Fiji: A Big and Necessary Jump
Fiji has been one of the worst kept secrets in a while. The chip has been talked about, written about, and rumored about seemingly for ages. The chip has promised to take on NVIDIA at the high end by bringing about multiple design decisions that are aimed to give it a tremendous leap in performance and efficiency as compared to previous GCN architectures. NVIDIA released their Maxwell based products last year and added to that this year with the Titan X and the GTX 980 Ti. These are the parts that Fiji is aimed to compete with.
The first product that Fiji will power is the R9 Fury X with integrated water cooling.
AMD has not been standing still, but their R&D budgets have been taking a hit as of late. The workforce has also been pared down to the bare minimum (or so I hope) while still being able to design, market, and sell products to the industry. This has affected their ability to produce as large a quantity of new chips as NVIDIA has in the past year. Cut-backs are likely not the entirety of the story, but they have certainly affected it.
The plan at AMD seems to be to focus on very important products and technologies, and then migrate those technologies to new products and lines when it makes the most sense. Last year we saw the introduction of “Tonga” which was the first major redesign after the release of the GCN 1.1 based Hawaii which powers the R9 290 and R9 390 series. Tonga delivered double the tessellation performance over Hawaii, it improved overall architecture efficiency, and allowed AMD to replace the older Tahiti and Pitcairn chips with an updated unit that featured xDMA and TrueAudio support. Tonga was a necessary building block that allowed AMD to produce a chip like Fiji.
Subject: Processors | September 30, 2014 - 06:02 PM | Josh Walrath
Tagged: arm, cortex, Cortex-A, cortex-m, 90 nm, 40 nm, 28 nm, 32 bit
Last week ARM announced the latest member of their Cortex-M series of embedded parts. The new Cortex-M7 design is a 32 bit processor designed to have good performance while achieving excellent power consumption. The M7 is a fully superscalar design with 6 pipeline stages. This product should not be confused with the Cortex-A series of products, as the M series is aimed directly at embedded markets.
This product is not necessarily meant for multi-media rich applications, so it will not find its way into a modern smart phone. Products that it is leveraged at would be products like the latest generation of smart watches. Industrial control applications, automotive computing, low power and low heat applications, and countless IoT (Internet of Things) products can utilize this architecture.
The designs are being offered on a variety of process nodes from 90 nm down to 28 nm. These choices are made by the licensee depending on the specifics of their application. In the most energy efficient state, ARM claims that these products can see multiple years of running non-stop on a small lithium battery.
This obviously is not the most interesting ARM based product that we have seen lately, but it addresses a very important market. What is perhaps most interesting about this release not only is the pretty dramatic increase in per clock performance from the previous generation of part, but also how robust the support is in terms of design tools, software ecosystem, and 3rd party support.
Cortex-M7 can also be utilized in areas where a more complex DSP has traditionally been used. In comparison to some common DSPs, the Cortex-M7 is competitive in terms of specialized workload performance. It also has the advantage of being much more flexible than a DSP in a general computing environment.
ARM just keeps on moving along with products that address many different computing markets. ARM’s high end Cortex-A series of parts powers the majority of smart phones and tablets while the Cortex-M series have sold in the billions addressing the embedded market. The Cortex-M7 is the latest member of that family and will find more than its fair share of products to be integrated into.
Taking a Fresh Look at GLOBALFOUNDRIES
It has been a while since we last talked about GLOBALFOUNDRIES, and it is high time to do so. So why the long wait between updates? Well, I think the long and short of it is a lack of execution from their stated roadmaps from around 2009 on. When GF first came on the scene they had a very aggressive roadmap about where their process technology will be and how it will be implemented. I believe that GF first mentioned a working 28 nm process in a early 2011 timeframe. There was a lot of excitement in some corners as people expected next generation GPUs to be available around then using that process node.
Fab 1 is the facility where all 32 nm SOI and most 28 nm HKMG are produced.
Obviously GF did not get that particular process up and running as expected. In fact, they had some real issues getting 32 nm SOI running in a timely manner. Llano was the first product GF produced on that particular node, as well as plenty of test wafers of Bulldozer parts. Both were delayed from when they were initially expected to hit, and both had fabrication issues. Time and money can fix most things when it comes to process technology, and eventually GF was able to solve what issues they had on their end. 32 nm SOI/HKMG is producing like gangbusters. AMD has improved their designs on their end to make things a bit easier as well at GF.
While shoring up the 32 nm process was of extreme importance to GF, it seemingly took resources away from further developing 28 nm and below processes. While work was still being done on these products, the roadmap was far too aggressive for what they were able to accomplish. The hits just kept coming though. AMD cut back on 32nm orders, which had a financial impact on both companies. It was cheaper for AMD to renegotiate the contract and take a penalty rather than order chips that it simply could not sell. GF then had lots of line space open on 32 nm SOI (Dresden) that could not be filled. AMD then voided another contract in which they suffered a larger penalty by opting to potentially utilize a second source for 28 nm HKMG production of their CPUs and APUs. AMD obviously was very uncomfortable about where GF was with their 28 nm process.
