Forum Home | Log in | Signup | My Page | Rules | FAQ | Boards | Members | Search
« POPE 5 Home | Email msg. | Reply to msg. | Post new | Board info. Previous | Home | Next

Re: A little good news about my grand daughter 

By: micro in POPE 5 | Recommend this post (1)
Fri, 31 Jan 20 8:42 PM | 28 view(s)
Boardmark this board | Pope 5 		Msg. 50432 of 62138
(This msg. is a reply to 50424 by Decomposed)
Jump:
Jump to board:
Jump to msg. #

OUTSTANDING!!!!

Yes, I am familiar with Moore's Law and those processors and the tprocesing threads. I suspect Zim is as well.

Congrats to the DAD who taught his son how to be an acheiver and not settle for being just another student....

that is awesome!

Thumbs Up Thumbs Up Thumbs Up Thumbs Up




» You can also:
- - - - -
The above is a reply to the following message:
Re: A little good news about my grand daughter
By: Decomposed in POPE 5
Fri, 31 Jan 20 6:29 PM 		Msg. 50424 of 62138

micro:

Re: "He wants her there on scholarship when she graduates high school"That's awesome! Every college these days has a girl's softball team. (Tell your granddaughter to apply to the ones that actually use girls on the girls' teams.) It's nice too that she's giving college some thought while she's still in 10th grade. Not many kids do that until it's time to apply, and then it's too late to really look into the schools.

Hey, I've got news pertaining to my son, too.

Over the break, Joey procrastinated on signing up for classes and, consequently, wound up taking Computer Architecture from a professor that everyone dreads - not because he's bad but because he's HARD. He's so feared that, even though the class is required for Computer Science students, the prof only had 30 students enroll. The sole other teacher has 125 with another 25 on her wait list. Everyone wants her because she's so much easier

So, Joey was very worried... and more so when he learned about how grading will work in that class and after talking to his buddy who had the professor last semester. He then learned that the instructor's office hours conflict with Joey's other classes, and that there is only one T.A. (The other professor has eight.) Seemingly, the only saving grace was that the T.A. is a very good one, a PhD student who teaches the Computer Architecture labs for both the undergrads and Masters students.

There was no avoiding this prof. Joey had to have the class this semester (it's a prereq for other classes Joey wants to take) and the easier teacher won't take more people.

Less than a month into the semester, all concerns are gone. Joey says he is the top student, hands down.

Out of the blue, Wednesday, Joey received a text from the T.A. asking him if he would be interested in a job. The T.A. had already spoken to the professor, so this is a legitimate job offer. The T.A. and the professor are working together with Intel corp. on new techniques in "Speculative Super-optimization."

You may be familiar with "Moore's Law." Moore's Law states that we can expect a doubling of the transistors on computer CPUs and a resultant doubling of speed and capability every two years. Many don't know that Moore's law ceased to be true a few years ago. Comparably priced computer processors are not likely to speed up at anything close to that rate in the years to come. However, Intel, AMD and other CPU designers have gotten around this by clustering their processors. (You've heard of "Quad-core" processors, I'm sure. That would be 4 clustered processors and is an example of what I'm describing.)

Speculative-Super-optimization is a methodology for breaking up and assigning portions of a larger job to the individual processors in the cluster. While one processor works on the initial portion of the job, a guess is made as to the likely result(s) and the next portion of the project is kicked off by processor #2 before the #1 processor has finished what it's doing. And so on with the other processors. If the guesses are correct, then the whole job completes far faster than if one processor had been tasked with doing the work in linear fashion.

Speculative super-optimization is, therefore, why computers are continuing to get faster. Naturally, the accuracy of the speculation is instrumental in the speed with which jobs can complete. It is key to what Intel is working on these day.

And that's what Joey is going to be doing - while still an undergraduate.

I'm excited for him. This is an opportunity for him to develop leading edge skills few have in an incredibly important field. I think that if he plays his cards right, it could result in good things down the road.

Oh. Joey asked "Why me?" The T.A. said that "in addition to being smart, you are widely respected by people the professor and I respect." Okay. That's cool.

« POPE 5 Home | Email msg. | Reply to msg. | Post new | Board info. Previous | Home | Next