TR Forums

K-L-Waster wrote:

DPete27 wrote:

K-L-Waster wrote:

Holy crap!

Now, to be clear, we haven't seen anything official from AMD on the pricing model mentioned in that article... but if it turns out that's legit, that sounds flat out illegal. I don't think anywhere allows bait-and-switch pricing.

How is that bait and switch? They DID sell cards at $499....even if it was only 275 of them. Giveaway games like that are never "free"

Telling the world and dog that the price is $499 with no mention of it being a limited introductory price leads to the expectation that that will be the price (oddly enough). It also leads review sites to post comparisons with competing products that are misleading.

For example, Jeff's scatter plot will tell a completely different story if the 64 is actually priced at $599. Instead of being performance and price competitive with the 1080, it's performance competitive with the 1080 but price competitive with the 1080 TI.

EDIT: or maybe it's all a misunderstanding -- in which case, no harm no foul.

http://techreport.com/news/32417/amd-we ... ontroversy

Nah, This is just consumer entitlement, plain and simple. As I posted in the comments on the related TR news piece:

Just because the way demand has worked in the past in these markets - where the value and demand for an item gradually diminishes after launch, in line with a static, and more often declining, price model with time, it doesn't mean that in a situation where high demand remains constant/increases and/or supply is constrained, prices can't increase.

AMD made a statement on what launch prices were and stuck to it. Why shouldn't they raise them if supply/demand warrants it? They are a business, not a government-subsidized commodity supplier. People don't seem to be as up in arms when DRAM manufacturers raise prices in the face of demand.

The reason for "misleading" price comparisons in reviews, again, isn't the fault of AMD, nor even of the reviewers. It's worked well enough in the past. Traditionally launch prices remain stable for a couple of months or so and then (if no other sudden disruptive competing product enters the market) prices generally gradually decline, reaching some plateau. But sure, when the supply/demand trajectory is atypical, these comparisons are suddenly less useful.

CuttinHobo wrote:

Considering the short windows where a GPU is the best way to chew through cryptocurrency - before the specialized ASICs come out - I it might be AMD's best bet to cater to that fleeting market immediately. It seems like most miners will gladly pay launch-day prices, whether they end up making an actual profit or not. If they had held the mining optimizations back until gamers get their fill, it may have been too late.

I'd go a step further than this.

AMD should have marketed ONLY mining and compute cards (at pre-inflated prices), and kept gaming-Vega as a replacement for Polaris about a year from now (for about the same price as Polaris when it was introduced).

I mean, it's not like their architecture is going to improve that much for next year, right? (..and the cut-down Volta version of the 1060 isn't going to be a "lightweight".)

cynan wrote:

AMD made a statement on what launch prices were and stuck to it. Why shouldn't they raise them if supply/demand warrants it? They are a business, not a government-subsidized commodity supplier. People don't seem to be as up in arms when DRAM manufacturers raise prices in the face of demand.

The reason for "misleading" price comparisons in reviews, again, isn't the fault of AMD, nor even of the reviewers. It's worked well enough in the past. Traditionally launch prices remain stable for a couple of months or so and then (if no other sudden disruptive competing product enters the market) prices generally gradually decline, reaching some plateau. But sure, when the supply/demand trajectory is atypical, these comparisons are suddenly less useful.

Sorry, but this is BS: AMD advertised these at a lower price than they really are. They supported that low price for a tiny little short time, by rebating retailers directly, because the advertised price is lower than what retailers have to pay the distributors. It's fake pricing.

And it changes the conclusion of the reviews. So reviewers are complicit, sorry Jeff, if they publish price-performance charts that reflect fake pricing.

Recreated TR chart:


Edited for reality, cards available IRL, Vega 56 @ same scaling versus MSPR from Vega 64:

See how this changes everything?

Now, I like AMD. Check my post history. I have an old R9 290X and a good Adaptive-Sync monitor. I'm not a NVIDIA or Intel fanboi looking to bash AMD. But this is deceptive business, and reviewer sites who wish to be taken seriously should call it out. It is not supply/demand-driven, the price is not being raised due to shortages, that's simply untrue.

The real price is distrie price plus reasonable margin, not the price factoring in the mayfly-like hidden rebate. At the real price, these cards are horribly uncompetitive IRL. (Whether reviews reflect RL or corporate spin is up to them.)

