JPR Q4 graphics marketshare

[n0x1ous]

they are gone, chump change for anyone that wants x86 and GPU IPs.

GPU IP yes - but the x86 license doesn't transfer if AMD is sold.

 

[96Firebird]

On the bright side, I hear that AMD finally clawed back some pro graphics market share.

I was wondering how those price cuts for the pro cards were going to effect AMD's pro market share. Any links to pro market share data?

 

[raghu78]

These numbers underline the fact that AMD does not have a clear strategy in the GPU market. AMD's biggest mistakes were to ignore the notebook discrete GPU market. After Pitcairn they did not have a decent notebook GPU in 2.5 years and Tonga was just not capable to compete with GM204 in terms of power efficiency. Whats more its only available in the iMac with Retina display.

AMD's problem is their GPU designs are not power efficient. GM204 is a breakthrough in GPU power efficiency. AMD can compete with GM204 with the help of HBM and a 3072 sp dedicated gaming chip (stripped of double precision and other features required only in workstation GPUs) . But the fact is AMD still is behind architecturally as when Nvidia gets to HBM by late 2016 or early 2017 AMD will have to address the massive power efficiency gap in architecture.

Nvidia's massive R&D investments are turning this into a no contest and the GPU market now is beginning to tend towards the CPU market. AMD's management were talking that they were protecting R&D but the results are not there. If AMD does not have a top to bottom R9/R7 3xx GPU stack with compeletely ground up new GCN 2.0 redesigns they can kiss the GPU market good bye. There is always a tipping point after which regaining market share becomes impossible. AMD are damn close to that point if not yet already. If Q2 and Q3 2015 does not reverse the market share trend its all over. Rest in peace AMD.

 

[blastingcap]

There is always a tipping point after which regaining market share becomes impossible. AMD are damn close to that point if not yet already. If Q2 and Q3 2015 does not reverse the market share trend its all over. Rest in peace AMD.

Don't get carried away. Comebacks are possible. Mac OS was almost dead in the late 90s and managed to survive just fine. http://www.theregister.co.uk/2014/0...d_laptop_market_share_dips_below_90_per_cent/

In fact Apple found greener pastures than desktop anyway:

 

[raghu78]

Don't get carried away. Comebacks are possible. Mac OS was almost dead in the late 90s and managed to survive just fine. http://www.theregister.co.uk/2014/0...d_laptop_market_share_dips_below_90_per_cent/

In fact Apple found greener pastures than desktop anyway:

How many comebacks like Apple do you hear in even a century. Today Apple is the largest technology company based on their strength in the high end smartphone market. Apple is over dependent on a single product - iPhone. Thats not good in the long run. Their iPad sales are falling though Mac is strong and outgrowing the larger PC market. But the big difference is AMD has never had the level of innovation, product design (aesthetics) and marketing that Apple possess. You only need to look at the AMD R9 290X reference cooler fiasco to know they are bad at product design. AMD's significant successes were K7 Athlon Thunderbird and K8 Athlon 64 Hammer and the initial Opteron (till 2005). From 2006 Intel came swinging back and beat AMD down to pulp where they bleed today and Nvidia is now doing the same in the GPU market. If AMD cannot pull off some impressive GPU products and CPU architectures (Zen/K12) by late 2016 then its all over. Anyway thats my strong view.

 

[RussianSensation]

As Silverforce noted in the past, being late hurts AMD the most. By the time R9 300 series launches, 750Ti will be almost 1 year old and GM204/206 between 6-9 months old. Given AMD's hot, power hungry and loud perception, and most gamers not reading reviews beyond launch reviews, much less searching for specific cool and quiet after-market 200 cards, the typical PC gamer buying new is thinking NV to start with. IMO, even if 300 series wins in every metric, the market is too favourable towards NV. In AMD's case I would have invested millions into low-end and mid-range laptop and desktop 300 cards. If the rumours of AMD only bringing out Fiji XT as the new chips become true, then it would be a major business miscalculation to focus on <5% is the desktop market.

