Three months ago SemiAccurate not only told you Nvidia’s Denver was dead, but technically why it was a goner. Last week some others are just figuring out that it is not coming but don’t seem to have a clue about why.
You might recall that SemiAccurate told you all the details of the SoC long ago, August to be exact. August of 2011 that is. Then we told you a bit about Denver itself and it’s long and tortured genesis. The news of it’s almost assured death is not just technically a foregone conclusion, it is a result of blinding mis-management rather than any technical missteps. We hold the engineers on this project in the highest regard, the fact that they made it this far in spite of having the rug yanked out from under them repeatedly is a testament to their abilities. If Denver is indeed gone like all the signs point to, it is a sad, sad day for any enthusiast.
Looking back the signs were obvious, from Nvidia promising the SoC on a process that could not possibly happen in time, to the transparently impossible numbers presented to a room full of press and analysts at CES. We then told you that the SoC mysteriously disappeared from the roadmaps without a trace not even a full quarter later, something some other sites fearful of endangering their ad revenue would not report. Some even went so far as to call the sleight of hand as a win for Nvidia, something we can not fathom.
On the technical side, SemiAccurate told you about the performance of the core itself and the bus situation that made the claims of Nvidia’s CEO at CES quite impossible. In short what they promised in January could not have been the actual plan, technically speaking there was no way it could have worked. Little signs like showing an admittedly faked die shot were red flags that no one seemed to notice. When Parker/Denver/T50 fell off the roadmap in place of Erista during GTC 2014, it should have come as no surprise to the technically aware. That said, what did Nvidia replace the long-suffering Denver core with?
Note: The following is for professional and student level subscribers.
Disclosures: Charlie Demerjian and Stone Arch Networking Services, Inc. have no consulting relationships, investment relationships, or hold any investment positions with any of the companies mentioned in this report.
Latest posts by Charlie Demerjian (see all)
- HyperX ships it’s 60 millionth enthusiast memory module - Oct 15, 2018
- Bittware/Nallatech water cools 300W of Xilinx FPGA - Oct 12, 2018
- More on Intel’s 10nm process problems - Sep 17, 2018
- Intel puts out another 14nm 2020 server platform - Sep 11, 2018
- Why Can’t Intel Supply Enough 14nm Xeons? - Sep 10, 2018