Maybe I can grant that the launch price announcement wasn't the most forthcoming. But again, it's hard to blame AMD all that much.

When your major/only competitor in this market is a 500 pound gorilla making money hand over fist on their gaming GPU markups (as Pascal is cheaper to make than Vega, the 1080/70 considerably so, and Vega being MIA for a year), AMD needs all the help they can get to stay viable in this market. That's to say, "the real price is cost plus a considerable margin" has no anchoring in reality when, at $400 for a GTX 1070/Vega 56, Nvidia is probably still making a healthy markup and AMD is probably barely braking even. And this after all the GTX 1070s sold without any competition.. Add to this the volatility of the mining demand, and sure, AMD, due to its reasonable relative desperation to stay profitable compared to Nvidia, can't really help but look to be playing it fast and lose with any pricing guidance they proffer. (Which is why I conceded that perhaps AMD maybe just shouldn't have made any statement regarding pricing at all - but again, this is at odds with marketing convention...)

I don't know why people are upset about the price/performance graph. Sure, it's not accurate. But with the limited Vega supply and mining demand, any prices AMD announced prior to launch were likely never going to be accurate for any length of time. This doesn't stop people form looking at Jeff's pricing plots and then making whatever mental adjustment is appropriate given current pricing.

techguy wrote:

Waco wrote:

Pville_Piper wrote:

So I guess all the good ones went to the reviewers?

I'm going to go with buyer's remorse and a large helping of hyperbole, myself.

Hawkwind is primarily active at Rage3d forums. You can read his "journey" with Vega in this thread: http://www.rage3d.com/board/showthread.php?t=34045172

His card is faulty. Not really buyer's remorse. His comment about a 750W PSU being inadequate are, unfortunately, a misdiagnosis from the mob.

Ouch. The comments there are utterly unhelpful in every possible way.

cynan wrote:

I don't know why people are upset about the price/performance graph. Sure, it's not accurate. But with the limited Vega supply and mining demand, any prices AMD announced prior to launch were likely never going to be accurate for any length of time. This doesn't stop people form looking at Jeff's pricing plots and then making whatever mental adjustment is appropriate given current pricing.

I'm not upset by it - I'm not invested in this at all.

HOWEVER,

The reason is that it's NOT ACCURATE.

Moreover, the competition's cards in the graph ALREADY REFLECT CURRENT OVER-PRICING. Both it and the review"s conclusions are completely wrong, and it's not just TR's but EVERY review source that's gone to the trouble of reviewing the card with the **** pricing.

Should the reader be making a mental adjustment for something that was pure fiction to begin with?

Should the author be content with fostering pure fiction?



..and really, it's not like AMD hasn't pulled **** similar to this before. (..and I find it rather unbelievable that readers and authors haven't mentioned this.)

Polaris launched with inadequate supply. Particularly touting their 4 GB version of the 480 for $199 which had the right price/performance at the time, even though almost no one could find that card. Anyone remember that?

cynan wrote:

Maybe I can grant that the launch price announcement wasn't the most forthcoming. But again, it's hard to blame AMD all that much.

Why is it hard to blame AMD? The launch price announcement was forthcoming, it was more than that: it was explicit.

Explicit and clear about a price which is not, and never was going to be, what the card would sell for after the first few minutes.

cynan wrote:

When your major/only competitor in this market is a 500 pound gorilla making money hand over fist on their gaming GPU markups (as Pascal is cheaper to make than Vega, the 1080/70 considerably so, and Vega being MIA for a year), AMD needs all the help they can get to stay viable in this market.

Yes. But it shouldn't lie, or trick reviewers into publishing price-performance plots which are lies.

cynan wrote:

That's to say, "the real price is cost plus a considerable margin" has no anchoring in reality when, at $400 for a GTX 1070/Vega 56, Nvidia is probably still making a healthy markup and AMD is probably barely braking even. And this after all the GTX 1070s sold without any competition.. Add to this the volatility of the mining demand, and sure, AMD, due to its reasonable relative desperation to stay profitable compared to Nvidia, can't really help but look to be playing it fast and lose with any pricing guidance they proffer. (Which is why I conceded that perhaps AMD maybe just shouldn't have made any statement regarding pricing at all - but again, this is at odds with marketing convention...)