I think some of you guys overestimate the impact of GM204. NV notes that GM204 sold just 1 million in 2.5 months from launch but every 3 months about 12-14 million AIB cards are sold. This is why the graphics card issues at AMD go well beyond Maxwell vs. GCN. It sounds like OEMs also favour NV to start with as well, probably because the volume sales and profits per unit are higher. Thus, OEMs would pick NV even if AMD wins in every metric because OEMs care about what the consumers buy not what's the best product per say. For example if 960 outsold a 290/290X by 3:1 or 5:1, OEMs do not care that a 290 is a 45% better card with double the VRAM. All they care about is volume sales and profits.

Ultimately it sounds like AMD has not convinced the end consumers that its cards are actually worth buying in any segment. OEMs follow the consumer's demands. AMD couldn't even get market share gains with $399 R9 290 vs. $499 780 or with $299 280X over much more expensive 770 2-4GB. This means the biggest thing killing AMD is consumer perception because 770/780 hardly outgunned 280x/290 on perf/watt by anything major and yet NV kept gaining market share. Since it took NV 6-9 months to roll out Fermi and Kepler top-to-bottom but AMD was still unable to gain market share overall, we have real world proof that when AMD wins on perf/watt and price/performance and absolute performance, it still does not get consumer switching to gain market share.

AMD needs to work way harder with OEMs to push their product and change consumer perception.

 

[Alatar]

As Silverforce noted in the past, being late hurts AMD the most. By the time R9 300 series launches, 750Ti will be almost 1 year old and GM204/206 between 6-9 months old.

750Ti is already over 1 year old right now

GM206 launched over 4 months after GM204 so I can't see how at any point in time GM204 and GM206 could both be in the 6-9 months age bracket.

Besides GM204 is already over 5 months old. Let's all remember how probably the most infamous enthusiast hardware delay disaster ever, Fermi, was only late by 6 months compared to its competitor, cypress, and managed a refresh earlier than the competing brand.

 

[Flapdrol1337]

Since it took NV 6-9 months to roll out Fermi and Kepler top-to-bottom but AMD was still unable to gain market share overall, we have real world proof that when AMD wins on perf/watt and price/performance and absolute performance, it still does not get consumer switching to gain market share.

AMD needs to work way harder with OEMs to push their product and change consumer perception.

When fermi was late the 5850 and 5870 were in high demand though, couldn't get a 5850 for msrp until months after launch.

 

[RussianSensation]

Let's all remember how probably the most infamous enthusiast hardware delay disaster ever, Fermi, was only late by 6 months compared to its competitor, cypress, and managed a refresh earlier than the competing brand.

Right but Fermi only launched the more expensive $350 and $500 GTX470/480 cards within nothing in the low-end and mid-range sectors. Look at the launch dates for HD5450/5550/5570/5750/5770/5830. AMD had the entire market to itself for at least 6 months before GTX460 and lower cards dropped and didn't gain market share by much.

Also, your point that NV was able to launch a refresh of GTX400 series is not at all comparable to R9 300 series. GTX580 was just an unlocked GTX480 with some minor changes to the transistor composition and FP16 integer performance/Z-culling. The performance gain was just 18-20% over the 480. R9 390/390X should be a significant departure from R9 290X in terms of design and performance. It would not be correct to recall R9 290X a refresh of 7970 or 390X a refresh of a 290X. That's why those designs are not comparable to NV refreshing Fermi.

----

EVGA's GTX960 that sells for $210 with $4 shipping on Newegg but already has 97% favourable reviews. Some gamers even bought $20 backplates for this card. If you read the comments, a lot of those gamers are upgrading from older NV cards like 9800GTX, GTS250, GTX550Ti, GTX460/460SE, GTX650Ti GTX660, GTX470. Only a few of those had R9 270X or HD7850.

^ Point is certain trends emerge and continue to show themselves if you follow GPU reviews on Newegg:

1) A LOT of NV gamers just upgrade to an NV card. This is an automatic thought process.

2) Most PC gamers don't do their research. Extending from #1, they pick a budget, and buy an NV card that fits that budget.

Ask yourself this if R9 290 was a hypothetical GTX960Ti with 45% more performance and double the VRAM of a 960 for $250-260, would an NV only gamer buy a $210 GTX960? But because R9 290 is an AMD card, it simply doesn't register on the radar. It seems most of these gamers have already decided they are replacing their NV card with yet another NV card and it appears almost irrelevant what the price/performance, perf/watt, absolute performance or VRAM a similarly priced $200-250 AMD brings in comparison to that 960.

When fermi was late the 5850 and 5870 were in high demand though, couldn't get a 5850 for msrp until months after launch.

That's 1 way to look at it another way is AMD failed to meet supply and the result is a shortage of HD5800 cards.

AMD lost market share from HD4850/4870 to R9 290 generation; and AMD pretty much lost market share in HD4000/5000 and 6000 generations (How much more proof is required that perf/watt, price/performance, first mover advantage were not enough for AMD to get customers to switch?)
http://forums.anandtech.com/showpost.php?p=37181068&postcount=218

A lot of gamers in this very thread who find the situation to be dire did not purchase a single AMD card since 2008. I find it very difficult to accept that any brand agnostic PC gamer who wasn't ONLY interested in single GPU performance crown or SLI only found that no AMD card was worth buying starting from HD4850/4870 era. Even when I got my HD4890, it traded blows with a GTX275 but cost just $30 more than a GTX260 216, while 275 was more expensive. When I purchased my HD6950 for $230 and I knew it would unlock to an HD6970, the cheapest GTX570 was $330-340! I actually skipped 5850/5870 but I grabbed each of my 470s for $210.

It's hard to understand today how a brand agnostic PC gamer would buy a $210 GTX960 over a $250-260 R9 290 that offers nearly 50% more performance and double the VRAM. In Canada just $50-60 separates a GTX960 and an R9 290 (and if a gamer is savvy enough he can find a 290 for just $30-40 more than a 960).

The point the average PC gamer doesn't do in-depth research, doesn't read post-launch reviews, doesn't research after-market cards closely. They buy what their friends think is good or what they hear or just keep buying the same brand they trust from 10 years ago - which happens to be NV. I am about to finish my contract working in Central Asia and the market is basically 99% NV. A colleague of mine purchased a GTX470 2 years ago for $400 to play World of Tanks. He doesn't know anything about videocards but he knows he will only buy NV since that worked for him. He picks a budget and just buys an NV card that fits his next upgrade.

Contrast this to a price/performance focused buyer. You cannot determine what card is the best at price/performance without reading reviews, comparing benchmarks and without comparing NV and AMD cards at various price segments. Naturally a gamer focused on price/performance would be way more informed than a typical gamer who upgrades every 2-4 years based on brand and what his/her friends recommend.

I also knew AMD's brand image was terrible when technically savvy PC gamers on our forum still didn't buy AMD cards to mine thousands of dollars that they could have used to buy NV hardware for a long time. Even if I were a hardcore NV user for decades, I would have bought a ton of AMD cards for mining so that I could have free NV upgrades when mining died for years to come. The shocking part is that even with the perk of AMD cards making $, PC gamers on our forum didn't buy AMD cards.

It also doesn't help that the same average North American PC gamers who don't do any research buy at places like FutureShop and BestBuy where AMD cards are often MIA and whatever AMD cards are for sale are grossly overpriced. AMD needs to work closely with retailers and OEMs to get supply and market pricing under control.

 

[3DVagabond]

When fermi was late the 5850 and 5870 were in high demand though, couldn't get a 5850 for msrp until months after launch.

Then they released the 480/70 and instantly started claiming back market share. The 460 came and they were right back on top. AMD had their entire stack out for months and months, but still people waited for nVidia.

 

[RussianSensation]

R390X and its cut-down SKUs is gonna make or break AMD. I have zero doubts about this, because their CPU/APU division is dead. Currently they are sustained by console royalties and for many years before that, by their GPUs. Without a strong GPU presence, they are gone, chump change for anyone that wants x86 and GPU IPs.

R9 390/390X and some cut-down derivatives are not going to break AMD. $350+ cards are less than 30% of the GPU market. GT720, 730, 740, 750, 750Ti, 960 and various 840, 850, 860M, 965M, etc. derivatives is what's wiping out AMD's market share the most. Per TPU, NV announced that about 1 million GM204s were sold in 2.5 months since launch. That's not a lot considering 12-14 million AIB sales a quarter.

