During my work on thesis, now I am in need of thread api. My default choice is pthread on C. Why? I know C since bachpan, and pthread seems to be the most matured of all the multithreading things present out there. Though its an overkill to actually understand everything in that posix thread standard. I think it'd be beneficial in long run. so, yeah ima sweating on pthreads...
heres a decent tutor for pthreads:
https://computing.llnl.gov/tutorials/pthreads/
Friday, April 29, 2011
Tuesday, March 8, 2011
ASM on 64 bit platform vs 8086 asm
Over some years ago, I had learned great deal of assembly, essentially the 8086 instruction set.
Part of the reason was, that I used in in my undergrad project.
Now,after like 4-5 yrs I am again trying to see the assembly code(thanks gcc for providing conversion) ,and MAN! that code I do not recognize, movl, movd, movq ,movzbl there are like a thousand variations of mov itself ...
although feels good to see at something you did long ago...
:)
Part of the reason was, that I used in in my undergrad project.
Now,after like 4-5 yrs I am again trying to see the assembly code(thanks gcc for providing conversion) ,and MAN! that code I do not recognize, movl, movd, movq ,movzbl there are like a thousand variations of mov itself ...
although feels good to see at something you did long ago...
:)
My genetic journy (ripoff by natgeo)
It is said, since ancient times, that all brahmins belonging to our subcast are decendants of ancient vedic rishis. These rishis have gotras in their honor. I belong to Gotra of Vatsa (sometimes called Srivatsa).
Now, I had this idea when I saw that natgeo is doing this genographic thing that could trace back my ancestors to a group, defined primarily by some thing called genetic marker. A marker is as they put it, a small mutation that occurs in this case, in the Y chromosome,and because this marker is passed unchanged from a man to his son, its perfect way to form trees such that the person at root is great ancestor of everyone of the leaves.
They lived in between pakistan afganistan border, between rivers of saraswati and sindhu(indus).
Ok, so after shelling out 110$ ,and after waiting for like 3/4 th of a year(ARRGH!) they finally posted my result. So, was it interesting ? Totally. Was it a rip off ? YES TOTALLY.
The thing is, the items that they promise on their webfront ,arent all delivered instead I got only one "online" page, that documented my dna as R1b1 p25(snp present) . And told me that these markers are found in abundance in Western europe and iberian region.I am supposed to be direct decendant of cromagnon.
But what i know : I am direct decendant of great rishis ?
Might it therefore be that cromagnons = great rishis ? [they are btw, always dipicted in coarse clothing and unshaved head tied into a knot, and unshaved beard].
THere is also some classification based on the way they tie the knot, (front,left,right,back).
So, finally I have decided to go further and get classified into one of the possible sub-clads of r1b1.(shelled out another 85$ ).
A quick search reveals that the tribe of Hazara and some people in iran seem to have this varity of r1b1, now I am hoping to be tested as one of them.
If at all, I belong to any other region , then it would for sure mean , that someone got scandalous...damn them whoever that is.
Now, I had this idea when I saw that natgeo is doing this genographic thing that could trace back my ancestors to a group, defined primarily by some thing called genetic marker. A marker is as they put it, a small mutation that occurs in this case, in the Y chromosome,and because this marker is passed unchanged from a man to his son, its perfect way to form trees such that the person at root is great ancestor of everyone of the leaves.
They lived in between pakistan afganistan border, between rivers of saraswati and sindhu(indus).
Ok, so after shelling out 110$ ,and after waiting for like 3/4 th of a year(ARRGH!) they finally posted my result. So, was it interesting ? Totally. Was it a rip off ? YES TOTALLY.
The thing is, the items that they promise on their webfront ,arent all delivered instead I got only one "online" page, that documented my dna as R1b1 p25(snp present) . And told me that these markers are found in abundance in Western europe and iberian region.I am supposed to be direct decendant of cromagnon.
But what i know : I am direct decendant of great rishis ?
Might it therefore be that cromagnons = great rishis ? [they are btw, always dipicted in coarse clothing and unshaved head tied into a knot, and unshaved beard].
THere is also some classification based on the way they tie the knot, (front,left,right,back).
So, finally I have decided to go further and get classified into one of the possible sub-clads of r1b1.(shelled out another 85$ ).
A quick search reveals that the tribe of Hazara and some people in iran seem to have this varity of r1b1, now I am hoping to be tested as one of them.
If at all, I belong to any other region , then it would for sure mean , that someone got scandalous...damn them whoever that is.
Thursday, March 3, 2011
New Scientific language ?
I have been reading quite a lot of journal papers now a days, and I seem to waste most of my time in reading,and understanding what the author is trying to say. Once you understand it, the concept seems very simple. So simple that you are amazed at the time you wasted in trying to figure it out.