What? You're seriously arguing that AMD has made an uncompetitive product so cannot help but lie about the price in order to get reasonable reviews?

cynan wrote:

I don't know why people are upset about the price/performance graph. Sure, it's not accurate.

I don't even.

cynan wrote:

But with the limited Vega supply and mining demand, any prices AMD announced prior to launch were likely never going to be accurate for any length of time. This doesn't stop people form looking at Jeff's pricing plots and then making whatever mental adjustment is appropriate given current pricing.

The plots exist to tell a story, it should be the true one.

Vega 64 does not have the performance, its selling point is price-performance -- but it doesn't have that either.

Reviewers wouldn't've published charts with the Radeon RX Vega 64 card at $1 on the x-axis if AMD had made secret rebates to retailers so that each could sell 1 Radeon RX Vega 64 card for $1 -- they would have smelled the rat and not misled the readers. This is different only in degree, AMD's Radeon Technologies Group supported the pricing with hidden rebates for a very small quantity of cards to be available at the review price, in order to get review conclusions of "competitive" when the truth is "uncompetitive".

Reviews only exist to tell the truth to readers, and validate or invalidate marketing claims. They do not do so when they do not tell the truth.

Topinio wrote:

Yes. But it shouldn't lie, or trick reviewers into publishing price-performance plots which are lies.

Trick reviewers into price performance plots? Nobody at AMD held a gun to reviewer's head (or as far as I know even suggested) that a price/performance plot was an integral part of a review of their product. Because of the supply/demand situation, the price performance plot was never going to be accurate. That doesn't mean that these plots aren't generally useful for GPU launch reviews, just not particularly so with Vega.

Topinio wrote:

What? You're seriously arguing that AMD has made an uncompetitive product so cannot help but lie about the price in order to get reasonable reviews?

That's exactly what happened. Strictly in terms of profit, relative to Pascal, Vega, at current performance, with current HBM2 prices/supply, for gaming, is absolutely uncompetitive. And if you think profit isn't the most important motivator for a publicly traded company... Sure, AMD could have been more upfront about the limited preorder quantities at launch prices, put saying that they lied is at least as disingenuous.

You yourself even state that Vega isn't competitive with Pascal, and then act surprised when I make that claim that this is the impetus for them wanting to get a reasonable amount of profit out of Vega by immediately deciding to raise launch price once they saw the demand was there... So I don't even either.

Topinio wrote:

Vega 64 does not have the performance, its selling point is price-performance -- but it doesn't have that either.

Topinio wrote:

Reviewers wouldn't've published charts with the Radeon RX Vega 64 card at $1 on the x-axis if AMD had made secret rebates to retailers so that each could sell 1 Radeon RX Vega 64 card for $1 -- they would have smelled the rat and not misled the readers. This is different only in degree, AMD's Radeon Technologies Group supported the pricing with hidden rebates for a very small quantity of cards to be available at the review price, in order to get review conclusions of "competitive" when the truth is "uncompetitive".

Reviews only exist to tell the truth to readers, and validate or invalidate marketing claims. They do not do so when they do not tell the truth.

Sure. And every reviewer is entitled to update their review should they feel that AMD has misled them, in big bold letters. Selling a few hundred cards at $499 was a promotional marketing gimmick only in retrospect. Had all the Vega launch units not sold, do you still think AMD would be raising the price to $599? Complaining that Vega had a successful launch (at least as much as possible given supply, from a business standpoint) is nothing more than consumer entitlement.

The fact of the matter, the truth, is that there is no static price for Vega right now. Newegg is currently selling Vega 64 for $690, with different prices at other retailers, when AMD themselves claimed that their regular price would now be the bundle price ($599). You really think that AMD, given, as stated, the uncompetative stance of Vega 64 at $499 (again, from a profit perspective), will just sit idly and let retailers like Newegg make $100 more profit on the $689 price when they could take it themselves, just so they can win some sort of dubious moral pricing mindshare with customers? To what end? So that they have to stop selling gaming GPUs altogether when their RTG division goes bankrupt?

cynan wrote:

Trick reviewers into price performance plots? Nobody at AMD held a gun to reviewer's head (or as far as I know even suggested) that a price/performance plot was an integral part of a review of their product. Because of the supply/demand situation, the price performance plot was never going to be accurate. That doesn't mean that these plots aren't generally useful for GPU launch reviews, just not particularly so with Vega.

