Every above-middle-grade automotive manufacturer based in Europe has a European Delivery Program - why does it exist? BMW and Mercedes have one for EVERY model - even the cheap-seats models like the CL-A and BMW 1 series. Audi expanded their own ERP to every model with the launch of the A3. In fact, BMW now have a USDP for their US-manufactured models; it's there because it value-added, and folks WILL pay extra for it. Get your profit margin where you can, and HOW you can (within legality).
 
NVIDIA to OEMs: Yes we are selling this board to you at cost. However the chip on it we are selling at a $200 profit.
 
jfreund said:
This whole thing just looks like another example of Nvidia marketing gone retarded. The 1080 and 1070 look like excellent cards, and AMD has nothing to go against them. Why doesn't Nvidia focus on that instead of introducing this "Founders Edition" nonsense?
Click to expand...
I highly suggest you vote with your wallet!
 
maverikv said:
I wonder if founder's ed. will be the only way to get a blower this gen
Click to expand...
Good question, I do not have any idea on that yet as I have not seen any AIB cards to date.

schizo said:
I think most of us plan to-- the price isn't what's annoying, it's that we need to wait that extra month. Oh well.
Click to expand...
I hear you. I do not have a problem with Founders Edition, but like I said in the article, the execution of this new plan has been badly executed at best. That last thing NVIDIA needs to do is alienate its loyal high end customer base.
 
Throwing my $.02 in here....

It concerns me that NVIDIA has always been marketing about having great performance, but there is a shift in marketing to "craftsmanship." Last I knew, there hasn't been a wide-spread problem with the quality of NVIDIA cards, or their partner cards, so where does this craftsmanship emphasis come from? Could it be that NVIDIA is anticipating a less-than stellar performance increase, or at least much less than anticipated, and is trying trump up value and sales early on?
 
I think the whole thing is a marketing scam.
As you get smaller and smaller in the lithography, you would expect a smaller tolerance for imperfections in the die. That's why yield rates are lower as you go smaller.
I expect the special binned engineering sample gpus to be almost equivalent to the non special gpus.
They just wanted to do a cash grab to enthusiasts who will spend a decent amount on a 1080 just to release the next titan a few months later when hbm2 is released and the cards are upgraded again.
Until i see some real specific differences, that's all speculation.
 
I purchased a reference design 780Ti (when they were new) because I had an Alienware PC with terrible air flow. I was so impressed at how cool and quiet it was (especially in such an awful case) I bought 2 reference 980Tis when I built my 6700k rig. The MSRP means nothing to me as new cards in Canada always sell well above MSRP a few days after launch until they put them on clearance when new models are announced again. This happened even when our dollar was at par with the USD. We hope for big sales just to pay close to MSRP lol.
 
Dawill said:
There are no alternatives. I don't think many of us even look at AMD cards these days.
Click to expand...
People always say this, but the truth is that most consumers have zero loyalty to red or green. Back when ATI was on top, I had a 5850. Loved it. Amazing GPU for the time. As soon as NV climbed the mountaintop, I came back. I will buy whatever is the best price/performance at the time, and most consumers are the same way. AMD simply hasn't released particularly competitive hardware for the past couple years.

I do think AMD will do extremely well with Polaris 10, but it isn't very interesting to most of us on the [H] forums as it's a mainstream part.
 
My one real concern with these new cards is power throttling, as the 970 has been notoriously known for being power throttled.

So the 1080 is only going to use a single 8-pin power connector and TDP is 180W. I fear that lower power draw could potentially be a bottleneck. I understand it is 16nm FinFET and it is supposed to naturally have 2x better perf/watt, but I still don't know. A few of us might not care about our electricity bill (lol) and want to bump that 2.1 GHz up to 2.4 GHz but can't because the power delivery is already maxed out. It doesn't matter how good or bad your cooling is if you can't deliver enough juice for the OC. The interesting thing was seeing the difference in VRM design, number of VRM chips, using high end copper heat sinks on the VRMs, etc. The MSI cards probably did the best job on power delivery IMO vs some other vendors.

Oh well time will tell.
 
After all this discussion, I'm feeling pretty good about getting a "Founders Edition" sometime in the future. I am looking to buy a monster GPU and drive it into the ground (5+ years use), and an Nvidia-made board seems like a pretty solid idea. Initially I was hoping for an EVGA hybrid-watercooled type board, but this was more for warranty and temps/power consumption than overclocking (when you buy a Bugatti Veyron, why tune it? DRIVE IT!).

As a stockholder of Nvidia and longtime fanboy (although last card was ATI Radeon 5850) I'd love to support them!

I have no idea when/if a TI version might drop, hopefully 6-12 months after this release; I'll have to look into the EVGA step-up program, not sure how that might work if i became impatient and bought a 1080...
 
You must log in or register to reply here.
Back
Top