AMD powers 83pc of the Global VR System market says JPR

[scaryjim]

Question is does anyone know ...

Well, the claim is repeated in the Hexus article and is therefore presumably in AMD's press release, and it would be foolish of them to include that if it wasn't true.

... or have a link to the actual JPR report? ...

Full JPR reports costs $1500, and will almost certainly land anyone who distributes it without permission in hot water, so I'm guessing not.

Fudzilla links to JPR's highlights of the report, but their interpretation of the publicly available figures is off; the table they describe as high end AIB market share is actually the total AIB market share. I assume the full report includes a more complete breakdown, which may well show that AMD's share of the enthusiast market is much higher. Without access to it we can't be sure; however it looks like enthusiast board shipments were less than 6M vs almost 40M total. That makes enthusiast AIBs only 15% of the market, meaning that AMD has plenty of overall share to be doing well in that segment.

The Fudzilla article makes a big hooha out of the supposed lack of processing power for Playstation VR and how AMD are out of line for claiming it as part of their market, but that ignores the fact that PSVR is a hardware-locked platform so MUCH easier to optimise for. It may not provide the same level of visual fidelity as a high end gaming PC, but it should still provide a fluid, low-latency solution with good optimisation. So that's not really a fudge by AMD - PSVR is absolutely a part of the addressable VR gaming market.

The other point Fudzilla misses is that AMD are talking about the total addressable VR market, not just the VR gaming market. So AMD may well be including all their GCN APUs in their figures, if they'll be able to address "entertainment, education, science, medicine, journalism and several other exciting fields" with them. We've already got a report of a Carrizo APU-based stand-alone headset, so that's not unrealistic, and those are fields that are potentially less performance oriented than gaming.

[DanceswithUnix]

Fudzilla links to JPR's highlights of the report,

That is *a* report, I don't think it is *the* report that AMD were referring to though.

[Corky34]

Well, the claim is repeated in the Hexus article and is therefore presumably in AMD's press release, and it would be foolish of them to include that if it wasn't true.

Good job we can depend on no one ever having interpreted statistics in a way to paint something in a different light then isn't it, what's the phrase?
Lies, damned lies, and statistics.

Full JPR reports costs $1500, and will almost certainly land anyone who distributes it without permission in hot water, so I'm guessing not.

So not something that can be easily verified by an independent third party then.

Personally i like this interpretation of what AMD are claiming, it seems fairly well balanced and objective to me but what do it know.

[kalniel]

So not something that can be easily verified by an independent third party then.

Plenty of independent tech sites have a JPR subscription.

[Corky34]

Just none that (from what my searches have turned up) have looked at the details of it, outside of what AMD claimed most sites see to be basing their reporting on either the press release from AMD or guesses.

[scaryjim]

... Personally i like this interpretation of what AMD are claiming, it seems fairly well balanced and objective to me but what do it know.

It's fairly well balanced if you're *just* looking at the VR gaming market. That's not what AMD said. I even quoted the list of other VR markets they're talking about. Gaming might not even be the biggest VR market.

But even if we do just look at the gaming market, the PS4 has hit over 30 million units shipped. The enthusiast GPU market in 2015 was just 6 million units. Even if nvidia had 100% share of the enthusiast GPU market, add the PSVR and AMD has 83% of the addressable market.

Huh, that was relatively easy after all...

[lumireleon]

AMD gets 83% becoz NVIDIA has a bad habit of forcing developers to use their in-house source code which has some moron like regulations of anti-sharing behavior(CUDA, G-SYNC, Phys-X) while AMD has full support for OPEN SOURCE CODE..(OPEN CL, VULKAN, FREE-SYNC) that's why SONY & MICROSOFT made their consoles from AMD and here we go again with VR.

[kalniel]

AMD gets 83% becoz NVIDIA has a bad habit of forcing developers to use their in-house source code

I don't think so. It's more likely the reasons already mentioned - consoles and APU are considered 'VR-ready'.

[Corky34]

It's fairly well balanced if you're *just* looking at the VR gaming market. That's not what AMD said. I even quoted the list of other VR markets they're talking about. Gaming might not even be the biggest VR market.

But even if we do just look at the gaming market, the PS4 has hit over 30 million units shipped. The enthusiast GPU market in 2015 was just 6 million units. Even if nvidia had 100% share of the enthusiast GPU market, add the PSVR and AMD has 83% of the addressable market.

Huh, that was relatively easy after all...

Obviously you're free to make up your own mind on how valid AMD's claims are but personally, and going on past experience of AMD's claims, I'm going to take their claim with a mountain of salt, it wouldn't be the first time that a company, person, or politician (are they people? ) has interpreted a report or statistics to better serve their needs.

[scaryjim]

Obviously you're free to make up your own mind on how valid AMD's claims are but personally, and going on past experience of AMD's claims, I'm going to take their claim with a mountain of salt, it wouldn't be the first time that a company, person, or politician (are they people? ) has interpreted a report or statistics to better serve their needs.

Unless we get the actual figures from the actual report we'll none of us ever know. OTOH they're quoting a source which means the figures are at least checkable by someone; that's better than a lot of people manage. There'll always be arguments over what is a valid interpretation: you can put whatever qualifications on "VR" you want an spin the figures that way.