Then do you think the reviewers knew that AMD was selling cards through the channel at well over $499 to the retailers, and rebating them to hit that price point on a very small number of cards?

Jeff, did you know that the regular price was $599 with "free" games for both the black and silver cards, $699 for the aqua card, plus taxes?

cynan wrote:

Sure, AMD could have been more upfront about the limited preorder quantities at launch prices, put saying that they lied is at least as disingenuous.

Lie of omission, and I disagree it's at all disingenuous to point it out. AMD said $499 MSRP for the RX Vega64, £450 in the UK and chose not to disclose that this was a ‘launch day only’ price. This was planned.

cynan wrote:

You yourself even state that Vega isn't competitive with Pascal, and then act surprised when I make that claim that this is the impetus for them wanting to get a reasonable amount of profit out of Vega by immediately deciding to raise launch price once they saw the demand was there... So I don't even either.

Nonsense. They planned it for $599. $499 was via a very short-lived rebate programme.

cynan wrote:

Topinio wrote:

Reviewers wouldn't've published charts with the Radeon RX Vega 64 card at $1 on the x-axis if AMD had made secret rebates to retailers so that each could sell 1 Radeon RX Vega 64 card for $1 -- they would have smelled the rat and not misled the readers. This is different only in degree, AMD's Radeon Technologies Group supported the pricing with hidden rebates for a very small quantity of cards to be available at the review price, in order to get review conclusions of "competitive" when the truth is "uncompetitive".

Reviews only exist to tell the truth to readers, and validate or invalidate marketing claims. They do not do so when they do not tell the truth.

Sure. And every reviewer is entitled to update their review should they feel that AMD has misled them, in big bold letters. Selling a few hundred cards at $499 was a promotional marketing gimmick only in retrospect. Had all the Vega launch units not sold, do you still think AMD would be raising the price to $599?

"Only in retrospect" for those of us who did not know the real plan, or the price that retailers (say they) had to pay for those $499 cards. Whether those writing the reviews knew only in retrospect is something only they can tell us.

cynan wrote:

The fact of the matter, the truth, is that there is no static price for Vega right now. Newegg is currently selling Vega 64 for $690, with different prices at other retailers, when AMD themselves claimed that their regular price would now be the bundle price ($599). You really think that AMD, given, as stated, the uncompetative stance of Vega 64 at $499 (again, from a profit perspective), will just sit idly and let retailers like Newegg make $100 more profit on the $689 price when they could take it themselves, just so they can win some sort of dubious moral pricing mindshare with customers? To what end? So that they have to stop selling gaming GPUs altogether when their RTG division goes bankrupt?

AMD can price it where they like, I have no problem with that. What I have a problem with, is claiming it's a $499 card and generating reviews with price-performance charts based on that claim, while making retailers pay over $499 for it so they cannot sell it for $499.

OK so I bit the bullet and paid another £135 for a top quality 850w PSU and I can now say that for 1440p, with a freesync monitor, Vega 64 AC is a great performer. In a like for like comparison with my old Fury Pro it's between 25% to 50% faster depending on the game which is fine by me. I probably needed a new PSU since my 7700K purchase so I'll chalk that down to experience. So it comes down to price really. I paid £450 for mine which I think is fair value. At £599 no chance I'd have gone 1080ti without a doubt. I'm still not happy that most of the new features are still disabled in the drivers but hopefully that will improve. With Vega 56 you'll need an AIB card and at least a 750w power supply to make it fly and with that combination it should smoke a 1070. Again it will all depend on the price and if it's dearer it'll be a harder sell. No doubt we'll see in time.

Lordhawkwind wrote:

So it comes down to price really. I paid £450 for mine which I think is fair value. At £599 no chance I'd have gone 1080ti without a doubt. I'm still not happy that most of the new features are still disabled in the drivers but hopefully that will improve. With Vega 56 you'll need an AIB card and at least a 750w power supply to make it fly and with that combination it should smoke a 1070. Again it will all depend on the price and if it's dearer it'll be a harder sell. No doubt we'll see in time.

Yeah, £450 is reasonable for the RX 64, and gz on getting one. Unfortunately, I was busy during the minutes it was available for that price

I might settle for a RX 56, though as you say it will all depend on the price: if it's under £350 long enough to be bought, I'm in; otherwise -- or if it's insanity and over £400 -- I'll stick with the R9 290X or chuck in a RX 580 for < £225 (if I get the delivery).

