>finally graduated CS >want to do PhD how to do it, where

>>57228063
Good luck on that. Should have went computer engineering instead if that's what you wanted to do.

>>57228104
CE. But good luck. CE makes cs look like a social science.

>>57228179
thank you very much

>>57228074
but all USA schools want CS. US PhD is the best thing you can do for CS.

>>57228179
>>57228074
I'm currently doing a PhD in CS, and thesis is about creating a framework for dynamic IO device sharing between multiple hosts in a PCIe cluster

It's all about the path you choose, there's no such thing as "this is a CE topic only" and "this is a CS topic only"

>>57228063
>how to do it, where to do it?
Find a project you want to do and apply for it. It's easier this way. The alternative is to sit down with someone and create a project proposal and then try to find funding for it, but that's hard.

Alternatively, if you have any industry contacts, you can approach a company and make a deal with them to do an industry PhD.

>>57228063
heeyyy, Parallel Computing is something I want to do
And kinda am doing, Im a second semester undergrad for CSCI who has build a Beowulf cluster with my friend and am working on one for the Physics department at my Uni

>>57228063

I'll give you my advice as a current Ph.D. student.

Find a professor you want to work with. Make sure their research interests align with yours. Contact them and try to schedule a phone call/interview talking about their research. In addition to the topic, you also want to make sure you can get along with the person (since you will be spending a lot of time with him).

Since you are interested in architecture, look at Electrical Engineering/Computer Engineering programs as well as CS.

Make sure that you have funding available before you start. I see this too often, where people start without any funding and spend too much time worrying about supporting themselves through school.

>>57233004
For anyone else, this is relevant to ANY Ph.D program or even undergrad studies.