I got to meet him in his office and have dinner with him once. It was an unforgettable and hugely influential experience.
Two fun anecdotes that have never left me:
- He taught me that IPIs (inter-processor interrupts) are inherently and hugely expensive. Knowing this has helped me with architectural choices more times than I can count.
- He quoted (I think from someone else) a rebuttal to the idea that Physics is the reality and math is just theory. It goes something like: Math is the reality that physicists sometimes discover. Love it.
Let me see if I can remember what Arvind said. This was like 20 years ago so I won’t get it exactly right.
He was comparing to two CPUs sharing data with shared memory. In the shared memory case, each CPU stays on the happy path but sometimes has to do some cache coherence, which may stall the CPU, but that’s the worst that can happen, generally.
But if you try to send a message to another CPU, then you’re asking it to raise an interrupt. That’s not the happy path. CPU will have to stop everything to receive the interrupt and then divert execution to some interrupt gate. Arvind’s point was that every impl of this is going to be much much worse than the worst case of cache coherence.
Most processors are designed with very expensive task switching, but I don't think I'd call that "inherent". A processor could implement an interrupt as a forced branch 200 cycles in the future. Then it's the same speed as a function call with callee-save registers. That's still slower than letting cache hardware manage the whole thing in the background, but it should be as little as nanoseconds slower.
Sorry for the naive question - but what is an example of “send a message to another CPU”?
In high level programming languages, isn’t this usually over shared memory as well? Are there constructs to directly “message” a CPU? When might one use it?
There's a fairly recent study looking into the cost of raising an interrupt. TLDR: the cost on conventional systems is quite high; a significant part of the cost can be eliminated by using a custom kernel; even with a custom kernel, there remains a substantial cost.
To my knowledge, the remaining cost could be decreased to approximately the same cost as a branch mispredict, but getting there would require changes to the chip hardware and software stack.
> To my knowledge, the remaining cost could be decreased to approximately the same cost as a branch mispredict, but getting there would require changes to the chip hardware and software stack.
Do it even need to be a misprediction?
If you are completely focused on latency then flushing everything else makes sense. But I would think that if you continue execution for now and put a branch instruction into the queue you'd reduce the cost per interrupt even further.
Mathematical concepts are abstracted so to be beyond universal (ie, another universe will still have the same peano arithmetic). Our universe exhibits some subset of all mathematical properties
This is about 25 yrs. back. As a grad student, I used to stumble into his papers pretty regularly. Later, I'd scan for author name and pay a lot more attention if it was by Arvind (also a few others). I'd always smile after looking at the author name. Growing up in India, your last name doesn't mean much.
> Growing up in India, your last name doesn't mean much.
My experience is the opposite. People often use only the last name, and first name is seldom used. E.g Mr. Kumaraswamy or Ms. Deshmukh in professional context and simply the last name in friend circles.
He probably means northern part of India. Surnames are more emphasised in the South than North.
I think it started when British started asking for surnames in job application in colonial India. South Indiana took their birthplace as surname along with caste name. In North, Singh and Kumar were used as surnames. Caste names became quite popular in late 20th century.
Before British, I am not sure if surnames were common anywhere in India. Would be a lovely study in history.
i think the reason people like arvind stopped using surnames was that, in india, surnames had come to mean far, far too much. now that he's dead he can't protest against mit news foisting a surname on him
Not to gloss over his accomplishments, but does anyone know how his mononym stuck? It’s one thing for an entertainer who’s marketed as such, but quite another for an academic.
When I first met him in 1978 and asked for his last name, he replied "Arvind is my name". He was very deeply against the caste system and felt that using last names (IIRC, which also indicated caste) was something that he refused to do from high school (his high school diploma name is just 'Arvind'). It seemed that this was a hassle his whole life, but one he continued because he felt so strongly.
He said that after 9/11, the US government came down very hard on MIT to force use of last names and that is the first time I ever knew what it was.
Deeply thoughtful, highly caring and warm to all the people he met, he will be deeply missed!
Indian naming traditions vary by region. It gets intricate upon emigration and matching to Western formats. e.g. this distinguished biophysicist also goes by one name:
Indeed. I had a professor in undergrad named Kanapathipillai Thirugnanasambanthan. He might've been from Sri Lanka rather than India proper, though.
He went by "Sam Thiru".
When I first began I asked a friend if he knew anything about this "Professor Thiru". He said "Well, that's not really his name. His real name wouldn't fit in the schedule." :-)
Later, a different professor used him as an example of why it was a bad idea to hard-code the length of name fields.
Great guy, and one of the best instructors I've ever had.
Lots of people from Andhra/Telugu people have many names, and are known by initials or drop them if they come to the US. That sometimes includes optional caste names or titular names that come at the end, with the family name leading (as in China and Japan). My uncle, for example, had a four-word given name, formally introduced his given name in India or in temples as the latter two, and then dropped the first of the two, and swapped family-first for first-family when he came to the states.
For example, if you ever fly into Bangalore or Mumbai, you'll see GVK on their airports. That's for Gunapati Venkata Krishna Reddy, who goes by GVK. Gunapati is the family name, Venkata Krishna is his given name, and Reddy is the caste name.
Other examples are NT Rama Rao, P.V. Narasimha Rao, and my favorite, is a person I know of who went by KRKVNS Sharma.