Remember AMD's market share was being eroded already when 750/750Ti launched. With 960 the damage will be even more in the next 3 months. 80-85% market share for NV by July 2015 is easily possible.

Also, it's premature to call AMD dead in the CPU/APU space. It takes 3-5 years to design a new CPU architecture and that's what AMD has been working on for Zen slated around Q3 2016. The failure of Zen is way more likely to kill AMD than R9 300 series. Right now R9 200 series doesn't make $ for AMD. Don't forget that AMD won the design for Nintendo's next gen console and PS4/XB1 will last at least until 2019.

Finally, Mubadala invested $10 billion in GloFo in early 2014. GloFo needs AMD to be one of its key customers and since Mubadala invested $10B into GloFo, there is no way Mubadala will allow for AMD to fail unless they will allow GloFo to fail too.

Then they released the 480/70 and instantly started claiming back market share. The 460 came and they were right back on top. AMD had their entire stack out for months and months, but still people waited for nVidia.

I tend to resell most of my GPUs in person. I always ask for my own curiosity what games the PC gamer buying my card plays for which this card is an upgrade, and what his recent GPU history is. Every NV card I sold was a PC gamer who only bought NV. Every AMD card I sold, the buyer said that the reason he got my card was that it offered most value than anything from AMD or NV on the market. This has been true since at least 2002 that I started paying attention to this. Obviously this is just my anecdotal evidence. I remember you mentioned some time ago that in New Zealand NV tends to have higher prices/premiums for similar performance but gamers still prefer NV. Based on my anecdotal evidence even slower NV cards tend to have higher resale value in Canada that faster AMD cards. A used 760/670 would sell for pretty much the same as a 280X/7970Ghz here.

An interesting situation happens when graphics are sold inside products like gaming consoles where the GPU maker's brand is removed completely from the non-savvy gamer. I always laughed that if Apple purchased AMD and put its logo on AMD cards, Apple videocards would gain market share against NV.

 

[Silverforce11]

R9 390/390X and some cut-down derivatives are not going to break AMD. $350+ cards are less than 30% of the GPU market. GT720, 730, 740, 750, 750Ti, 960 and various 840, 850, 860M, 965M, etc. derivatives is what's wiping out AMD's market share the most. Per TPU, NV announced that about 1 million GM204s were sold in 2.5 months since launch. That's not a lot considering 12-14 million AIB sales a quarter.

Remember AMD's market share was being eroded already when 750/750Ti launched. With 960 the damage will be even more in the next 3 months. 80-85% market share for NV by July 2015 is easily possible.

Low volume but high margin parts are responsible for most of the revenue & profits. Thought you knew that by now.

 

[Head1985]

We can only hope, that 380/380x will be complete new SKU and not renamed 290/290x, bucause renamed 290/290x will people never ever buy vs renamed GTX980 for 400dolars.
NV can easily rename GTX980 to 1070 or whatever just like they did it with GTX680/770 and still make tons of money(we all knows GM204 is only mainstream GPU sold by HiGH-End prices)

if 380/380x is renamed 290/290x i think AMD is done and it will be same like in CPU scene vs intel...

 

[RussianSensation]

Low volume but high margin parts are responsible for most of the revenue & profits. Thought you knew that by now.

I don't think it's that simple. The Xbox One APU is a 363mm^2 die, compared to 348mm^2 for the PS4&#8217;s APU, and AMD sells those for $100-110 to MS/Sony. NV sited that there profit margins they desire are not there in consoles (> 50% profit margin). However, AMD sells WAY more Xbox One and PS4 APUs than R9 290 cards. What do you think is the revenue on each R9 290/290X? I doubt it's much more than $130 but the chip is 438mm2.

So far 30 million XB1 and PS4s sold. How many R9 390 sales would AMD need to have to match the revenue and profits of the consoles? I think you overestimate the impact of high-margin low volume cards like GTX780Ti and $550 290X. AMD would be way better off having solid products from $75-450 than to have very poor line-up from $75-450 but the world's best SKUs at $550 and $650.

Also, it's not cost efficient or viable to use R9 390/390X in laptops which means it's even worse if AMD put all of its eggs into 2 SKUs. If they did, they deserve to fail.