Hello GTX 1080 in three days. I really wanted to purchase Vega, and maybe if I were feeling irrationally patient, I would. But I wanted the best deal now and for me that was not Vega. Will still consider AMD in future CPU/GPU purchases...

I've been seeing multiple posts from people saying that GPUs are not primarily about gamers anymore, that mining just means we can expect the market to be affected by those who are willing to pay more and gamers are just going to have to deal with it for the foreseeable future.

Dysthymia wrote:

Hello GTX 1080 in three days. I really wanted to purchase Vega, and maybe if I were feeling irrationally patient, I would. But I wanted the best deal now and for me that was not Vega. Will still consider AMD in future CPU/GPU purchases...

I've been seeing multiple posts from people saying that GPUs are not primarily about gamers anymore, that mining just means we can expect the market to be affected by those who are willing to pay more and gamers are just going to have to deal with it for the foreseeable future.

I heard someone joked about who would even buy AMD cards anymore if they lost in both gaming AND mining...With Vega I can actually see that happening.

Topinio wrote:

AMD can price it where they like, I have no problem with that. What I have a problem with, is claiming it's a $499 card and generating reviews with price-performance charts based on that claim, while making retailers pay over $499 for it so they cannot sell it for $499.

Do we actually know that AMD is charging retailers higher than MSRP for these cards, or is this just a wild claim? I'd be extremely surprised if this was true.

Well one thing I think we can say is:

Goodbye Vega *56*.

(Edited: I see that the release date hasn't been met yet, still - I think it's unlikely there will be much on offer when it is released. Probably a minor required supply from each vendor and then that's it, at least for awhile.)

CScottG wrote:

I've no real interest in their video cards, but frankly this is off-putting enough that I wouldn't want *anything* from AMD (..which is too bad considering that the industry needs the competition).

I kind of feel sorry that Scott is actively involved in a Co. doing this sort of thing..

This is a truly staggeringly mystifying, self foot-shooting position to take.

Let me get this straight. You want competition, but if the competition isn't neck-and-neck as successful as the competition (ie, Nvidia) then you'd rather no competition.

A very real perspective is that consumers should be happy about any competition. For them, as far as long-term value is concerned, anyway, any competition instead of no competition can only be a win.

Does that mean you have to like the company? Like their currently more desperate pricing/marketing strategy (again, relative to the current competition, anyway) of supply-constrained, expensive-to-make, sold out products? No.Of course not.

But at the end of the day, it's about competition. And if that means that in 6, 12, 18 months from now AMD has a video card (or other product) that you deem seems to be a good value for the level of performance you're interested in, why wouldn't you consider it?

cynan wrote:

Let me get this straight. You want competition, but if the competition isn't neck-and-neck as successful as the competition (ie, Nvidia) then you'd rather no competition.

It was edited, yes? (as a screw-up on my part for not remembering the delay for the "56" version. I was looking through various manufacturer's sites and hadn't seen any "56's" and had a WTF? moment. This of course prompted me too go back and re-read the reviews and see my error.)

As for my intense dislike: that's a factor of a perceived and continued (..though not yet accurate, time will tell there) marketing method that is flat-out deceptive. Again though, it may not turn-out that way, "56's" may be far more than simply a token marketing ploy. Of course this has nothing to do with success, rather its the manner that business is being conducted. Moreover I'm not saying I'd rather have no competition. Instead it's just a personal choice that I would not do business with a Co. anytime soon that, regardless of product, conducted itself in a manner I find repugnant.)

Since this thread branched out on topic a bit, I'll chime in.

1080's are only $10 over MSRP. The 1080 Ti sell at MSRP. Both are in stock and widely available. Vega 64 was supposed to sell at 1080 pricing, and Vega 64 Liquid is priced the same as 1080 Ti cards.

Vega 64 Liquid vs a 1080 Ti isn't even a contest in most games, yet the Vega part draws >100w more power and runs louder for it. There isn't any saving grace for this card. the last time AMD tried to launch a product with this bad a matchup it was called the HD 2900 XT, and they were justifiably ripped for doing so. Yet the blowback over Vega 64 Liquid has been pretty muted this time around.

Kougar wrote:

Since this thread branched out on topic a bit, I'll chime in.

