But R9 290/X released nearly a year ago. You think there was no separate team that worked on their successor? You realize AMD and NV have different teams working on different GPUs concurrently. You are assuming the same team which designed Tonga GPU is the one responsible for Fiji and Bermuda XT? You are also assuming when Tonga was designed that AMD focused specifically on performance/watt but in reality it could just be a test bed for:
1) 40% memory bandwidth efficiency through delta color compression
2) Newer video encoding engine
3) Nearly doubling of Tahiti's geometry performance which is a glance at things to come with 390X
Even if AMD doesn't have a response to GM204 in the next 5-6 months, they will in the next 12. Since ATI/AMD's existence, ATI/AMD always had a response; it's just a matter of time. Assuming that if in 5-6 months they show nothing, then well they are SOL is a simplistic viewpoint of how business works.
I am not.
1. Given how few mobile design wins AMD had with HD6000/7000/8000 and R9 200 mobile cards, they have hit rock bottom in that space in 3-4 years or are coming very close to it imo. Why? 7970/8970/R9 290M is basically the same chip and in the mid-range AMD has hardly improved from 8790M. It's total stagnation, meaning that whatever remaining market share AMD has in the notebook space is coming AMD's advantage in price in segments where NV can't compete or chooses not to. The Maxwell chips you mention will just recapture existing wins of Fermi/Kepler at similar price levels but NV is unlikely to lower prices much more as their mobile GPUs tend to be very expensive. I don't see cards like GTX970M or 980M being cheap either giving AMD an ability to undercut them in laptops by $200-400 even if they are inferior in performance/watt and absolute performance.
You also forgot that we are entering the 2nd half of Haswell CPU generation. There tends to be a lot of upgraders when a new generation of Intel CPUs come out but Broadwell and Skylake are not out in volume until Spring/Summer 2015. By that time AMD could have a response to all the chips you mentioned. NV already missed the back to school laptop buying season with most of Maxwell's mobile lineup and Q1/Q2 2015 are historically weak for mobile GPU sales. That gives AMD plenty of time to develop R9 300 series.
2. The doom and gloom in this thread doesn't even take into account that if AMD lost 50% of its desktop GPU market share and 50% of its notebook GPU market share to move both to 20%/20%, it would still survive long enough for a lower node and GCN 3.0.
So really, I am just eating popcorn at this point.
R9 290 after-market sold for $340-380 for months, I mean nearly 4-6 months and people still bought $480-550 780 3-6GB cards. As I said, most of the NV users selling their 570/670/680/780 cards for 970/980 wouldn't even consider a 50% faster R9 390X for $399. They would only buy NV by waiting longer for NV's response.
AMD needs to diversify beyond desktop and notebook graphics since the overall market for the discrete GPUs is declining due to lack of next generation PC games that have forced PC games to upgrade longer, and cheap next generation consoles which are making console gaming more affordable than during PS360 eras.
"While the actual sales figures of discrete graphics cards for desktop PCs are yet to be revealed, DigiTimes web-site reports that shipments in the Q2 2014 were down 30 – 40 per cent compared to the first quarter of this year. According to Jon Peddie Research, total shipments of graphics cards in the Q1 2014 dropped to 14 million units, a decline of 0.8 per cent compared to the same quarter a year ago. In case the drop of demand is so significant, this may be the worst decline in GPU sales in the recent years."
Source
It's no surprise AMD is spending millions on SoC development and APUs. Yes, graphics still matter but year after year it's becoming a distant spec of what the industry used to be. Even NV is moving into automotive and other sectors since graphics is just not cool anymore like 5-10 years ago. Go back to 2000 or even 2007. In 3 years flagship GPUs would be junk. My 7970s are still very fast for 1080P and even 1440P despite being nearly 3 years old. That's why upgrade cycles are extending to 3 years and beyond. That's why AMD and NV now have 6-12 months to respond to each other's new products since it's not as if when a new GPU comes out, 30 million people upgrade. Those days are long gone.
And obviously people have limited funds. When you are faced with a decision to buy a $700-800 Note 4 and/or iPhone 6+ that you will use for many hours every day for leisure and work vs. spending $600-700 on a new GPU to play console ports 40-50% faster, the choice is clear for most consumers. If you read forums online, many people are asking if they should buy a $400 PS4 or a 970 because despite 970 being vastly superior to a PS4, all we get are shoddy console ports like Dead Rising 3 and Watch Dogs. Wake me up when we have next gen PC graphics as a result of which I am actually excited about a GM210 and not upgrading to 970 SLI cuz I am bored. And I bet you a lot of PC gamers feel this way.
^ This is why AMD has plenty of time to respond since while 970 is a killer card, there are still no PC games worth upgrading for if you passed on a 780/780Ti since a lot of PC gamers will still keep waiting for the next wave of PC games to upgrade to because cards like 7950/670/680/7970 are still fast enough for them.
Last edited: