Friday, August 25th 2017

AMD RX Vega Supply Issues to Persist At Least Until October - Digitimes

DigiTimes is reporting, through "sources from the upstream supply chain", that AMD's current shortage of RX Vega cards to distribute to the retail market will continue at least until October. The tech reporting site says that sources are pointing towards the package integration of HBM2 memory (from SK Hynix or Samsung Electronics) and the Vega GPU (manufactured on Global Foundries' 14 nm FinFet) as being at fault here, due to low yield rates for this packaging effort. However, some other sources point towards the issue being with the packaging process itself, done by Advanced Semiconductor Engineering (ASE) through use of SiP technology. Whichever one of these cases may be, it seems the problem lies with AMD's choice to use HBM2 on their Vega graphics architecture.

As a footnote to its story, DigiTimes is also reporting that according to industry sources, NVIDIA has, in light of RX Vega's performance, decided to postpone the launch of Volta-based GPUs towards the first quarter of 2018.

Sources: Digitimes, via HardOCP
Add your own comment

92 Comments on AMD RX Vega Supply Issues to Persist At Least Until October - Digitimes

#1
the54thvoid
AMD have kind of deceived their way through this whole release. Disappointed in the consumers friend that they once were.
Posted on Reply
#2
INSTG8R
the54thvoid said:
AMD have kind of deceived their way through this whole release. Disappointed in the consumers friend that they once were.
From someone that's been using ATI/AMD exclusively I feel the entire Vega saga was juggling and sleight of hand act from the get go. I'd still buy a 56 if I do upgrade though..
Posted on Reply
#3
the54thvoid
INSTG8R said:
From someone that's been using ATI/AMD exclusively I feel the entire Vega saga was juggling and sleight of hand act from the get go. I'd still buy a 56 if I do upgrade though..
No doubts the 56 is the better option. The 64 seems like a knackered warhorse pushed to its last breath.
Posted on Reply
#4
Kursah
As a friendly reminder as some of the BS has already been cleaned up here.

Everyone has a right to voice their opinions, but not to openly offend, troll or abuse others. If you cannot simply follow the TPU Forum Guidelines and be respectful to others, then please do not bother posting. Infractions will be handed out to those that fail to adhere to this.

Nobody here is entitled to attack anyone else. Period.

:toast:
Posted on Reply
#5
Norton
Moderator & WCG-TPU Captain
The shortage may have something to do with Project 47 also. Wasn't that set up to use 80 Vega cards per unit?
Posted on Reply
#6
TheinsanegamerN
Once again AMD let down by their decision to pursue hyper advanced tech.

They could have had a DDR5 powered polaris chip with 4096 cores last year, with similar perf/watt to VEGA, without wasting all that capital. The worst 480 pulled maybe 180 watt, so double that for a 512 bit 4096 core polaris would still be 50-100 watt less then vega 64 likes to pull.

When will AMD management learn?
Posted on Reply
#7
IceScreamer
Well if you look at it from a different perspective these don't really look like average consumer graphics cards, even in this non Frontier form, so the majority of supply probably went to other, non gaming areas, like @Norton pointed out.
Posted on Reply
#8
Xzibit
Norton said:
The shortage may have something to do with Project 47 also. Wasn't that set up to use 80 Vega cards per unit?
I'm not sure you can blame one because has there been any shortage on the earlier Vega products Pro & Instinct ? Just could be they take priority over RX and demand being what it is in the market not enough chips are being deligated to RX since they are being put together at 2 different places SK and TW, Report didn't touch on that.
Posted on Reply
#9
Assimilator
I know AMD doesn't have the R&D budgets of Intel or NVIDIA, but damn, sometimes it feels like AMD are purposefully making things even harder for themselves. After all, if AMD had expected interposer yield issues this round again, it would've been prudent of them to opt for a memory controller capable of driving GDDR5/X as well as HBM2, similar to Pascal's. That would have allowed them to sell Vega SKUs with the cheaper, more reliable memory. Considering Polaris does alright with GDDR5, I still don't buy the "AMD high-end GPUs will self-immolate if they're not tied to insanely wide memory busses" - I feel like HBM was little more than a halo/product differentiator with Fiji, and that hasn't changed with Vega.

If it really is the interposer that's the problem, that means good Vega chips are being binned because of its failings, which is essentially the same as AMD burning their money. I'm sure the situation will improve by year end, but man, that's the exact opposite thing they (and especially RTG) need right now.