Vega 64 Liquid vs a 1080 Ti isn't even a contest in most games, yet the Vega part draws >100w more power and runs louder for it. There isn't any saving grace for this card. the last time AMD tried to launch a product with this bad a matchup it was called the HD 2900 XT, and they were justifiably ripped for doing so. Yet the blowback over Vega 64 Liquid has been pretty muted this time around.

Yes. Unless you are already invested in FreeSync, or for some reason find FreeSync highly desirable, then absolutely, at current pricing, Vega 64 should be a failure. With gamers, anyway.

But if they can sell their limited supply at current from-a-gaming-perspective-uncompetitive prices then this poses a sticky situation for AMD that the non-supply constrained Nvidia, who set the current pricing structure based on better performing, cheaper-to-make GPUs, isn't facing.

AMD could have simply come right out and stated. "Look, Vega is kind of a gaming failure, it's expensive, and we don't have a lot of supply due to HBM2, so we're just going to market the thing to GPU miners while we can". But then this obviously erodes their gaming mind share. And, as they probably fear, will undermine current inroads with consumers, game developers, monitor makers (FreeSync) , and most importantly, their position as a major source of PC gaming hardware.

AMD could have just taken a bath on Vega, and sold all of their limited supply of Vega 64 at gaming-competitive prices (ie, the $499 price advertised for launch). This makes little sense from AMD's perspective, because if people (ie, miners) are wiling to pay more for these cards, then all that will happen in the near term is the retailers will make the profit for themselves - as AMD has had to deal with in the past with polaris and, in the bitcoin boom, Tahiti, etc.

So instead, they have chosen a much-maligned-here middle road. Which, objectively, has been to price a small proportion of their limited Vega supply competitively at launch through retailer rebates. This gives AMD the flexibility of immediately increasing their sorely needed revenue from Vega sales if the demand is there (which it seems to be) or, if Vega isn't selling (which seems not to be the case), making the launch rebate permanent (to the disgruntlement of many prospective AMD customers, as voiced in this thread).

If taking AMD's perspective, I find it hard to begrudge them taking something along the lines of that last option.

cynan wrote:

Kougar wrote:

Since this thread branched out on topic a bit, I'll chime in.

Vega 64 Liquid vs a 1080 Ti isn't even a contest in most games, yet the Vega part draws >100w more power and runs louder for it. There isn't any saving grace for this card. the last time AMD tried to launch a product with this bad a matchup it was called the HD 2900 XT, and they were justifiably ripped for doing so. Yet the blowback over Vega 64 Liquid has been pretty muted this time around.

Yes. Unless you are already invested in FreeSync, or for some reason find FreeSync highly desirable, then absolutely, at current pricing, Vega 64 should be a failure. With gamers, anyway.

But if they can sell their limited supply at current from-a-gaming-perspective-uncompetitive prices then this poses a sticky situation for AMD that the non-supply constrained Nvidia, who set the current pricing structure based on better performing, cheaper-to-make GPUs, isn't facing.

AMD could have simply come right out and stated. "Look, Vega is kind of a gaming failure, it's expensive, and we don't have a lot of supply due to HBM2, so we're just going to market the thing to GPU miners while we can". But then this obviously erodes their gaming mind share. And, as they probably fear, will undermine current inroads with consumers, game developers, monitor makers (FreeSync) , and most importantly, their position as a major source of PC gaming hardware.

AMD could have just taken a bath on Vega, and sold all of their limited supply of Vega 64 at gaming-competitive prices (ie, the $499 price advertised for launch). This makes little sense from AMD's perspective, because if people (ie, miners) are wiling to pay more for these cards, then all that will happen in the near term is the retailers will make the profit for themselves - as AMD has had to deal with in the past with polaris and, in the bitcoin boom, Tahiti, etc.

So instead, they have chosen a much-maligned-here middle road. Which, objectively, has been to price a small proportion of their limited Vega supply competitively at launch through retailer rebates. This gives AMD the flexibility of immediately increasing their sorely needed revenue from Vega sales if the demand is there (which it seems to be) or, if Vega isn't selling (which seems not to be the case), making the launch rebate permanent (to the disgruntlement of many prospective AMD customers, as voiced in this thread).

If taking AMD's perspective, I find it hard to begrudge them taking something along the lines of that last option.

