Intel manufacturing woes to lead to increased market share for AMD

[Underseer]

https://marketrealist.com/2018/09/amd-stock-rises-on-intels-troubles

Although the rate at which AMD will gain marketshare now looks to be slower than originally anticipated

https://www.marketwatch.com/story/a...l-street-is-expecting-says-keybanc-2018-09-25

This is good news for AMD, which just released the second generation Ryzen CPUs. At the moment, the 2700 and 2700x are the only CPUs that deal with both Meltdown and Spectre at the hardware level. Granted, that's mostly a concern to people who would actually use the larger number of cores Ryzen offers, while Intel's current offerings do better at tasks like gaming, which at the moment are not as aggressively multithreaded as some other apps (Civ 6 excepted).

Anyway, one way or the other, Intel's inability to manufacture as many CPUs as previously expected will ultimately lead to more sales for AMD, and more pre-built systems being offered with AMD CPUs in them.

 

[phands]

I'm happy that AMD is getting a chance to be a second player again.

For over 10 years now, I've been building massive data centers for cloud computing. It was always a concern to have all our eggs in the Intel basket. Even though we bought from multiple suppliers like Dell, HP and Supermicro, they all ran basically the same Intel chipsets. Diverse data center locations. Diverse network paths. Diverse power and cooling systems. Diverse cable and fiber suppliers. Diverse RAM. Diverse storage. Intel CPUs.

And when you have more than a million multi-CPU servers, and they all get Meltdown or similar, that's a lot of maintenance logistics.

 

[bilby]

I'm happy that AMD is getting a chance to be a second player again.

For over 10 years now, I've been building massive data centers for cloud computing. It was always a concern to have all our eggs in the Intel basket. Even though we bought from multiple suppliers like Dell, HP and Supermicro, they all ran basically the same Intel chipsets. Diverse data center locations. Diverse network paths. Diverse power and cooling systems. Diverse cable and fiber suppliers. Diverse RAM. Diverse storage. Intel CPUs.

And when you have more than a million multi-CPU servers, and they all get Meltdown or similar, that's a lot of maintenance logistics.

Yes, a single point of failure is a big risk to any system where you need five nines of uptime.

(Unless it's a Microsoft environment, where five nines means that it's up for nine seconds every minute, nine minutes an hour, nine hours a day, nine days a month, nine months of the year).

 

[phands]

Yeah....Azure sucks dead bears out loud.

Interesting snippet...Salesforce spent 2 years trying to get their product running on Azure - I worked there at the time - and failed...it just wouldn't run. Once that was over, they tried AWS. Up and running just great in 3 months. M$ sucks.

I think ARM will be the next big name in servers at scale, because it's not an x86 architecture, and it is low power by design unlike Intel OR AMD.

 

[phands]

And now it seems as if AMD might shoot up to a 30% market share....

IT TURNS OUT that making decent processors that can challenge Intel has unsurprisingly worked out well for AMD.


The chipmaker is set to snaffle 30 per cent of the desktop CPU market, according to Taiwan's DigiTimes, which talked to sources familiar with AMD's upcoming fourth quarter 2018 results.

https://www.theinquirer.net/inquire...et-to-grab-30-of-the-desktop-processor-market