I have had to work a lot in the past few days(MS/r) I learned quite a few new things. First that there are more then 15 ways of doing binary multiplication. Second there are at-least 10 ways of doing binary exponentiation!! Boy people do have a lot of time on hand :) And the worst thing to happen was that a conference date was advanced by a full 20 days :( Now I have to work even harder !
The Olympics are over and it's fair to say that none of us expected 3 medals for India. It's an outstanding achievement.On he whole just when we thought that it couldn't get any better then Phelps, along came Usain Bolt. WOW is the only word I have to describe him. He ran the 100m as if it was a morning jog.
This was the first time I really followed the Olympics. I was getting used to life at IIIT when the last Olympics happened and was too ignorant in 2000. I think our generation has seen something of a miracle in the performances of Phelps and Bolt. Odds are we won't see anything as spectacular anytime soon. I'm glad I was around to watch Beijing 2008.
PS: I watched only the last 10 mins of Liverpool vs Boro :) Awesome awesome awesome
Monday, August 25, 2008
Friday, August 22, 2008
Where does nVidia go after the 4870x2?
AMD/ATI recently launched the 4870X2. It's basically two 4870's on one physical card. Benchmarks show that it totally outperforms nVidia's flagship GTX280. Which begs the question: what was nVidia thinking when it launched the GTX280? When launched it was almost always outperformed by the 9800GX2. It only beat the 9800GX2 with AA turned on at high res. Not to mention the initial price of $650. This untimely launch by nVidia has seen it lose the crown to ATI after almost two years. The last I think time ATI was on top was with the x1950XT.
In all this mess, does there lie any salvation for nVidia?? Yes there does. CUDA works amazingly fast with the new GT200 core. CUDA based Applications are springing up pretty fast and they work really well. Also 3 way SLI with GTX280 is still the fastest solution commercially available for gaming.
Where does nVidia go from here? Well I don't see nVidia being able to put anything more on the GTX280. What would be a good idea though would be a 9800gx2 2GB Edition. The main problem with the 9800GX2 is with using AA. This is more of a memory problem then anything else. Doing AA @ 1920x1200 or 2560x1600 with 512 MB of memory is not really feasible. ATI realized this as well and put in extra memory on the 4870x2. I think a 2GB 9800GX2 would really put nVidia back in competition with the 4870X2. At present the 1GB 9800GX2 retails as low as $265. Adding a extra GB of memory will push this upto not more then $350 and still keep it to within 5% of the $570 4870X2.
In all this mess, does there lie any salvation for nVidia?? Yes there does. CUDA works amazingly fast with the new GT200 core. CUDA based Applications are springing up pretty fast and they work really well. Also 3 way SLI with GTX280 is still the fastest solution commercially available for gaming.
Where does nVidia go from here? Well I don't see nVidia being able to put anything more on the GTX280. What would be a good idea though would be a 9800gx2 2GB Edition. The main problem with the 9800GX2 is with using AA. This is more of a memory problem then anything else. Doing AA @ 1920x1200 or 2560x1600 with 512 MB of memory is not really feasible. ATI realized this as well and put in extra memory on the 4870x2. I think a 2GB 9800GX2 would really put nVidia back in competition with the 4870X2. At present the 1GB 9800GX2 retails as low as $265. Adding a extra GB of memory will push this upto not more then $350 and still keep it to within 5% of the $570 4870X2.
Monday, August 18, 2008
Why IIIT will not have 50 PhD's anytime soon
There is just one reason why many people don't want to pursue a PhD at IIIT or any other Indian university - Money.
American Universities pay PhD students anywhere between $1500 and $2500 per month. You get paid something like Rs.9000 in India. There is just no comparison. For those wondering why money is that important: A PhD would start at-least at the age of 23 and would take almost 5 years. By the age of 28-29 you would come up against at-least one of these issues if not more.
1. Repayment of Loans
2. Family Obligations
3. Marriage
If you have any of these things to deal with you can't afford to earn the paltry 9K per month that a PhD gives you. You could however, possibly do all of the above while doing a PhD abroad. Just think of it like this. As a PhD student abroad, it would cost you a quarter of your monthly pay to buy a laptop. In India it would take you 5 months pay to buy a laptop. In India, we have the universities, we have the faculty and we have the level of work. What we don't have is a realistic stipend structure. Somehow academics in India think that there is some novelty in Poverty. There is *no* novelty in poverty.
IIIT is a place where things can happen. It just needs some will. Still paying TA's and RA's peanuts and not paying MS students like myself and many others anything, doesn't help. If any faculty member is reading : Fix this !
PS: The "Novelty in Poverty" phrase was originally put forward to me by Jimmy.
American Universities pay PhD students anywhere between $1500 and $2500 per month. You get paid something like Rs.9000 in India. There is just no comparison. For those wondering why money is that important: A PhD would start at-least at the age of 23 and would take almost 5 years. By the age of 28-29 you would come up against at-least one of these issues if not more.
1. Repayment of Loans
2. Family Obligations
3. Marriage
If you have any of these things to deal with you can't afford to earn the paltry 9K per month that a PhD gives you. You could however, possibly do all of the above while doing a PhD abroad. Just think of it like this. As a PhD student abroad, it would cost you a quarter of your monthly pay to buy a laptop. In India it would take you 5 months pay to buy a laptop. In India, we have the universities, we have the faculty and we have the level of work. What we don't have is a realistic stipend structure. Somehow academics in India think that there is some novelty in Poverty. There is *no* novelty in poverty.
IIIT is a place where things can happen. It just needs some will. Still paying TA's and RA's peanuts and not paying MS students like myself and many others anything, doesn't help. If any faculty member is reading : Fix this !
PS: The "Novelty in Poverty" phrase was originally put forward to me by Jimmy.
Subscribe to:
Posts (Atom)