Yeah, instead AMD is destroying what little good will the enthusiast community may have remained for RTG. That includes the press and those of us that actually buy these products. You can say "AMD had no other choice", but there are always choices. The problem is that in GPU design, the choices are often made years in advance. When Vega was being developed Pascal was still in the pipeline as well so the performance target for Vega was simply "greater performance than Maxwell". AMD made BAD decisions when developing Vega. They handcuffed themselves to HBM2, and GloFo's shyte 14nm process which are the primary reasons Vega is:

late

expensive

power-hungry

under-performing

Vega 56 is unavailable as of yet, and thus the only conversations about its availability/pricing are speculative
If you want to speculate about it, we should look at the precedent set by Vega 64 to inform said speculatory discussion, because the reality is that Vega 56 does not exist in a vacuum.

With that, Vega 64 can be summed up in the following way:

It's slower than 1080 Ti by a mile, while consuming more power

It's also not available for purchase outside of a bundle, which means consumers have to pay large sums (read: at least $1000 on Newegg right now) to buy one

It ~ matches GP104 on average (wins DX12, loses DX11), a ~ 15 month old, 165-180W, $379-549 MSRP product family

Face it, this is going to go down as one of the worst launches in GPU history. There is no reason to buy Vega, unless you're some kind of hardcore AMD fan who thinks Saint AMD is a charity, or you already own a Freesync monitor and aren't just satisfied with high refresh rates and feel like you have to take advantage of variable refresh rate or you will have wasted your money on said monitor purchase.

I'm tired of the spin, and tired of the obtuseness. I'm sure I'm not the only one. These are the reasons I barely post on forums anymore. It's worse than politics/religion. No one can have an objective conversation anymore.

techguy wrote:

I'm tired of the spin, and tired of the obtuseness. I'm sure I'm not the only one.

Probably out of context for the point you were making, but in regards to AMD spinning Vega for over a year and its "poor Volta" campaign, AMD's marketing does seem to have reached Trump levels of spin. Partly why I an so vocal about deriding them over it.

Vega was delayed 8 months. Sure AMD invested in HBM2, but that doesn't mean they have to use it if it doesn't provide any benefit. An 8-month delay is long enough to respin and design a GDDR5X model that they can offer at more competitive pricing, but they chose to not do that.

Something I didn't see mentioned yet is don't forget that HBM2 was said to save something like 40w over GDDR5X memory. AMD really wanted that ~40w power savings to use back on the GPU-side. It's mind-boggling. Even with a 40w savings from HBM2 factored in Vega still draws over 150w more than a 1080.

Kougar wrote:

Vega 64 Liquid vs a 1080 Ti isn't even a contest in most games, yet the Vega part draws >100w more power and runs louder for it. There isn't any saving grace for this card. the last time AMD tried to launch a product with this bad a matchup it was called the HD 2900 XT, and they were justifiably ripped for doing so. Yet the blowback over Vega 64 Liquid has been pretty muted this time around.

There are no bad cards, only bad prices. I'd take Vega any day of the week and twice on Sundays for $200.

I don't know if current pricing is completely an artifact of miners, or is also a bad price point. I guess we'll see eventually, since in the medium term AMD will simply increase card production (as much as they can).

Vhalidictes wrote:

There are no bad cards, only bad prices. I'd take Vega any day of the week and twice on Sundays for $200.

Yep. It's only been a week since release, and what with all this mining crap, I don't understand the negativity.

The Vega cards are significantly more appealing if you're also in the market for a monitor. The savings of going with Freesync versus an equal G-Sync model can be as much as $300 (as stated in the TR review), and if your choice also happens to be on the (upcoming) list for the "Radeon Red Pack", you can effectively net yourself another $100 off, for a total savings of $400.

There are apparently four package variants of Vega https://videocardz.com/72173/there-are- ... u-packages

I think most would agree how bad it is when the ASIC height differs between variants and filling it in with extra TIM is not a good thing... no wonder third party cards are delayed launching.

I am totally not regretting my 1080Ti purchase at launch right now.

I think AMD is digging themselves into a huge hole. Their Vega GPUs were late, they misrepresented performance, and they misrepresented price. With all these cards being bought for mining, their share of the high-end GPU market (and possibly mid-level, with Polaris MIA at reasonable prices) is going to further slip. In an industry where software optimizations make or break a product, you don't want to get the off-brand (which I experienced using my 290X, when some games just didn't play well on AMD). Compounding a poor showing, AMD comes off as really shady from a marketing and communications perspective, further solidifying their perception as the cut-corners-product vendor.