I had an Indian colleague that only had one name. To fit in with the company she went by a double first name. i.e (I can't remember her Indian name) Jane Jane.
There's a noted chemist named Warren Warren. Washington State had a football player named Samoa Samoa. Hawaii had a great volleyball player named Allen Allen.
I really don't understand why people give names like this. I understand how they happen when you're trying to force a name into a naming convention that isn't the one it comes from, but why start your kid out that way?
I think the naming convention of naming your kid William (as an example) just so they can be called Bill the rest of their life is even more peculiar. Why walk that extra mile?
I think it’s mostly so they can have some degree of choice over what they are called.
You might call them Bill, but they can decide to call themselves Billy, or Will, or Willy, or William - without the hassle of actually changing their name.
There is also a class thing where naming a kid with a short form is seen as being lower class.
Class is absolutely part of it in my experience. The difference between Chuck and Charles, between Chad and Chadwick, Bob and Robert, and so on. The old money folks I’ve known were very particular about that sort of thing, while few others seemed to care too much.
Walla Walla University (formerly college). Really annoying because there's a relatively much better known WWU on the other side of the state. Your ordering would have been a better choice when it renamed.
In publications from colleges/universities, that usually means "graduate class of year". When no institution is specified, it usually means "this institution".
So in this case, it means Divakar (graduate of MIT class of 2001) and Prabhakar (graduate of MIT class of 2004).
To add to your clear response, listing a year without giving any other detail also indicates a Bachelors degree.
If instead they had earned a graduate degree, that would be indicated with some abbreviation for the granting school/college. (I don't know the MIT terms, but at Yale it could be SOM '01, MED '04, LAW '05, etc.)
The best example I know for India is a Dr. Devika Icecreamwala, who is not an ice cream seller but a dermatologist. (It's her married name; she was born Patel, which is about as boring as you can get for an Indian surname.)
That’s a restaurant chain that serves Parsi food afaik! Like a modernized Brittania. I remember that they hired mostly hearing impaired wait staff and have instructions for interacting with them. It’s a really cool concept and the food is good too.
I scanned all responses to see if someone had posted this link! One of the great blog posts about the intersection between computer science and humanity.
Item 40 is the best one. Of course, the majority of items on this list will create problems in society long before they are an issue with computer systems.
When you are the only Indian at CSAIL at that time, everyone knows you by your first name. These days it wouldn't work as Indians are no longer scarce in the US.
Not only are his personal accomplishments amazing, but the number of people he taught and mentored who then went on to do the same with the next generation is immense.
Sad news, I've learned a lot from some of his old course notes and would come across papers coauthored by him. Take care Arvind, I'll write some Bluespec this week in your honor!
I was very fortunate to have worked on Arvind's Monsoon project as a contractor for Motorola. He was the very model of what a professor ought to be: an active and outgoing researcher, an natural educator, a good organizer, and a lively, witty, and gracious man in public. What a loss.
The last time I saw him, at a dinner in Champaign IL around 2005, he told this joke, apropos of MIT's reputation for arrogance:
Someone called a faculty meeting at MIT and said, guys, we have a problem. People say that we're arrogant. We need to come up with some ideas for being less arrogant. Who has an idea?
One of the faculty members stood up and said, I know! I'll teach a course on humility!
The meeting organizer said that's a great idea! A course on humility is exactly what we need! And the meeting was adjourned.
At the end of the semester, a follow-up faculty meeting was called.
The organizer asked, so, what about our course on humility? Let's have a report about it.
The faculty member who taught the course stood up, and said:
I just learned of the passing of Prof. Arvind. He was not only my professor, but also a warm human being who opened his home to a hungry and homesick undergraduate. I will never forget the Diwali at his home where he invited me and my friend among others from the MIT community. After the aarti, he told us stories of his graduate school days where he often longed and dreamt of pakoras from home. He then invited us to feast, and we ate well that night. Dear prof Arvind, in addition to your wisdom, it is your kindness that we will pass on to future generations.
>Life expectancy for men in the US is about 74 to 76 years
that's life expectancy at birth. The life expectancy of a man who is 70 yo today, having already not died of all the causes that harvest young men, is 13+ years more (without accounting for other aspects of their demographic)
It is a little young, the impression is right. Statistics is hard. "Life expectancy" being an average includes all of the reasons to die when you're young, and young deaths have a large effect on the average.
The most frequent age at death is in your mid to late 80s. I'd like to post the statistic on the 25th and 75th percentile of age at death or maybe 40th, 60th, but I can't readily find the data with the amount of effort I'm willing to put in.
Not that naturally - until fairly recently it was bimodal. Fortunately we've dramatically reduced the number of infant and childhood deaths in the last ~100 years.
That's a misleading number, knowing that he was a lauded prof at MIT we can expect that he was significantly wealthier than the median Mass. resident and with that significantly higher life expectancy.
@dang Black band, and change HN post title to the article's current title, "Arvind, longtime MIT professor and prolific computer scientist, dies at 77"?
I got to meet him in his office and have dinner with him once. It was an unforgettable and hugely influential experience.
Two fun anecdotes that have never left me:
- He taught me that IPIs (inter-processor interrupts) are inherently and hugely expensive. Knowing this has helped me with architectural choices more times than I can count.
- He quoted (I think from someone else) a rebuttal to the idea that Physics is the reality and math is just theory. It goes something like: Math is the reality that physicists sometimes discover. Love it.
reply