During all of this time GF was working to get their Luther Forest FAB 8 up and running. Building a new FAB is no small task. This is a multi-billion dollar endeavor and any new FAB design will have complications. Happily for GF, the development of this FAB has gone along seemingly according to plan. The FAB has achieved every major milestone in construction and deployment. Still, the risks involved with a FAB that could reach around $8 billion+ are immense.
2012 was not exactly the year that GF expected, or hoped for. It was tough on them and their partners. They also had more expenses such as acquiring Chartered back in 2009 and then acquiring the rather significant stake that AMD had in the company in the first place. During this time ATIC has been pumping money into GF to keep it afloat as well as its aspirations at being a major player in the fabrication industry.
Get Out the Microscope
AMD announced their Q1 2012 earnings last week, which turned out better than the previous numbers suggested. The bad news is that they posted a net loss of $590 million. That does sound pretty bad considering that their gross revenue was $1.59 billion, but there is more to the story than meets the eye. Of course, there are thoughts of “those spendthrift executives are burying AMD again”, but this is not the case. The loss lays squarely on the GLOBALFOUNDRIES equity and wafer agreements that have totally been retooled.
To get a good idea of where AMD stands in Q1, and for the rest of this year, we need to see how all these numbers actually get sorted out. Gross revenue is down 6% from the quarter before, which is expected due to seasonal pressures. This is right in line with Intel’s seasonal downturn, and in ways AMD was affected slightly less than their larger competitor. They are down around 2% from last year’s quarter, and part of that can be attributed to the continuing hard drive shortage that continued to affect the previous quarter.
Subject: Editorial | March 9, 2012 - 11:45 AM | Josh Walrath
Tagged: TSMC, tahiti, process node, nvidia, kepler, amd, 28 nm
Charlie over at Semiaccurate is reporting that TSMC has closed down their entire 28 nm line. Shut down. Not running wafers. This obviously cannot be good.
Apparently TSMC stopped the entire line about three weeks ago and have not restarted it. This type of thing does not happen very often, and when it does, things are really out of whack. Going back we have heard mixed reviews of TSMC’s 28 nm process. NVIDIA was quoted as saying that yields still were not very good, but at least were better than what they experienced with their first 40 nm part (GTX 400 series). Now, part of NVIDIA’s problem was that the design was as much of an issue as the 40 nm process was. AMD at the time was churning out HD 5000 series parts at a pretty good rate, and they said their yields were within expectations.
AMD so far is one of the first customers out of the gate with a large volume of 28 nm parts. The HD 7900 series has been out since the second week of January, the HD 7700 series since mid-February, and the recently released HD 7800 series will reach market in about 2 weeks. Charlie has done some more digging and has found out that AMD has enough product in terms of finished boards and packaged chips that they will be able to handle the shutdown from TSMC. Things will get tight at the end, but apparently the wafers in the middle of being processed have not been thrown out or destroyed. So once production starts again, AMD and the other customers will not have to wait 16 to 20 weeks before getting finished product.
NVIDIA will likely not fare nearly as well. The bulk of the stoppage occurred during the real “meat and potatoes” manufacturing cycle for the company. NVIDIA expects to launch the first round of Kepler based products this month, but if production has been stopped for the past three weeks then we can bet that there are a lot of NVIDIA wafers just sitting in the middle of production. Charlie also claims that the NVIDIA launch will not be a hard one, and NVIDIA expects retail products to be available several weeks after the introduction.
The potential reasons for this could be legion. Was there some kind of toxic spill that resulted in a massive cleanup that required the entire line to be shut down? Was there some kind of contamination that was present while installing the line, but was not discovered until well after production started? Or was something glossed over during installation that ballooned into a bigger problem that just needed to be rectified (a stitch in time saves nine)?
MIA or Simply Retired?
It is awfully hard to deny the value proposition of the AMD HD 6970 graphics card. The card overall matches (and sometimes exceeds) the NVIDIA GTX 570 at a slightly lower price, it has 2 GB of frame buffer, and AMD is consistently improving not just gaming performance for the new VLIW 4 architecture, but also adding to its GPGPU support. Throw in the extra happiness of a more manageable power draw, pretty low heat production for a top end card, and it is also the fastest single GPU card when it comes to bitcoin mining. With all of these positives, why hasn’t everyone gone out to buy one? Simple, they simply are hard to come by anymore.
¿Dónde están las tarjetas gráficas?
Throughout Winter and Spring of this year, the HD 6970 was an easy card to acquire. Prices were very reasonable, supply seemed ample, and most every manufacturer had one in a configuration that would appeal to a lot of people. The HD 6950 was also in great supply, and it was also in a few unique configurations that adds more for the money than just the reference design. This Summer saw the pool of HD 6970 cards dry up, not to mention the complete lack of HD 6990 cards in retail altogether.