And then there's Volta, which now has an extra half a year to be tested and tweaked to ensure that it pushes out its full performance and impresses everyone maximally on launch day. If Vega had been competitive we would've seen Volta launched with GDDR5/X, but now NVIDIA has time to go back and ensure it works perfectly with GDDR6, and the memory manufacturers have time to get production of the new memory to decent levels. Of course, should GDDR6 flop, NVIDIA already has the finalised GDDR5/X designs for Volta taped out, so they have all their bases covered.

Navi's launch in a year's time is looking both very optimistic and very far away for AMD...
Posted on Reply
#10
diatribe
I hope that this blunder doesn't destroy AMD's future in the GPU market.
Posted on Reply
#11
mcraygsx
For an occasional gamer like myself, I was really excited to switch from 1070 to RX 64 SILVER but prices are outrageous atm. $699 for a regular black version is too much. I could easily buy a 1080 Ti for similar price.

Thanks to AMD, NVidia is in no rush to release VOLTA.
Posted on Reply
#12
Fluffmeister
mcraygsx said:
For an occasional gamer like myself, I was really excited to switch from 1070 to RX 64 SILVER but prices are outrageous atm. $699 for a regular black version is too much. I could easily buy a 1080 Ti for similar price.

Thanks to AMD, NVidia is in no rush to release VOLTA.
Pretty brutal isn't it? If people think they will be able to buy themselves a Vega 56 for a bargain price too then they need to go join Emma Stone and Ryan Gosling in La La Land.
Posted on Reply
#13
Xzibit
Fluffmeister said:
Pretty brutal isn't it? If people think they will be able to buy themselves a Vega 56 for a bargain price too then they need to go join Emma Stone and Ryan Gosling in La La Land.
So with winners? 212 wins and 247 nominations
Posted on Reply
#14
Hood
Every day they release new lies designed to mitigate the fallout from the previous pack of lies. All the lies were designed to call attention away from the main original lie - that Vega's performance was between a 1080 and a 1080 Ti, for $500. All subsequent lies only made it worse, until even die-hard fanboys gave up in disgust. AMD is going from being the much-loved underdog to being the much-hated liar, that caused thousands of people to wait for a pipe dream that never happened. There's a thin line between love and hate, and many former AMD fans are crossing that line now. I'm glad that I prefer Nvidia, I may have to pay high prices, but at least they deliver what they promise, without all this drama and BS.
Posted on Reply
#15
ViperXTR
mcraygsx said:
For an occasional gamer like myself, I was really excited to switch from 1070 to RX 64 SILVER but prices are outrageous atm. $699 for a regular black version is too much. I could easily buy a 1080 Ti for similar price.

Thanks to AMD, NVidia is in no rush to release VOLTA.
Then again, Volta is just right on track like reported long ago, 2017 for their HPC tesla chip and 2018 for general consumer chip
Posted on Reply
#16
Fluffmeister
Xzibit said:
So with winners? 212 wins and 247 nominations
And after all the hype.... it was basically a song and dance about nothing.

I guess people are easily pleased.
Posted on Reply
#17
Fouquin
Hood said:
Every day they release new lies designed to mitigate the fallout from the previous pack of lies. All the lies were designed to call attention away from the main original lie - that Vega's performance was between a 1080 and a 1080 Ti, for $500. All subsequent lies only made it worse, until even die-hard fanboys gave up in disgust. AMD is going from being the much-loved underdog to being the much-hated liar, that caused thousands of people to wait for a pipe dream that never happened. There's a thin line between love and hate, and many former AMD fans are crossing that line now. I'm glad that I prefer Nvidia, I may have to pay high prices, but at least they deliver what they promise, without all this drama and BS.
That's all well and good but I have yet to actually see anything beyond speculation that AMD has in fact lied about anything. Lots of crying and moaning from the *PC Gamers* and very little proof.
Posted on Reply
#18
Xzibit
Fouquin said:
That's all well and good but I have yet to actually see anything beyond speculation that AMD has in fact lied about anything. Lots of crying and moaning from the *PC Gamers* and very little proof.
Pretty much even their own sources are telling speculating on 2 different theories.

