Reply to post: Re: Sigh

AMD is a rounding error on Intel's spreadsheet and that sucks for us all

Nate Amsden

Re: Sigh

Wouldn't of helped. AMD has been doing nothing in the server space for 5 years that is their own damn fault. They must of lost a lot of engineering talent around the time when opteron 6000 launched. They had little roadmap to follow up with after that (at the time as a 6000 customer I was eagerly awaiting more. I still like those chips and still have a bunch in production).

Then after being silent for a while AMD killed the newer chips that customers were anticipating.

Then AMD jumped on the ARM bandwagon for servers. That right there told me they lacked x86 skills in server space.

Then as most expected ARM continues to be stillborn in the datacenter much like intel is stillborn in mobile.

So it seems like AMD re tasked some of their desktop crew along with what was left of their server crew to make Zen.

They release some interesting benchmarks against intel but they rig them. A month or two later intel comes out with a new generation of chip (from what I recall AMD was testing against previous gen) which is much faster. Meanwhile AMDs launch is still months away.

Quite sad to see. I suppose I am over it now. I was very angry at AMD for what they did to themselves back in the opteron 6000 days.

If Zen for server is competitive with a lot of cores then I will deploy it. I really fear that it will not be though based again on those benchmark claims.

Having a CPU that is half the cost doesn't really mean much when the total cost of the server+software stack is more than 30k a pop for the systems I deploy.

I just want more real cores (today we use 18 core intel for vmware and most recently am building 22 core intels for LXC), all on the HP DL380Gen9 platform.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon