Pricing Discussions around AMD VEGA *over-flow*

[entity279]

Yes, I think there is also some headroom left, because i.e. for Polaris cards there are some modifications "avalilable" allowing them to be more optimized and thus achieving higher than usual MH rate. At the moment those are not yet available for Vega [...]

For polaris there are people on the web who can tell you for the any memory part number exactly what hashrate you can reach and how. Some will also make the bios file for you. There's such a large part of the mining comunity behind AMD atm, some with substantial resources..

HMB2 is a more complex memory system, but maybe there are enoungh enthusiasts (or greedy companies out there, not that they would share their results) to reverse engineer theyr way to that theoretical 65 MH/s hashrate

[..]

And they should have not have released specialized mining logic in their drivers if the primary target of the cards was gamers.

No one is to blame but greedy AMD .....

It's not like they don't need all the money they can get. It's not like the shareholders and their dogs won't sue them for deliberatly hindering the sales of their own products

 

[Deleted member 2197]

Well there's a difference between short term and long term with regard to shareholders, product supply and the dogs.
Most shareholders are "long" in the market and the dogs most likely aren't starving.

 

[leoneazzurro]

I never spoke about 65 MH/s hash rate, I think however that 45 could be achievable, maybe a little more.

 

[entity279]

I was reffering to that post on reddit where the supposebly knowledgeable miner programmer was stating that 65 mhs should be the teoretical maximum for the HBM2 memory as deployed on Vega.

Of course they may be other (subtle) bottleneks and that may never be achievable,

 

[CarstenS]

And they should have not have released specialized mining logic in their drivers if the primary target of the cards was gamers.

That was purely a fix for the DAG regression that mainly (but not only) affected polaris cards already in the wild. Neither (Nvidia nor) Hawaii GPUs were affected and Fiji also to a very minor extent.

With the driver, Vega just returned to the level previously achievable with Vega FE and it's older drivers with older DAGs.

For polaris there are people on the web who can tell you for the any memory part number exactly what hashrate you can reach and how. Some will also make the bios file for you. There's such a large part of the mining comunity behind AMD atm, some with substantial resources..

HMB2 is a more complex memory system, but maybe there are enoungh enthusiasts (or greedy companies out there, not that they would share their results) to reverse engineer theyr way to that theoretical 65 MH/s hashrate

Then why were those people never able to do the same for Fiji+HBM, which even has greater memory bandwidth?

The reddit guy only calculated according to raw bandwidth under the assumption of:
"1 Ethash requires 8KB of memory bandwidth, 1 MH requires ~7.8 GB of memory bandwidth"

https://www.reddit.com/r/Amd/comments/6rm0cs/vega64_mining_is_unlikely_to_exceed_65_mhs/

He is under false assumptions concerning the memory and it's OC.
edit: And his math is off as well. 483,8/7,8 = 62,02

 

[Kaotik]

-edit: wrong thread, re-posted here: https://forum.beyond3d.com/posts/1997386/

 

[w0lfram]

Mabye Pascal feels old and dying to you because NV released it over a year ago? Sure one can always wait for the AMD driver team to magically unlock VEGA, but then that is the same team that barely could make fully functioning launch driver. I have the feeling that at the time AMD might have improved VEGA drivers to a meaningful level, VEGA will no longer be facing Pascal but Volta and that will be hugely one sided.

I see no issues with AMD drivers.

Matter of fact, I see high quality and leading technology in everything AMD is doing right now. I am not a bradnwagoneer and in the past Nvidia had better drivers & a better software suite. But not any more and my gaming has turned to Radeon. The quality of the brand has grown considerably in the last 3 years and now far outpaces their competition and the industry.


AND Vega is the latest technology on the streets. That is why it requires a premium.

Vega vs Pascal isn't about raw game numbers, it is about about building blocks for the future games. If RX Vega can run last years games, equal to their competitors gtx1070 & 1080, then it is a far better value to the end user and Gamer.

Again, not being attached to any brand and logically looking at now and the future, any unbiased Gamer building/buying a new rig over the upcoming holiday are not going to pass up the technology Radeon-Vega brings, all because 2 year old Pascal can crunch certain outdated games extremely well?

Latest tech = Premium



Honestly, what would buying a Pascal card today, get you.. ? over buying a Vega card..? And in the future..?

Even at equal cost, Vega is a much better choice for the Premium Gamer, because RX supports the most standards, better compute, latest VR tech, prepped & ready for future game titles and game engines. AMD has announced many Gaming partners and pranced many of them on-stage at many of their events. They always mention their "partners" in tweets, etc. AMD has a lot of collaboration in the gaming industry and what seems like a lot of Development studios are on board and behind them.

I suspect AMD is holding back, and Dr Su is a Gamer herself. I think there are several huuuge announcement yet to be made about RXv64. Anyone remember that cube? And what ever happened to "infinity fabric" and why are we locked out of it, or kept in the dark..? What is AMD brewing?

I believe there are mystical RX Vega drivers... I have seen them in my dreams. (And on presentation slides)




I am not worried about $50 here, or there. I see RXv as the latest tech and demands a premium. We have all seen Pascal, Polaris, Fiji and other GPU uArchs get mystical drivers during their life times too. So, perhaps it isn't mystical, but logical to think that refinement will come for RX Vega (fine wine?)