I'm not necessarily defending AMD's definition of VR, simply pointing out that the people shouting "rubbish" are simply doing their own interpretation of the statistics: it's no more or less valid than AMDs.

[watercooled]

Erm, has Asus not got a GTX 950 with no power connector and using only 75w max. That won't look as great if AMD compared it to a more comparable product. Wonder what that performance per watt comparison looks like......

As others have said they were measuring wall power consumption, not just the GPU. But far from being advantageous to AMD, this way of testing would actually make the difference in power consumption look smaller than it really is. Once you subtract the rest-of-system power consumption, you'll find the difference (in terms of percentage) actually looks far larger.

[CAT-THE-FIFTH]

Personally i like this interpretation of what AMD are claiming, it seems fairly well balanced and objective to me but what do it know.

The problem is that they have made a very basic error in that article - they say AMD discrete marketshare is 20% and the issue is there are a few problems with that.

What JPR reports each quarter is not installed userbase marketshare but share of shipped units. The actual installed userbase of installed AMD and Nvidia cards is probably nowhere near a 80/20 share.

JPR shipped units is vague on whether laptop dGPUs are included too IIRC ,and if you have been following JPR for the last year,it is AMD laptop dGPU share which has suffered - I suspect AMD share of actual shipped desktop cards is a higher percentage looking at what JPR has said for the various dGPU segments,but even if it is only desktop cards, there is no way of saying what the split is when it comes to the enthusiast segment!!

JPR has said that a total of 9 million have been sold in the last two years,ie,those around £250 and above,but AFAIK have not said what the split is.

They call GameWorksVR a low level API which is again a bit hit and miss. LiquidVR is low level since it is based on Mantle - there is no indication that GameworksVR is anywhere near that level as a whole API.

Then you add the fact that every AMD GPU released after the R9 290 series supported VR specific features a year BEFORE Nvidia did. It took until Maxwell V2 for Nvidia to actually match the same level of VR features as the R9 290 series. The latest GCN GPUs like Tonga and Fiji support another tier of VR features like fine grained pre-emption which AFAIK,are only supported by them.

So,that is another complication to things too.

So then to make up a figure that AMD only has 50% of the addressable market comes a bit across as back of the napkin calculations - since they have based it on faulty userbase measures,ignored the fact AMD has far more non-gaming VR tie ups with companies like Adobe,Associated Press,Creative Artists Agency,etc and educational companies. Basically a whole lot of content creation and teaching companies. Nvidia is nowhere to be seen outside gaming.

Then add the fact AMD,is involved in producing other types of VR headsets too like that AR one,which is addressing a somewhat different market,then that article's numbers don't really add up.

Add 36 million PS4 consoles to the mix too.

So taking JPR figures of 9 million £250 class and above GPU sold over the last two years.

Even if Nvidia had 100% of that market,that would give AMD 80% of the total VR compatible consoles and PCs.

The only problem is in the first year only the R9 290 series were really VR compatible until the GTX970 was sold. Hence not all of those 3 million estimated enthusiast cards sold in 2014 fit the bill.

Of those 9 million cards,a far chunk of them,ie,all those GTX780,GTX780TI and Geforce Black Edition cards are not really considered adequate for VR due to lacking essential features which Maxwell have.

Even if you assume 20% of those 9 million cards were AMD ones,it would give AMD around 84% of the total VR capable PCs and consoles. But even then,a far number of the remaining Nvidia cards like the Kepler ones are below the bare minimum.

JPR has taken all that into consideration I suspect as part of their actual research into the field,and I don't think they have been that unfair to Nvidia in their numbers.

I don't see why there is any argument over what JPR has reported,its almost like people are terrified AMD might have a larger potential userbase for VR,but at the same time would never debate any of the Nvidia figures of how much of the car market or tablet market that Nvidia bumpf also reports every year,and are quite happy to use companies like JPR too!

[Xlucine]

The PS4 fleet is bound to be essential to mass market adoption of VR - it's substantially cheaper than a PC meeting VR recommended specs, and much more common. I also wouldn't be at all surprised to see PS4's turning up in large numbers in corporate VR uses - just like back when PS3's were all the rage for cheap high performance compute farms

[Corky34]

The problem is that they have made a very basic error in that article - they say AMD discrete marketshare is 20% and the issue is there are a few problems with that.

What JPR reports each quarter is not installed userbase marketshare but share of shipped units. The actual installed userbase of installed AMD and Nvidia cards is probably nowhere near a 80/20 share.
<Snip>

Going on the Steam hardware survey that 80/20 split it fairly accurate, either way unless this report can be independently verified all we have to go one are best guesses and what little nuggets AMD may have cherry picked from it, as you yourself said shouting "rubbish" is simply people doing their own interpretation just as AMD have, to take either opinion as facts is wrong IMO, this claim from AMD has no more validity than the people saying it's not true.

[lumireleon]

game developers are more focused on consoles than PC becoz its harder to pirate console games....true or false?

[kalniel]

game developers are more focused on consoles than PC becoz its harder to pirate console games....true or false?

Off topic, but false.