Telegrams for Guests: Server Engineer

Hello fellow server engineer! I loved your insights into the areas of computing you are working into. Something that I found really interesting was your sudden change in consideration between AMD and Intel from the first time you appeared and episode 25. Is the enterprise industry finally changing their purchasing decisions based on the 10nm issues at Intel and the performance/watt uplift in Rome and Milan? Thanks for the interesting conversations.

Eduardo Villanueva

Can docker enable hardware (processor) independent software deployment? Is enterprise software written in HW independent languages? If not, why not? Is there a trend to go hardware independent?

node55

How would you describe enterprise (large financials) AMD purchase mix vs intel of the last 12 months vs. planned use in the coming 12 months? E.g., trail and test, experiment in trial use cases , dual source just in case, proper dual source, conversion to AMD

Andrew

Hello Mr. Server Guy! Glad to see your return, just have a few questions, in broken silicon 25, you said it takes about 5 years for a software stock to mature for a new product. Has this been accelerated for AMD because if the continual issues Intel has faced (delays, security mitigations etc)? Also how has your 7002p been doing over the past year? I hear various people had issues with virtualization on the EPYC platform.

PotatoesAreLyfe

Do you think Amd with these new gpus are going to get back in the virtual gpu space? Referring to vdi type deployments as that's the field I work in. They have basically left that market to Nvidia. Thanks

You you see ARM servers take a serious chunk of the market within 5 years outside of cloud hyper scalers like AWS that create their own chips?

dj5k

Hey there, As far as I'm aware, Intel has had a much higher server chip market share than AMD for quite some time. As AMD continues to deliver more compelling products in server, do you foresee companies adopting them given the greater inertia in this space compared to the consumer market? If so, over what sort of timescale? Cheers Tom, really appreciate your content and tech-journalism.

Authoritarrr

What wins (if any) do the advancements that Sony has made with the ps5 custom storage controller bring to the server/business side?

Josef Thörn

What does Xilinx bring to the table in the Data Center Market? I see they have "Data Center Acceleration" listed as a solution their website and it seems to center around AI. How prevalent are Xilinx solutions in the data centers? How would AMD's portfolio benefit from the acquisition of this IP?

1) How much of a role do you see ARM chips playing in the server platforms of the future, outside of the Hyperscalers' custom silicon? Will they challenge X86 anytime soon? How does nVidia's pending purchase of ARM change things? 2) Any thoughts on Intel's recent missteps (security issues, manufacturing woes, political infighting etc.)? What (if anything) are they doing well? Can you see them catching up to AMD in the near-medium term? 3) Any thoughts on SiFive/RISC-V?

panic582d

Hello, Why are cpu's and frankly anything of the sort vulnerable to over the air hacking? Why can't every component/avenue of attack, in a CPU, be encrypted with something like an rsa key?

Dominik Koc

What are your thoughts on CCIX vs. CXL from a vendor-agnostic, technical perspective, and what do you think the implications of the "Post-PCIe Bus War" having been won by Intel & CXL are?

Sayonara

Have you seen anything in the pipeline that would make Radeon viable in AI, or Professional development? Almost all the applications I use, all seem to either gain performance with CUDA or outright require CUDA cores to function. I currently work with 3D modeling and Photogrammetry.

Vipast

Has the cost to benefit ratio for converting to AMD happened yet or is buying Intel still more cost-effective?

Vipast

Hello Mr Engineer. I have a two-parter for you: In your last appearance (or one of them, anyway), you mentioned that it takes time to adopt new generations of CPUs in the server market. Would you say that enough time has elapsed since Epyc Rome’s launch for it to be seriously considered as an upgrade to a professional grade or “heavy duty” server? If so, why? And if not, why? I would also like to ask for your thoughts about Zen3 (and the upcoming Epyc Milan chips) and the utility that Zen3’s improvements over Zen2 (latency, unified cache/CCX, etc) could bring to the server space. Thank you very much for your time, and keep up the good work, Tom! 😁

Jake_ Dude_23

A sequence of questions about cloud/remote gaming/software: 1) What are the challenges in making programs/games that can accept and respond well to remote inputs. 2) Are there practicable solutions in the offer, I stress practicable, wishing for fishes does not count. 3) The speed of light given as a constant, what is the best latency to be expected from the middle east and Africa to European and American servers. 4) Are there real, end user applicable, mitigation for excessive server/gateway hops that do not require third party software. 5) Is there any honest internal discussion of the weakness of remote computing, within professional server community?

Woolly

Regarding server and cloud infrastructures, is there anything to comment on with intel’s Optane or micron’s QuantX technologies?