We can only hope, that 380/380x will be complete new SKU and not renamed 290/290x, bucause renamed 290/290x will people never ever buy vs renamed GTX980 for 400dolars.
NV can easily rename GTX980 to 1070 or whatever just like they did it with GTX680/770 and still make tons of money(we all knows GM204 is only mainstream GPU sold by HiGH-End prices)

if 380/380x is renamed 290/290x i think AMD is done and it will be same like in CPU scene vs intel...

Pretty much because NV could just lower 970 to $249-259 and the entire stack of R9 370X/380/380X are irrelevant. Also, the re-badging theory doesn't all address how AMD will have new products for laptops. You don't need to know GPUs, but understanding business is enough to know that you simply do not spend hundreds of millions on 2 products that will be priced at $400+ and neglect 95% of the laptop and desktop market by hoping you can coast another 1.5 years to 14nm with 3.5 year old Pitcairn and Tahiti re-badges. R9 285 doesn't sell well at $180, R9 290 at $250 and R9 290X at $300. What would it accomplish to re-badge these 3 cards as R9 370/380/380X and price them at $149, $199 and $249? $179 960 and $259-269 970 will wipe them out and then AMD would have nothing for laptops for yet another 1.5 years? I don't buy this.

 

[nvgpu]

Why would anyone buy an outdated, 3 year old Pitcairn/Tahiti that doesn't support variable refresh rate for gaming, doesn't support 4K H.264 decoding, doesn't support 4K HEVC decoding and pretty much far less power efficient than GM206 GTX 960?

Nvidia released 3 full top to bottom lineups, Fermi, Kepler & Maxwell while AMD can only rebrand Pitcairn and Tahiti not once, but twice, HD8000 and R9 200.

 

[SimianR]

Why would anyone buy an outdated, 3 year old Pitcairn/Tahiti that doesn't support variable refresh rate for gaming, doesn't support 4K H.264 decoding, doesn't support 4K HEVC decoding and pretty much far less power efficient than GM206 GTX 960?

Maybe it's just me, but I don't think that someone buying a $200 video card is probably worrying about 4k content or displays, expensive g-sync monitors, etc. Most people buying these gaming cards are concerned about about gaming performance at 1080P and the R9 280/x is still a good competitor to the 960 in that regard.

 

[Silverforce11]

I don't think it's that simple. The Xbox One APU is a 363mm^2 die, compared to 348mm^2 for the PS4’s APU, and AMD sells those for $100-110 to MS/Sony. NV sited that there profit margins they desire are not there in consoles (> 50% profit margin). However, AMD sells WAY more Xbox One and PS4 APUs than R9 290 cards. What do you think is the revenue on each R9 290/290X? I doubt it's much more than $130 but the chip is 438mm2.

So far 30 million XB1 and PS4s sold. How many R9 390 sales would AMD need to have to match the revenue and profits of the consoles? I think you overestimate the impact of high-margin low volume cards like GTX780Ti and $550 290X. AMD would be way better off having solid products from $75-450 than to have very poor line-up from $75-450 but the world's best SKUs at $550 and $650.

Why did you bother to bring console into this discussion, there's zero relevance in what we're discussing.

It's already acknowledged that consoles are the bulk of AMD's current revenue. Its the only thing keeping them on life support because their CPU division has died for so long and now the GPU division is bleeding fast.

When you brought up numbers on 970/980, over a million sold in a few months is a big deal when you compare potential revenue & profit to the 950ti or 960. Even if they lower parts out-sell the high end stuff 5:1 or 10:1, the margins & revenue on the high end is ridiculous. They are selling a tiny chip for flagship prices.

Even in AMD's own share holder briefing and report in the past show that most of their profits & revenue come from mid-range and above dGPU and not mainstream or low end.

Do not underestimate high-end products on the bottom line, they do generate the bulk of the profits. But they won't help much on marketshare numbers due to volumes.

 

[Pneumothorax]

Don't get carried away. Comebacks are possible. Mac OS was almost dead in the late 90s and managed to survive just fine. http://www.theregister.co.uk/2014/0...d_laptop_market_share_dips_below_90_per_cent/

In fact Apple found greener pastures than desktop anyway:

I don't think Su is anywhere near Jobs in the salesman/magician category. To bring Apple back from the brink took major luck, a super friendly press, and some ingenuity. Unfortunately AMD doesn't seem to have much of the first 2. To really turn things around, the 390x needs to be as good as the Athlon 64 vs. Pentium 4.

