Retail RX 480 4GB to 8GB Memory Unlock Mod Works

H

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
Over the long holiday weekend some Radeon RX 480 4GB owners were claiming that their cards actually shipped with 8GB of memory but they were unable to access all the memory due to BIOS limitations. Well, the crew at techPowerUP decided to flash the BIOS on a retail RX 480 4GB (that had 8GB of physical memory on the card) and it worked. Hit the link for all the information and benchmarks.

Earlier this week, we heard reports of some early adopters of the 4 GB variant of AMD Radeon RX 480 claiming that their cards shipped with 8 GB of memory physically present on their cards, but their graphics card BIOS somehow prevented the GPU from addressing more than 4 GB of it. In its Reddit AMA, the company presented a vague answer to the question of whether such 4 GB cards are moddable to 8 GB by flashing it with the BIOS of the 8 GB variant, by stating that the ability to mod is restricted to review samples. This is both true and false. Short answer: retail 4 GB RX 480 can be flashed to 8 GB, and the modified card performs on par with the 8 GB variant.

Adding more fuel to this whole "8GB in 4GB sheep's clothing" story are these pictures pulled off Imgur.
 
There was earlier reports that reviewers were sent an 8GB card with 2 bios files.

The sticker on the RX480 4GB box could just be that it was cheaper/easier to use stickers than have a second run of boxes printed.

Good luck to those who buy a 4GB and flash it.
 
Is there a way to tell how many chips the board has without removing the cooler and voiding the warranty?
 
It has been a while since there was so much drama over graphics cards. I think Fermi was the last one that I recall being super divisive. Still, this has happened before but usually is related to cards with faulty memory, or memory that just didn't perform up to par. I wonder if that's the case here or if they just ran way too much production of 8GB units?
 
I'm very confused....

I'd say this was a packing error, hence the stickers on the boxes, but the BIOS is even limited to 4GB. Why the 'free' memory on those boards? There's gotta be a reason why.....a high chance of possibly faulty memory?

I'd be curious to see how those that flashed the BIOS for the extra 4GB fare in the long run.
 
funks said:
Is there a way to tell how many chips the board has without removing the cooler and voiding the warranty?
Click to expand...

Its not the number of chips (which is the same in all cases) but the part number used. I think you have to pull the cooler.
 
AMD always has a surprise when new cards come out from them. Just like I was able to flash my 290's to 290x's with ease when they first came out.
 
Enabling another 4 gigs, that is average 3.4w per chip so 3.4x4 = 13.6w according to tomshardware, and you just took your card put of PCI-E specs. Bahahahahaha. I think the 4gb cards were probably close to 150 unless all 8 chips were being powered but not being seen. Hmmmm
 
NKD said:
Enabling another 4 gigs, that is average 3.4w per chip so 3.4x4 = 13.6w according to tomshardware, and you just took your card put of PCI-E specs. Bahahahahaha. I think the 4gb cards were probably close to 150 unless all 8 chips were being powered but not being seen. Hmmmm
Click to expand...

I've been thinking along those lines as well (and also wondering if the 470 was the original target of this Polaris chip).
 
Reminds me of the business model for computer printers. Pay more for a firmware that enables faster printing speed!
 
Sometimes it easier to manufacture 1 item and lock it down for cheaper clients. Especially if it is new and not enough operating capital to cover both methods of manufacture until some original stock gets moved and money is flowing.
 
So with AMD 4=8 , with NV 4=3.5 ?

Seriously I'm actually just having a laugh. I think its just early ref boards that were sent to reviewers, maybe same batches that can be bios controlled to either version. Most likely time constrained and only 1 pcb ref version made it out of the gate in time.

I doubt faulty memory , time will tell if anyone reports that.

And power draw arguments are fails at best relevant for only the crappiest of motherboards if that, that issue will be fixed shortly in drivers if it was an issue. Seems its not an issue people are worried about as they are selling out everywhere by all reports.
 
Thats insane, I might grab a 4GB version in this case... Well ok no.
 
Free 4gb on vram! Awesome! Wonder how many cards are 8 gigs out there masquerading as 4 gig.
 
Sonicks said:
I'm very confused....

I'd say this was a packing error, hence the stickers on the boxes, but the BIOS is even limited to 4GB. Why the 'free' memory on those boards? There's gotta be a reason why.....a high chance of possibly faulty memory?

I'd be curious to see how those that flashed the BIOS for the extra 4GB fare in the long run.
Click to expand...

More likely it was simply cheaper to pay for the extra 4gb of ram on the lower end version then design,print and assemble a second line for the first run. Memory is a lot cheaper then the hardware guys would like folks to believe.

I suspect when the manufacturer gets to a second production run they will setup a second line and put out cards that really do only have 4gb.

