Rant: Why parallel processing will never take off

Miscellaneous Topics
#1

…because the academics who are actually researching it seem like they’ve never programmed anything in their life.

A little extract from the MPI documentation:

int MPI_Scatter ( 
        void *sendbuf, 
        int sendcnt, 
        MPI_Datatype sendtype, 
        void *recvbuf, 
        int recvcnt, 
        MPI_Datatype recvtype, 
        int root, 
        MPI_Comm comm )

Now documentation is scarce to say the least, but when you look at that it suggests sendcnt is the total number of data elements to scatter, and recvcnt is the number of elements received by each process.

Except its not! In fact it actually explicitly states that “send count and receive count must always be identical”. WTF? Heres an 8 param function, which really only accepts 7 values.

:frowning: >:( >:(

#2

Orangy, you are completely wrong.

My wife does parallel processing each day, because she talks on the phone, yells at me, and changes diapers all at the same time.

You are just looking for it in the wrong place.

#3

Academics will never be the hold up for parallel processing systems. The hardware is the limitation.

BTW, the reason graphics processors are so fast is because they are parallel systems.
Whatever you’re examining that is irritating you is not the sum of parallel processing development :slight_smile:

#4

[quote]Whatever you’re examining that is irritating you is not the sum of parallel processing development :slight_smile:
[/quote]
I hope not or we’re all doomed. :wink:

It just somewhat annoys me in that this is a major piece of software / API, produced by university researchers (and publicly released no less, not some internal teaching software), and yet they make glaring, basic mistakes that practically anyone could point out.

#5

Academics. They are living they own isolated lives. Let them be. You aren’t forced to use their work ain’t ya? ~_^

Can you post a link or other clue leading to that group of people that never listened word documentation, or worse, real live programming?

#6

[quote]Academics. They are living they own isolated lives. Let them be. You aren’t forced to use their work ain’t ya? ~_^
[/quote]
Well, except that almost everything we use in graphics programming comes from academic research, as well as programming languages and computers in general. Of course you aren’t forced to use this work, but without a computer, I wonder how you could post in this forum? ::slight_smile:

[quote]Can you post a link or other clue leading to that group of people that never listened word documentation, or worse, real live programming?
[/quote]
You judge from some lines of code to the habits of a whole group?

#7

He did say it was a rant! I’ve read through enough researcher code to have a sympathic ear, but I’d stop short of damning all of || computing.

#8

[quote]He did say it was a rant!
[/quote]
Raghar didn’t say this.

[quote]I’ve read through enough researcher code to have a sympathic ear, but I’d stop short of damning all of || computing.
[/quote]
Yes, that’s true. Computer science developes rapidly and is still a very young science, so most of the older (and some of the younger) researchers have problems to catch up with the current development. Some profs are quite amazed, if you show them how easy and powerful software can be today. And some make mistakes in practise, although they have good theoretical skills.

#9

After 6 years of unemployment period, with no academic education, I have no understanding for academics.

I think somehow, that Ada was somewhat different from current academics. If by nothing else than knowledge about wide array of things. So this sorts of that computer problem.
I created my own language. This sorts the language problem.
And graphic research? If nothing would be created by academics, people that would need to do graphic programming would develop it themselves.

Yes parallel programing is fine, but (most) academics sucks.

[quote]You judge from some lines of code to the habits of a whole group?
[/quote]
Qui, mademoiselle. C’est la vie. What did you expected?

Actually I was curious if any other code is in the spirit of posted one. I wasn’t disappointed. (By viewing HTML part of documentation)
I also don’t try to say it was a crap. I seen worse, but It doesn’t reach a reasonable standard.