 

[monstercameron]

Just needs 1 killer app...

 

[RussianSensation]

Why would anyone buy an outdated, 3 year old Pitcairn/Tahiti that doesn't support variable refresh rate for gaming, doesn't support 4K H.264 decoding, doesn't support 4K HEVC decoding and pretty much far less power efficient than GM206 GTX 960?

Nvidia released 3 full top to bottom lineups, Fermi, Kepler & Maxwell while AMD can only rebrand Pitcairn and Tahiti not once, but twice, HD8000 and R9 200.

Because a 2GB card is obsolete for gaming in 2015 to keep for 2 years? Also, most GPU gamers could care less about 4K decoding because we lived without it for 20 years. A 960 is worthless for 4K anyway. Fact is R9 290 can be purchased for $50 USD or $50-70 CDN more and it annihilates a 960 by up to 50% in games and has sufficient VRAM to last for 2 years. However, given your username, no one is going to take anything you say seriously regarding GPU selection. If you love NV GPUs, you would have paid attention and seen how 8800 GT 256 Mb, 8800 GTS 320mb, GTX470/570/580 1.28-1.5GB became paperweights. 960 is simply an awful product given its specs and price. If it was $149, would be more acceptable.

@ Silverforce11,

The reason I brought up consoles was to show you that Low Volume + High margin is not the only way to make money for AMD. To get back to my point you honestly think $400+ R9 390/390X on the desktop will save AMD? How many of those cards will sell? Historically the breakdown of flagship AMD vs. NV cards above $400 is at least 70% in favour or NV. Even if 390X beats GM200 in everything, it won't matter for sales. We know 390/390X won't because it will be limited to 4GB of VRAM and chances are GM200 will overclock better based on great overclocking of 960/970/980. AMD needs to focus more on $75-400 segments for laptops and desktops. I guess you and I will need to agree to disagree but if were my firm I would make sure there are great replacements for Pitcairn, Tahiti and Hawaii. If I had money left over, only then I would spend it for the $600 card performance crown.

 

[nvgpu]

You're wrong, plain and simple, 4K H.264/HEVC decoding is done fully on the GM206's full fixed function decoder, it's not worthless for 4K video decoding and it will outlive Pitcairn and Tahiti when G-Sync monitors start using ASIC hardware instead of the FPGA currently used.

GTX 960 4GB cards will be out in March and neither Pitcairn nor Tahiti can handle 4K gaming either. And stop comparing to the 290/290X, that card is worthless with all the miners selling off their Hawaii cards once the mining bubble burst back in 2014.

http://www.channelregister.co.uk/2015/02/12/amd_stops_chip_shipments/

Beleaguered chip underdog AMD is not shipping any new gear to channels this quarter as it bites the bullet to clear existing stock swilling around the industry, estimated to have been in excess of $100m.

This is the only reason you can buy Hawaii dirt cheap, because it's not selling at all.

 

[shady28]

Not a great article given the way they keep doing sequential quarter comparisons (as opposed to same qtr of previous year) and mixing every kind of iGPU into the numbers.

Then you see the chart that says Intel has 71% of the GPU market, even though the article itself says over 30% of PCs have dGPUs (and Intel doesn't make a dGPU).

Impossible to separate dGPU / APU for AMD. Then there's the spectre of how much of Nvidias market share is attributable to Tegra series, since they apparently included tablets.

That said one interesting fact came through :

"30.18% of PCs had discrete GPUs, which is down -2.16%."

30% have dGPU, but that # is declining.

 

[MrK6]

I don't think Su is anywhere near Jobs in the salesman/magician category. To bring Apple back from the brink took major luck, a super friendly press, and some ingenuity. Unfortunately AMD doesn't seem to have much of the first 2. To really turn things around, the 390x needs to be as good as the Athlon 64 vs. Pentium 4.