DigiTimes
Some sources pointed... and hence resulted in low yield rates

However, some other sources claimed that the issue may be
Usually you want your sources corroborating the story not complicating it by speculating.
Posted on Reply
#19
RejZoR
Interesting enough, RX Vega 64 with reference cooler is still available on many European shops from what I can see. It's just AiO versions that are gone. Price wise, they are roughly 60-70€ more expensive than aftermarket GTX 1080. Which sucks quite a bit considering they are loud and crappy reference cooled compared to beefy AORUS and GamingX GTX cards...
Posted on Reply
#20
Hood
Fouquin said:
That's all well and good but I have yet to actually see anything beyond speculation that AMD has in fact lied about anything. Lots of crying and moaning from the *PC Gamers* and very little proof.
First they said they were delaying launch to make sure there are no issues with supply, and the delay was almost A YEAR. This is after there were several press releases about HBM-caused delays. Then, when supplies ran out on day one, their eventual answer was "we SHIPPED tens of thousands of cards, and 'logistical problems' kept them from reaching distributors (logistical problems involves physically being unable to get to their destination - were the damn trucks hijacked? I don't recall any blizzards, floods, or hurricanes that week). Then all the crap with prices started - bundles, "launch" prices, rebates to retailers for launch only, SEP double-talk vs MSRP. Strange, isn't it, that all these bundles are trying to sell Ryzen CPUs and boards - is there a problem with Ryzen sales? Normally hardware sold in bundles is over-stocked last generation stuff, trying to get rid of them before they're stuck with a loss. So another lie was, "the bundle strategy is designed to thwart miners". Funny how these dozens of bundles all have Vega cards available, but there's no cards available alone. Then, just yesterday, "we've worked out the "logistics issues", and all stores will be fully stocked THIS WEEK". Now today they say supplies won't normalize until October. So which is the real story? It doesn't even matter, because only one story can be the real story, and by definition, the other contradictory stories HAVE to be lies. Take your pick, but I am fairly sure that NONE are the real story, and ALL are lies. It's a high-stakes game they're playing, and people who believe them do so because they desperately refuse to hear anything negative about AMD. I think it will explode in their face very soon.
Posted on Reply
#21
the54thvoid
Hood said:
First they said they were delaying launch to make sure there are no issues with supply, and the delay was almost A YEAR. This is after there were several press releases about HBM-caused delays. Then, when supplies ran out on day one, their eventual answer was "we SHIPPED tens of thousands of cards, and 'logistical problems' kept them from reaching distributors (logistical problems involves physically being unable to get to their destination - were the damn trucks hijacked? I don't recall any blizzards, floods, or hurricanes that week). Then all the crap with prices started - bundles, "launch" prices, rebates to retailers for launch only, SEP double-talk vs MSRP. Strange, isn't it, that all these bundles are trying to sell Ryzen CPUs and boards - is there a problem with Ryzen sales? Normally hardware sold in bundles is over-stocked last generation stuff, trying to get rid of them before they're stuck with a loss. So another lie was, "the bundle strategy is designed to thwart miners". Funny how these dozens of bundles all have Vega cards available, but there's no cards available alone. Then, just yesterday, "we've worked out the "logistics issues", and all stores will be fully stocked THIS WEEK". Now today they say supplies won't normalize until October. So which is the real story? It doesn't even matter, because only one story can be the real story, and by definition, the other contradictory stories HAVE to be lies. Take your pick, but I am fairly sure that NONE are the real story, and ALL are lies. It's a high-stakes game they're playing, and people who believe them do so because they desperately refuse to hear anything negative about AMD. I think it will explode in their face very soon.
It doesn't matter how logically you try to put it, your argument is always ignored and a sly, troll comment is put in to attack you.
Let people believe what they want and see what happens. I'm damn happy I didn't wait.
Posted on Reply
#22
Vayra86
Hood said:
First they said they were delaying launch to make sure there are no issues with supply, and the delay was almost A YEAR. This is after there were several press releases about HBM-caused delays. Then, when supplies ran out on day one, their eventual answer was "we SHIPPED tens of thousands of cards, and 'logistical problems' kept them from reaching distributors (logistical problems involves physically being unable to get to their destination - were the damn trucks hijacked? I don't recall any blizzards, floods, or hurricanes that week). Then all the crap with prices started - bundles, "launch" prices, rebates to retailers for launch only, SEP double-talk vs MSRP. Strange, isn't it, that all these bundles are trying to sell Ryzen CPUs and boards - is there a problem with Ryzen sales? Normally hardware sold in bundles is over-stocked last generation stuff, trying to get rid of them before they're stuck with a loss. So another lie was, "the bundle strategy is designed to thwart miners". Funny how these dozens of bundles all have Vega cards available, but there's no cards available alone. Then, just yesterday, "we've worked out the "logistics issues", and all stores will be fully stocked THIS WEEK". Now today they say supplies won't normalize until October. So which is the real story? It doesn't even matter, because only one story can be the real story, and by definition, the other contradictory stories HAVE to be lies. Take your pick, but I am fairly sure that NONE are the real story, and ALL are lies. It's a high-stakes game they're playing, and people who believe them do so because they desperately refuse to hear anything negative about AMD. I think it will explode in their face very soon.
The real story is this: AMD talked itself into producing Vega with HBM for so long and so often that there was no turning back. They had to release the card, knowing it would never reach the 1080ti anyway, but still would require to do so to extract enough of a margin on it to make it worthwhile. They also HAD to release some sort of high end offering at this point; not being present for even longer in that segment would have marginalized RTG entirely. They desperately tried to stretch performance; evident by all those BIOS versions they employ; if they had a well rounded product they would have given us one optimized BIOS setting that covers all bases.

I don't buy 'low yields' arguments anymore btw, having a yield issue this late in the game just doesn't happen. Yields don't just change for the worse, unless you've got drunk employees handling the silicon. What really happens is that HBM supply is funneled to the high margin markets first, which is super obvious and the story of HBM's short life up till today.

So here we are, AMD employs every stalling trick in the book to keep supply low and keep the product scarce to drive up its price to a manageable level. You can be sure that even Project 47 (Hitman, anyone?) is a dedicated attempt to increase margin on Vega cards.
Posted on Reply
#23
HD64G
Hynix failure to be on time with mass production HBM2 forced AMD to pay much more to Samsung and get half of their needs. That made stocking more difficult and selling for lower price impossible. That's why the packs with monitors, ryzen builds or games. Some of us have written here about this problem months ago but most kept on blaiming AMD only. AMD's own failure was to have good drivers with all features enabled. But we are used to it for 5 years now. Some here also think it is easy to change the memory type of a GPU ready to get to mass production just a few months before that when AMD learnt from Hynix about their fail on production of HBM2. Not possible at all.

As for the reference design's cons, when was the last time that a ref design from any maufacturer was GREAT for high-end GPUs that didn't need the custom ones to get the best out of it for performance-noise-thermals?
Posted on Reply
#24
Assimilator
HD64G said:
Hynix failure to be on time with mass production HBM2 forced AMD to pay much more to Samsung and get half of their needs. That made stocking more difficult and selling for lower price impossible. That's why the packs with monitors, ryzen builds or games. Some of us have written here about this problem months ago but most kept on blaiming AMD only. AMD's own failure was to have good drivers with all features enabled. But we are used to it for 5 years now. Some here also think it is easy to change the memory type of a GPU ready to get to mass production just a few months before that when AMD learnt from Hynix about their fail on production of HBM2. Not possible at all.
As I stated in my previous reply, AMD already had problems with HBM supply with Fiji, so it would have been prudent (I would argue, necessary) for them to anticipate the same with HBM2 and build a GPU with a memory controller that could handle multiple memory types. That AMD chose not to do so is entirely their fault and represents an absolute failure in terms of risk-mitigation.

HD64G said:
As for the reference design's cons, when was the last time that a ref design from any maufacturer was GREAT for high-end GPUs that didn't need the custom ones to get the best out of it for performance-noise-thermals?
There's a difference between "acceptable" and "terrible". Since Hawaii, NVIDIA's reference coolers have always fallen into the first group, while AMD's have always fallen into the second. AMD's failure to improve their reference cooler - over multiple GPU generations - is, again, entirely their own fault.

Fanboys making endless excuses for AMD isn't going to make AMD fix its s**t. Negative publicity and voting with your wallets will. No company should get a free pass on repeated failures, regardless of how small and underdog-y it is.
Posted on Reply
#25
iO
Assimilator said:
... to anticipate the same with HBM2 and build a GPU with a memory controller that could handle multiple memory types. That AMD chose not to do so is entirely their fault and represents an absolute failure in terms of risk-mitigation.
Not even Nvidia with their massively bigger R&D budget managed to do this and chose to build the same GPU twice with two different memory controllers.

And I really doubt adding another ~50W+ for GDDR5 memory and its subsystem plus another potential bottleneck by reducing the memory bandwidth would actually help the card...
Posted on Reply
Add your own comment