Ab jetzt poste ich nur noch in original ohne Übersetzung. Dann kann sich jeder seine Übersetzung selbst machen.
Die originale Verfassung:
ECOSYSTEM! read to whole comment before commenting please. It's all about a software/hardware ecosystem that allows big business to sustain long-term high profit margins. DIY market will never have as much volume nor margins as DC HPCs do in Enterprise accounts. Having said that, If you think that AMD is not going to command an ecosystem, then by all means SELL now and find a better investment. If you think AMD will command an ecosystem then BUY/HOLD because that will take AMD across $100. This is how Intel built it's empire with lousy products, it was only when sandy-bridge was released that Intel actually had a better product but even still IBM's PowerPCs were performance kings. Broadcom is and ecosystem-play, Qualcomm same. Nvidia with CUDA same. AMD had zilch.
HISTORY LESSON: So back in the day Jerry Sanders and his goons didn't care about an ecosystem and were happy to rely on Intel's libraries and x86 platform compatibility and sit back and make good hardware, they always had great engineers because AMD was and still is an Engineering-culture driven company. Once Bulldozer fell out of the sky ( architecture wasn't entirely at fault, it was Global Foundries that failed and delayed as Bulldozer was meant to be an architecture that could hit 7GHz and fit into an APU heterogenous SoC) and when Intel took architecture and process-node advantage with Sandy-Bridge it was game-over. Nvidia in the meantime built an ecosystem around their CUDA development kits. AMD was left to die after being bulldozed by Intel anti-trust practices and the expensive ATI acquisition that became a financial burden for a company that expected to rightfully sell the better performance Athlon cores that was sabotaged by Dell, HP and lenovo. But then along came Dr. Su and Papermaster, both from IBM and they KNOW ecosystem very well. I invested in AMD because I clearly saw what was holding AMD back, and I determined that Dr. Su has the right mind-set to fix all of these, and SHE DID EXACTLY THAT. Move production away from GloFlo, under-promise over-deliver, keep firm to product and tech releases, get rid of personalities like Raja, Scott and others, embrace Papermaster's intuition, no ego, establish trust, Encourage engineers, patience and perseverance and let time heal the wounds while staying true to delivering great products! ONLY after proving great products and solid 5 year unwavering road-maps can AMD claim that they are now fit for an Ecosystem-play. This is starting to happen as we speak and 2020 will be the year of products and solutions that are built on scalable ecosystems, preferably open-source heterogeneous ecosystems. Everyone is welcome to compete in this ecosystem and Dr. Su feels confident that on level playing fields neither Intel nor Nvidia can take her and her team and her ally partners, on.
So there we have it. That's why I'm holding AMD. Folk may hold $intc and $nvda for other reasons. e.g. @Mario aka Dos holds Nvidia because he's a fanboy of a fanboy (his kid) and that was his justification along with his cherry picked analysts that spew garbage. Gotta watch out for those. Yahoo is not the most intellectual board but we have some super smart people like @kris and others that make me smarter every time I read their comments.
Anyways, Zen 3 and BIG Navi is expected next year. I think AMD released PCIE-4 for the reason that they're cooking some big MCM chips for HPC and Gaming, that will rely on this, as well as Infinity-fabric based heterogeneous architectures. AMD has the lead on this tech (MCM + 7nm + respective scalable RDNA architecture), and it's a big ecosystem play imo. 7nm+ is best suited because it will offer 40% better die-size and 10% more efficiency. RDNA is already very efficient but with RDNA2 tweaks we should be able to squeeze multiple GPUs into a package without breaking the power-budget...but the end-result will probably be 40TFLOP GPUs that can do photo-realistic real-time rendering. Combine this with Google Stadia or VR headsets and edge-inference it's a whole new era entertainment and business/productivity use-cases.