Not singling you out in the least because it's been mentioned by several others, but AMD's problem is not just not having a 390X. Although flagship graphics cards can be used to advertise a new line of cards (which you may be arguing here), they're a sliver of actual shipments. AMD's problems are a lack of a new line-up of cards in over a year, neglecting the mobile space, and poor marketing in general. To the point that mobile is key, notice how Intel is wiping the floor with both AMD and nvidia.

 

[SimianR]

You're wrong, plain and simple, 4K H.264/HEVC decoding is done fully on the GM206's full fixed function decoder, it's not worthless for 4K video decoding and it will outlive Pitcairn and Tahiti when G-Sync monitors start using ASIC hardware instead of the FPGA currently used.

GTX 960 4GB cards will be out in March and neither Pitcairn nor Tahiti can handle 4K gaming either. And stop comparing to the 290/290X, that card is worthless with all the miners selling off their Hawaii cards once the mining bubble burst back in 2014.

http://www.channelregister.co.uk/2015/02/12/amd_stops_chip_shipments/



This is the only reason you can buy Hawaii dirt cheap, because it's not selling at all.

None of what you said takes away from the fact that the 290 is really the best card you can get right now at it's price point. I've also never heard one person list 4K decoding as a feature they were looking for or were concerned about when shopping for a gaming GPU, not that it takes away from it being a bullet point on the 960's box.

 

[RussianSensation]

You're wrong, plain and simple, 4K H.264/HEVC decoding is done fully on the GM206's full fixed function decoder, it's not worthless for 4K video decoding and it will outlive Pitcairn and Tahiti when G-Sync monitors start using ASIC hardware instead of the FPGA currently used.

GTX 960 4GB cards will be out in March and neither Pitcairn nor Tahiti can handle 4K gaming either. And stop comparing to the 290/290X, that card is worthless with all the miners selling off their Hawaii cards once the mining bubble burst back in 2014.

http://www.channelregister.co.uk/2015/02/12/amd_stops_chip_shipments/

This is the only reason you can buy Hawaii dirt cheap, because it's not selling at all.

1. Modern systems such as i7 4790K handle 4K video just fine. Not that it matters because what fool will have a 4K monitor and a $200 960? You keep talking about this as if it's relevant for gamers. Even 970/980 don't have native 4K decoding and they are selling well.

2. The prices I specified are for a new after-market 290 card. It makes no difference that there are many used 290 cards for sale for someone cross-shopping a new $200-250 USD (or $260-300 CDN one).

3. You didn't at all address how 960 is a big compromise in games that use > 2GB of VRAM: Titanfall, Wolfenstein NWO, Shadow of Mordor, FC4, Dead Rising 3, Evolve. That list will keep getting bigger over the next 2 years. Chances are someone on a budget is not going to keep a 960 for just 6 months. You think $50 extra for 50% more performance and double the VRAM makes 290 "worthless" just because some used 290s on sale were mining cards? Ya like I said bring some logical arguments to the table instead of hyping up NV.

Last time I checked a $200 280X beat a 960 by about 15% and had 3GB of VRAM. Just because NV buyers ignore these advantages, doesn't mean they don't matter for brand agnostic gamers. Not even sure why you talked crap about Pitcairn as sub-$150 Pitcairn destroys a 750Ti by 31-44% in games for barely more money. Sounds like you just don't care about performance, which is fine as not everyone does; but blatantly ignoring how 750ti and 960 get dropped by R9 270/270x and 290 is not going to change reality.

If R9 290 was a $250-260 NV GTX960Ti card, 960 would likely not sell many units. How many NV gamers would not pay $50-60 extra for 45-50% more performance and double the VRAM over the 960? Ya right...

Also, even if 960 has 4GB in March, it still gets owned by a 280X and 290 in performance. The article you linked discussing AMD clearing inventory channel is not surprising. They overestimated the demand for cards due to crypto-currency dying on GPUs and they will EOL R9 200 series by Q2 because R9 300 series should replace those cards. It sounds like you don't follow the GPU market closely. NV had 120,000 GTX570 chips to clear which pushed back GTX660's launch:
http://www.kitguru.net/components/g...-gtx570-overstock-delays-gtx660-until-august/

Having excess inventory is not ideal since it means projections exceeded demand but it's far from unusual and doom and gloom as you have implied in your post.