I am buying RX Vega because it won't hinder my gaming.

 

[Rootax]

Raja, is that you ?

 

[w0lfram]

Don't hate.
If either were sitting on a table and you could take one to keep...

 

[Geeforcer]

I suspect AMD is holding back, and Dr Su is a Gamer herself. I think there are several huuuge announcement yet to be made about RXv64. Anyone remember that cube? And what ever happened to "infinity fabric" and why are we locked out of it, or kept in the dark..? What is AMD brewing?

I believe are mystical RX Vega drivers... I have seen them in my dreams. (And on presentation slides)

Those are all good points. AMD has Nvidia right where they want them, having lulled them into sense of security with seemingly an epically muddled launch. Then, when Nvidia least expects it, they will spring their trap. I mean, it would be pretty idiotic to show your product in a good light during the launch window, that's exactly what they expect you to do. You make you HUUUUUGE announcements LONG AFTER the reveiws have been published. And yes, Holocube is a game changer. Sure, to the untrained eye it may look like a marketing gimmick that neither ships with the card nor has any impact on it's perfromance, but it's fairly clear it will bring on huge improvements to Vega. How, you may ask? Infinity Fabric, of course. Having evolved beyond a mere data interconnect infrastructure, the technology has transcended our limited understanding of electromagnetic and is now fully capable of connecting any implement to the GPU while multiplying it's power 10-fold. To fully grasp this unprecedented capability, just glance around the room. See that fridge, with it's 3 hp compressor? By the power of Greysk... Infinity fabric it links straight into Vega rendering pipeline, at 10x power. Can you imagine what kind of polygon throughput you can attain with 30 horsepower thrown in? And that's just from one generic appliance; Holocube can render reality in less then 12 parsecs (less than 1.2 parsecs after 10x Infinity Fabric Boost).

That doesn't even get into what AMD is brewing, which, spoiler alert, is an authentic Trappist ale (firm, earthy head, nutty nose, honey and nutmeg tone, smooth finish with a hint of allspice and ripe plum, score: 99), which is the only way to refresh yourself, having witnessing reality-warping graphics performance of Vega.

 

[Kaotik]

https://www.pcgamesn.com/amd/amd-rx-vega-launch-price-reduction

“Our SEPs, and the price tag that we announced,” Youngblood says, “is our full intention of where we would suggest the product be priced. Not just for launch, but ongoing.”

 

[CarstenS]

Nice of AMD to clarify their earlier statement a couple of days after still without clearly ... ah well. We'll see.

 

[xEx]

At what point has AMD lied about MSRP? It's $499 for just the card, $599 for Radeon Pack and $699 for Radeon Pack Aqua, it was that at launch, it is still that and it will be that in the future.
Now we have even AMD rep re-confirming it https://forum.beyond3d.com/posts/1997432/

Just to answer your question. It seems that the vega VGAs cost retailers 509.13 dollars(they claim to make 15% profit at 599) so the mrsp of 499 is just not possible, that would require the VGA to cost them 424.15 and that would be about 30% profit for the 599 which they say is too much(under normal circumstances).

 

[Kaotik]

Just to answer your question. It seems that the vega VGAs cost retailers 509.13 dollars(they claim to make 15% profit at 599) so the mrsp of 499 is just not possible, that would require the VGA to cost them 424.15 and that would be about 30% profit for the 599 which they say is too much(under normal circumstances).

There are no just one type of "Vega VGAs", there's different SKUs which cost different amount for the retailers too

 

[xEx]

There are no just one type of "Vega VGAs", there's different SKUs which cost different amount for the retailers too

Well the "499 SKU" were only given to few retailers around the world and only in very limited quantity. If it is compelling with the false marketing laws IDK but it certainly was in the limit.

 

[Kaotik]

Well the "499 SKU" were only given to few retailers around the world and only in very limited quantity. If it is compelling with the false marketing laws IDK but it certainly was in the limit.

It was in really limited quantities, but they will bring out more of them

 

[Deleted member 13524]

Youtuber Jayz2cents claims to have a source that directly contradicts Gibbo's statements about AMD directly making those rebates.

 

[CarstenS]

This? "Now if what my sources telling me is true then it makes what Gibbo was saying sound a little bit hard … to differentiate the facts between - I don't know what the motive would be - to actually call this a backend rebate if it's not the case. So this is ground zero for the problem that we have today."
Not sure if I understood correctly, but what does it mean?

 

[Malo]

Not sure if I understood correctly, but what does it mean?

Basically that according to his source, whoever that may be, Gibbo is wrong and there was no rebate.

 

[Deleted member 2197]

I thought the AIB's backed up Gibbo ....
AIB Partners to GN: Gibbo Was Right About AMD Price Changing

In speaking with different sources from different companies that work with AMD, GamersNexus learned that “Gibbo is right” regarding the AMD rebate expiry and subsequent price jump. AMD purportedly provided the top retailers and etailers with a $499 price on Vega 64, coupling sale of the card with a rebate to reduce spend by retailers, and therefore use leverage to force the lower price.

http://www.gamersnexus.net/news-pc/3023-aib-partners-to-gn-yes-vega-price-will-change