|On September 1, GlobalFoundries played host to the first annual Global Technology Conference (GTC for short). While there were a fair number of partner presentations on display, the conference was primarily driven by GlobalFoundries executives and announcements. When we spoke to GlobalFoundries in March we remarked on the company's aggressive roadmap; we were curious to see if the company would still be on track six months later.
Based on comments made by company CEO Doug Grose, GlobalFoundries is on track to realize some $4 billion in revenue for the year 2010 and plans to double its size within two years (it's unclear if this refers to revenue, wafer starts, or customer base). The company's long-term plans for Charter semiconductor have been clarified—despite the age of certain Chartered factories, many of them are well-tuned and optimized for their existing production lines. GF intends to ramp different nodes and technologies at its three international locations, as detailed below:
Chartered is capable of up to 210,000 200mm wafer starts per month (split across five fabs) and 50,000 300mm / 65nm wafer starts per month. These process nodes might seem old to PC enthusiasts, but they still account for an enormous amount of revenue—in Q2 2010, the majority of TSMC's revenue (55 percent) came from its 150+ nm nodes (28 percent) and its 65nm/55nm node (27 percent). At present, Chartered's Fab 7 produces some 40nm tech; this capacity will be transferred to Dresden. Fab 7 will be dedicated to 65/55nm production on 300mm wafers, and will ramp up to the 50,000 WPM (wafers per month) we mentioned earlier.
At Dresden, the fabs GlobalFoundries inherited from AMD are both slated for overhauls and expansions of their own. GlobalFoundries intends to boost the total output of the two fabs up to 80,000 WPM, up from an original 50K. Both facilities will be dedicated to sub-45nm production with a near-term focus on scaling both 32nm SOI and 28nm bulk silicon. Company executives dropped hints that there's some very early work on 22/20nm technology also being done at Dresden, but the bulk of that production will likely be handled at Fab 8 in New York.
Speaking of Fab 8, the foundry's product target has nearly doubled from 35,000 to 60,000 WPM. This additional capacity carries a cost; Fab 8 was originally expected to ship wafers for revenue in 2011 with full production in 2012. Now those dates have been pushed back a year, shipments will begin in 2012, with full production in 2013. Fab 8 will still launch using a 28nm process and 300nm wafers, but we expect Fab 8 to lead the company in 20/22nm production as those processes come online.
The company's general roadmap through 2010.
GlobalFoundries' long term goal is to provide potential customers with an array of processing nodes and technologies while simultaneously offering cross-foundry standardization. The Common Platform Alliance isn't new, but it was a major topic of conversation at GTC 2010. On the manufacturing side of things, the CPA is a joint agreement between Samsung, IBM, and GlobalFoundries. The three companies have agreed to deploy compatible bulk CMOS technology at their various facilities and to collaborate on semiconductor research.
According to GF, foundry customers benefit from this relationship in several ways. First, it offers synchronized production at multiple locations, thus smoothing out the capacity restrictions that might occur at any single fab. Second, it allows vendors to build a single design in multiple locations. This isn't the norm—if Company X wants to change from UMC to TSMC, its microcontroller will likely need to be re-optimized. This is true even if the process technology and wafer size are held constant. In the wake of TSMC's 40nm woes, customers are evidently interested in exploring more flexible manufacturing options, not to mention contract terms that don't leave them at the mercy of a single foundry.
|Ties That Bind: GlobalFoundries, Bulldozer, (And Bobcat?)|
|Now that we've discussed GlobalFoundries' plans in general, let's take a gander at what we can expect from the company's 32nm SOI and 28nm bulk silicon processes. Both processes will use what's known as high-k metal gate technology. High-k can be built either through the use of a gate-first or a gate-last approach. Intel—the only company to date that's shipped high-performance / high volume products using high-k—favors the gate-last approach. GlobalFoundries, IBM, and Samsung are championing gate-first technology.
Here's how the nerdfight breaks down. Proponents of gate-first technology argue that it allows for less-restrictive designs, smaller dies, and is easier to build. It's also cheaper—GF expects to save ~$75 million over four years by opting for gate-first rather than gate-last. Gate-last proponents (Intel, TSMC) claim that gate-first technology isn't a long-term solution and can't deliver all the benefits of the gate-last approach. We won't actually know which approach makes more sense until we see shipping silicon.
What we do know is that the shift to high-k metal gate manufacturing has significant implications for Bulldozer and, by extension, Bobcat. If you remember Intel's Prescott, you should also remember that the CPU's high temperatures and significant power consumption made it a lose/lose proposition for enthusiasts almost across the board. One of the reasons for this was the amount of current that 'leaked' out of the CPU—at just 1.2nm thick on a 90nm process, the gate dielectric material was unable to prevent electron tunneling.
Prior to 90nm, both Intel and AMD shrank gate thickness by ~0.7x per process, right up to the point where gate leakage, which was already present at 130nm, skyrocketed. Intel didn't change gate thickness at 65nm, but was able to do so again at 45nm because switching from Poly/SiON to high-k reduced the amount of gate leakage exponentially. AMD's processors up to and including the Phenom II, have yet to make this switch. Here's GlobalFoundries own prediction for gate scaling at 32nm and below.
AMD's curve looks much the same as Intel's, with the exception that AMD held gate lengths constant through 90nm, 65nm, and 45nm. At 32/28nm, however, power leakage will drop while gates again scale downwards. Over the past few weeks, we've wondered how Bulldozer's power consumption and performance will compare to Sandy Bridge. We're nowhere near filling in that particular puzzle, but we do know, at least, that Bulldozer will be fighting with an advantage previous generations of AMD chips haven't had.
When Will Bobcat Make The Leap?
AMD's first generation of Ontario processors (dual-core Bobcat + GPU) will be built by TSMC on that company's 40G (high performance) process. Left unanswered is the question of whether or not AMD will stick to TSMC at the 28nm node or move to GlobalFoundries. The peculiarities of 28nm technology make this a more important question than it might have been at previous nodes. TSMC currently plans to launch two 28nm processes. One will use high-k metal gate and a gate-last approach while the other is built around conventional Poly/SiON. There are questions regarding the suitability of Poly/SiON at 28nm and TSMC has, as we've noted, opted for the more difficult gate-last approach.
Even if we assume that TSMC executes both approaches perfectly, the inherent differences between gate-first and gate-last mean that there's no simple way to jump from one to the other. This difference will also apply to Radeon chips; switching from one foundry to the other at or below 28nm will necessitate significant and time consuming design/material changes.
AMD may still be keeping its options open and watching to see how the two approaches play out in the real world, but we'd be surprised if the company doesn't opt for one or the other before the end of the year.
|We saw slides and details on GlobalFoundries roadmap that we aren't allowed to post at present; what we can tell you is that the foundry's manufacturing timetable remains aggressive through the next 4-5 years. GF has already adopted double patterning for 32/28nm production and is researching extreme ultraviolet lithography (EUV). There are certain problems with EUV that have yet to be licked, but there's time enough to address the problem—double-patterning should be effective down to 20nm technology and it's not the only multi-pattern option.
GlobalFoundries isn't talking much about catching up to Intel (at least not yet), but if the company remains on schedule the gap between the two should slowly start to narrow. Regardless of how much that particular gap changes, GlobalFoundries is going to rattle the balance of power in the worldwide foundry industry over the next 12 months. It's been years since TSMC faced much competition for first-source designs—we'll see if it remembers how to fight.