How low can their marketshare go?

That's funny. I had completely the opposite experience with my Radeon R9-290. I bought it for MSRP on launch day, and it still provided good gaming performance for several years after the contemporary NVidia cards were no longer able to keep up with the latest titles. Other than being noisier that I would like, it was a complete success.

Other than occasionally-buggy NVidia drivers and issues with the resolution of my secondary monitor, I have also been quite pleased with my GeForce GTX980Ti. Maybe I'm just easier to please than some folks.

Went full on with AMD a few years ago and picked up a MG279Q 144Hz monitor (35-90Hz VRR) at the time to get the eventual benefits of FreeSync.
I then paired it with a R9 390 (instead of a Fury) and have managed to tame the temperatures (primarily undervolting) and get great use out of it, but now want higher performance.

I do still have a nVidia GTX 760, but that's where I left the green team, since I wanted to go with a VRR monitor, but was unwilling to pay the gsync tax when all nVidia had to do was adopt the VRR tech of DP1.2a and above.

The RX 480/580 was not enough of a jump for me even though they used less power then my R9 390, but only offered similar FPS and frame time performance.

Thanks to TR for the great review of the Vega 56/64, and I have no issues with these cards even with the higher power requirements comparitively to the green teams offerings.

What I take exception to is the crazy pricing AND availability. There is no true or consistent price discovery and hence performance/dollar discovery right now.
It feels like trying to buy something from a seller with product listings but no advertised prices or true prices.
If I ask "how much?" - the answer from the seller is "how much are you willing to offer?" - sorry, out of stock.

I would buy at something close to MSRP if possible - no luck so far.

I think this is damaging for AMD's reputation, but it seems like desperation to try anything to maximize profits, which includes variable pricing schemes/bundles with availability guessing games.
Ultimately AMD will do whatever it takes to maximize profits - likewise, I do my best to minimize my gaming hardware expenses. My R9 390 still works.
AMD is not a charity and likewise, I'm not a philanthropist for AMD either.

Delays are to be expected when OEMs will have to figure out what to do about the z-height difference between molded and un-molded Vega chips, then also what to do about the even/uneven HBM2 difference on top of that. Going to take time for OEMs to make 2-3 different baseplate designs and match them up with the correct Vega chips.

That AMD would be using so many different Vega chip variants speaks to how much of a shortage of HMB2 there must be. I wonder if they regret their plan to spur HBM2 adoption yet...

Kougar wrote:

Delays are to be expected when OEMs will have to figure out what to do about the z-height difference between molded and un-molded Vega chips, then also what to do about the even/uneven HBM2 difference on top of that. Going to take time for OEMs to make 2-3 different baseplate designs and match them up with the correct Vega chips.

That AMD would be using so many different Vega chip variants speaks to how much of a shortage of HMB2 there must be. I wonder if they regret their plan to spur HBM2 adoption yet...

I agree, Kougar. I can forgive AMD for trying to make a profit off miners. I can understand how there is going to be availability problems in any case with a new GPU. If AMD can't source enough GRAM from a specific supplier and needs to scrounge that's just how it has to be.

But different packages?! There's no good reason for that to be happening and it's clearly in AMD's court. And it's worse than people are making it out to be. Someone, somewhere, is going to have either a badly engineered or badly mounted heatsink. It's just going to happen, and all those failures are going to give Vega a bad reputation.

JustAnEngineer wrote:

Maybe I'm just easier to please than some folks.

Thanks for the jibe. Factual issues I had with my 290X:
-Getting FreeSync to work with two monitors (one fullscreen game, one extra) was extremely hit-or-miss and require both voodoo rituals and trial-and-error
-On top of that, FreeSync would respond oddly to individual titles (Diablo 3 required fullscreen windowed for FreeSync, others required FullScreen, one title only seemed to work in windowed)
-Some gaming titles (e.g. Battleborn) suffered huge penalties with Radeon hardware
-Intermittent crashing in certain titles (e.g. FFXIV DX11)

But yes, I'm tough to please in expecting my hardware features to work well across all gaming titles. Or maybe I just want to boot up a game after work and not fiddle with unplugging monitors or fiddling with settings to get FreeSync working. I guess I just have less free time than you?