The only issue I can see with this... is perhaps a handful of people expecting to get a 8gb card and being disappointed if/when the manufacturer starts shipping actual 4gb cards.
 
$40 peel off rebate for an upgraded card.

AMD hasn't officially released anything, but there are some modders that already adjusted it towards the 6 pin so no reason to expect it won't be adjusted in the near future.
 
Sonicks said:
I'm very confused....

I'd say this was a packing error, hence the stickers on the boxes, but the BIOS is even limited to 4GB. Why the 'free' memory on those boards? There's gotta be a reason why.....a high chance of possibly faulty memory?

I'd be curious to see how those that flashed the BIOS for the extra 4GB fare in the long run.
Click to expand...
Volume it likely is a cost of production move set up 1 line to produce cards and just bios lock out cards to fill the 2nd tier at least for early production. Likely in a month or so they'll have a proper line to assume 4gb models till then it's just about filling street date volume. This isn't the first time early production models from AMD were just bios locked or something similar of a higher tier card. Esp if it's just memory, gddr memory isn't exactly breaking the bank on the profit margins of the card.
 
Memory costs less than $5 per gigabyte. Getting people to upgrade from 4GB at $200 to 8GB at $240 nets $20 of extra profit (or at least $20 less loss - this is AMD we are talking about here) per card.

So it's not beyond imagining that the 4GB cards were never intended to sell in any volume, and instead just exist to bait buyers into looking at RX 480 cards. When they see that the 4GB cards are not available, but the 8GB cards are, they'll hopefully say "well, it's only $40 more for twice the memory ..."

Of course, they have to have some 4GB cards out there to avoid being accused of fraud, but maybe not enough to justify setting up a separate production run, even of box art. No problem: AMD can just limit a (very) few of the 8GB cards to 4GB using the BIOS, slap a sticker on the 8GB box to read 4GB, and they are done. And that is what they did.

TL;DR: the 4GB RX 480 cards may be a near-phantom version meant only to generate consumer interest in the 8GB cards.
 
I don't understand why people complain even when they get more than they paid for. I would be extremely happy if I got 8 for the price of 4.
 
I don't think anyone would be complaining if they bought the 4GB card and discovered they could mod it to have the full 8GB.

I was speculating on the possible dissatisfaction of the ones that bought the more expensive 8GB version, only to discover that they could have just bought the cheaper 4GB version and modded it.

IE, $30 for the BIOS.
 
babochee said:
what a terrible business blunder... pure fail... who's getting fired?
Click to expand...


Yeah, horrible business blunder!!!

It's not like nVidia didn't fuck over 970 owners by selling 3.5GB cards in 4 GB clothing or anything.

Oh wait they did! Nvidia explains GeForce GTX 970's memory performance issues, admits error in specs

Correlate this to taking a girl home from the bar:

AMD makes you feel like you took home a 10 to bed when you wake up the next morning.
Nvidia makes you feel like you woke up with a tranny in bed.

Yeah that's bad business.
 
griff30 said:
Yeah, horrible business blunder!!!

It's not like nVidia didn't fuck over 970 owners by selling 3.5GB cards in 4 GB clothing or anything.

Oh wait they did! Nvidia explains GeForce GTX 970's memory performance issues, admits error in specs

Correlate this to taking a girl home from the bar:

AMD makes you feel like you took home a 10 to bed when you wake up the next morning.
Nvidia makes you feel like you woke up with a tranny in bed.

Yeah that's bad business.
Click to expand...

nVidia's 970gate was really a negative marketing blunder (IE customers bought 'less' than advertised, for now let us ignore the whole 3.5+0.5 vs 4GB argument).

This is a different kind of blunder (customers bought 'more' than advertise), but I really don't see myself believing that this was an unintentional blunder, I think AMD would have to have known that the two cards were physically identical and the difference only exist in the BIOS.
 
quiktake2009 said:
Reminds me of the business model for computer printers. Pay more for a firmware that enables faster printing speed!
Click to expand...

Ha. Way back in the day, HP was trying to charge like $15 or $20 for a driver update that was MMX enabled. It made their super slow print rendering about 4x faster.

Now that is what you do when you are trying to rip off people.

It could be that it was cheaper/easier to just make a massive amount of 8GB cards and put 4GB BIOS on them instead of having two different lines with different components on them.
 
chenw said:
nVidia's 970gate was really a negative marketing blunder (IE customers bought 'less' than advertised, for now let us ignore the whole 3.5+0.5 vs 4GB argument).

This is a different kind of blunder (customers bought 'more' than advertise), but I really don't see myself believing that this was an unintentional blunder, I think AMD would have to have known that the two cards were physically identical and the difference only exist in the BIOS.
Click to expand...

Sorry I couldnt find the sarcasm button, you'd think you could have seen that in my post.
 
You must log in or register to reply here.
Back
Top