A few weeks ago, a rumor regarding TSMC experiencing issues with NVIDIA’s 28nm manufacturing process for GPUs spread across the Internet. Not long after that rumor, a new rumor emerged that stated AMD would beat NVIDIA to 28nm. More information has now reached us in the form of a quote from AMD’s CFO Thomas Seifert who stated the following during the Q2 2011 Earnings Call:
“We also passed several critical milestones in the second quarter as we prepare our next-generation 28-nanometer graphics family. We have working silicon in-house and remain on track to deliver the first members of what we expect will be another industry-leading GPU family to market later this year. We expect to be at the forefront of the GPU industry’s transition to 28-nanometer.”
With some talk of the new Radeon HD 7xxx series coming out as early as September, the statement that AMD has working silicon is a good sign that they are on track to deliver GPUs on the small 28nm manufacturing process. Whether they can beat NVIDIA to it; however, still remains to be seen. With that said, things are looking good for AMD’s GPU division. What are your thoughts on this, will NVIDIA pull a Hail Mary and get Kepler out of the gate before AMD can ramp up production of the new Radeons?
Nividia’s strategies remains
Nividia’s strategies remains making the biggest best performing chips first and then scale it down to make performance, main stream and then value chips later.
If anything, they will have A LOT of trouble keeping up the launch schedule.
I think they might launch it in 2Q 2012, with availability in 3Q. Of which AMD will be ready with their 2nd gen 28nm chips in 4Q.
28nm I think we need to stick
28nm I think we need to stick two under one die. I thinks this is the next step, but the ram or cache needs to get smaller in size or bigger in volume. More ram than you need is better. Just look at the 580 or 590, when you try to push them across mostly displays in duel or triple gpu config.
AMD may make it by the end of
AMD may make it by the end of September however, from what I have heard this will only be the lower end parts. The high-end may have to wait until next Christmas or 2012.
If it’s the low end, what’s
If it’s the low end, what’s the point? High end customers are willing to pay the big bucks to get a next-gen card that has improvements over their current card. I’d understand if they wanted laptop GPUs with 28nm out early, because laptops are what most people buy these days.
They need to get some balls
They need to get some balls and release a flagship product first and then scale down, not dip their toes in the water to see if its hot or cold.
You may remember that ATI/AMD
You may remember that ATI/AMD didn’t do that before in the past. If look back to the 4xxx series and ATI’s first forray into the 40nm part with the 4770. I can see them releasing a reasonable mid-range product and working the kinks out with that and then going with a full on flagship chip the following quarter or two down the road right when nVidia is about to release it’s next flagship model.
While enthusiasts may want to
While enthusiasts may want to see the “flagship” models released first, financially this isn’t the best approach. The enthusiast market is by far the smallest and in reality, the least profittable for chip makers. The largest profits come from the entry, budget and mid-range markets. The enthusiast market doesn’t see the sales that the other markets do due to pricing. More people are willing to spend the $100-$250 necessary to obtain an entry-level, budget or mid-range card than there are those willing to spend $300 and up for a high-end or enthusiast product.