I believe this happens due to lack of current languages to "crisply" describe an idea. Languages that describe ideas are either way to bloated (background,premises etc,and finally the new concept), or way to short to intuitively understand stuff (eg. mathematical equations that describe principle. it could be as simple as ohms law , or as complicated as eg. maxwell's equations).
When I try to compare this with what happens in computers, it essentially reminds me of i/o transfer rates. Far far ago, there was a processor 8088 which had 16bit processor but 8bit data bus.
As a result of this, its processing speed would be handicapped by the transfer rate.
Human brain is a much more complicated structure,from my personal experiences, when this happens, ie. theres not enough input from matter at hand to brain, brain tends to occupy itself in other thoughts, which then will slowly kill your original thought, because its the new thought is already in process and does not need any new input from outside world.
If only there werent so many layers on human communication, we could have achieved some sort of efficiency in reading stuff.
I once had read that our ancestors separated from our primordial ape-races due to a mutated gene that caused the muscles of jaw to reduce, giving more space for brain to grow, and this made us capable of producing speech* or a scheme of conveying ideas/events.
Essentially the "overhead" as we call it in computer jargon , is far too much for current languages.
We cannot really distribute 6billion ideas of 6 billion people very quickly unless we try and get some mechanism like this in place.
In other words, we waste most of our capabilities due to a simple limitation that the languages put on us.
I believe this happens due to lack of current languages to "crisply" describe an idea. Languages that describe ideas are either way to bloated (background,premises etc,and finally the new concept), or way to short to intuitively understand stuff (eg. mathematical equations that describe principle. it could be as simple as ohms law , or as complicated as eg. maxwell's equations).
When I try to compare this with what happens in computers, it essentially reminds me of i/o transfer rates. Far far ago, there was a processor 8088 which had 16bit processor but 8bit data bus.
As a result of this, its processing speed would be handicapped by the transfer rate.
Human brain is a much more complicated structure,from my personal experiences, when this happens, ie. theres not enough input from matter at hand to brain, brain tends to occupy itself in other thoughts, which then will slowly kill your original thought, because its the new thought is already in process and does not need any new input from outside world.
If only there werent so many layers on human communication, we could have achieved some sort of efficiency in reading stuff.
I once had read that our ancestors separated from our primordial ape-races due to a mutated gene that caused the muscles of jaw to reduce, giving more space for brain to grow, and this made us capable of producing speech* or a scheme of conveying ideas/events.
Essentially the "overhead" as we call it in computer jargon , is far too much for current languages.
We cannot really distribute 6billion ideas of 6 billion people very quickly unless we try and get some mechanism like this in place.
In other words, we waste most of our capabilities due to a simple limitation that the languages put on us.
Progress iptables, mangling and all
As described in my last post, i have been trying to get a fully working twice-nat with dns-alg setup.
I finally chose lubuntu (LXDE ubuntu ) as the operating system. puppy is b.s ,so is dsl. both have major problems during installation and package updates.
So, finally after wasting like 2 days on getting the right distro, was on track. Then I shifted my focus to selecting the real platform which the experiments should be carried out on. I mailed two profs, and both of them seem to be more leaning towards the DETER testbed , it supports dynamic configuration of network ,measurements etc.
It appears to be a better lead then eg.planet lab.
Lets see, right now, this iptable,netfilter stuff seems incredibly vast, lets see if i get lucky and find a simple tut.
meanwhile ima already hittin on server fault and stackchange websites.
Lets see how it goes, I am planning to complete everything by April mid. So, if I get to defend by say eg. June, then Ima all set damit!.
Only two major things of my life remain ,and they choose where I go, lets see if given a choice as dire as this, am I able to propel myself to achieve it.
Dash!
I finally chose lubuntu (LXDE ubuntu ) as the operating system. puppy is b.s ,so is dsl. both have major problems during installation and package updates.
So, finally after wasting like 2 days on getting the right distro, was on track. Then I shifted my focus to selecting the real platform which the experiments should be carried out on. I mailed two profs, and both of them seem to be more leaning towards the DETER testbed , it supports dynamic configuration of network ,measurements etc.
It appears to be a better lead then eg.planet lab.
Lets see, right now, this iptable,netfilter stuff seems incredibly vast, lets see if i get lucky and find a simple tut.
meanwhile ima already hittin on server fault and stackchange websites.
Lets see how it goes, I am planning to complete everything by April mid. So, if I get to defend by say eg. June, then Ima all set damit!.
Only two major things of my life remain ,and they choose where I go, lets see if given a choice as dire as this, am I able to propel myself to achieve it.
Dash!
Friday, February 25, 2011
socio-personal notes
FUCK this! I am sick of all the marriage invitations that try to present themselves as "hip" by providing some 2 bit banal bull shit about
1. movie /trailers
2. wicket gone/ end of life
and other such analogies, the whole point of doing it was to do something new, and when people repeat the same idea, how is it new...its not.
I am also sick of every other noob, focusing his camera at arbit things and acting like a full professional photographer. One particular picture is to focus at yourself by using a mirror ,and putting it as a dp. WTF!
A quick calculation : rare images in are supposed to be "RARE", to capture them you need to think unlike others. However when you have people who do things because its in vogue , it defeats the purpose. I am not sure they could provide rare images because most of them would try to mimic each other resulting into uniformity rather than rarity.
Some times I am so amazed at level of un-individualistic thoughts represented by some people.Its as if they have no shred of independent thinking instilled in them.
This although seems to be good in a narrow scope of life, its not. Once you dont have any individuality left, you are merely a slave to society (read those who control it).
However, I am not entirely sure if its all that bad.
-
Also, i was just thinking about an interview question that was to count the number of words in a given string .
If scopes are fixed, this is a pretty simple problem, but throw into it unicode, internationalization and bunch of memory allocation schemes and VOILA !
you have got yourself a real tough problem to handle.
Shit that puppy is still installation phase, can we ghost the installations ?
I think this should be trivial, since a file is hdd, we could probably make multiple copies.
Later!
1. movie /trailers
2. wicket gone/ end of life
and other such analogies, the whole point of doing it was to do something new, and when people repeat the same idea, how is it new...its not.
I am also sick of every other noob, focusing his camera at arbit things and acting like a full professional photographer. One particular picture is to focus at yourself by using a mirror ,and putting it as a dp. WTF!
A quick calculation : rare images in are supposed to be "RARE", to capture them you need to think unlike others. However when you have people who do things because its in vogue , it defeats the purpose. I am not sure they could provide rare images because most of them would try to mimic each other resulting into uniformity rather than rarity.
Some times I am so amazed at level of un-individualistic thoughts represented by some people.Its as if they have no shred of independent thinking instilled in them.
This although seems to be good in a narrow scope of life, its not. Once you dont have any individuality left, you are merely a slave to society (read those who control it).
However, I am not entirely sure if its all that bad.
-
Also, i was just thinking about an interview question that was to count the number of words in a given string .
If scopes are fixed, this is a pretty simple problem, but throw into it unicode, internationalization and bunch of memory allocation schemes and VOILA !
you have got yourself a real tough problem to handle.
Shit that puppy is still installation phase, can we ghost the installations ?
I think this should be trivial, since a file is hdd, we could probably make multiple copies.
Later!
Internet simulation on your laptop and your single IP ISP
I have planned to create a simple simulation that will allow me to actually get the feel of implementing the SI framework on planet lab before hand, the main work for my thesis.Its very late ,but i intend to deliver a solid experiment and analysis , no matter what!Because this is one of those parts of masters, that really makes the course different than undergrad courses.
Behold I am doing everything from my laptop . THANK YOU SO MUCH AGAIN, virtualization , the best thing that ever happened in CS.
Above image describes a rough drawing of how things will work.The hp laptop would host all the virtual network described by the devices inside the circle.
One of the virtual machines would possess 2 nics , one for internal networking and another for external.
I am going to let it be bridged nic, so that i can use ip aliasing to get all the ips i need to simulate pooling of ips.
One machine on the virtual network will act as a dns server, and another as a world wide web server.
When the setup is complete , the expected result is I would try to access a www server of virtual domain from my hp laptop. Hopefully , it should return a different ip each time i try to hit the website with a different ip.
And this is how , ladies and gentlemen , I propose you could do awfull network simulations with minimal hardware.
The only downlfall is this: I cannot just move my laptop to another location and start doing experimentation. This is because, my virtual router (one machine with 2 nics) would require the same changes that my pc would, and I am not sure till now how can I do that in case of virtual box.
One awwwwwwwfully lame experience is installing operating systems using virtual box. It takes billion years to install, and you need to get a good distro so that your machine wont be overloaded. I used two candidates damn small linux, and puppy linux.
DSL is faster ,based on knoppix and all, but it sucks at package mgmt.
puppy is a bit slower than DSL, but apparently its better at packet mgmt.
I choose puppy cause need to install stuff for admins eg. dhcp scripts and all...
If your a new guy in masters, i suggest you to go for a big processor and memory, you will need it because you will more or less have to use virtualization.
Bye till later
Behold I am doing everything from my laptop . THANK YOU SO MUCH AGAIN, virtualization , the best thing that ever happened in CS.
Above image describes a rough drawing of how things will work.The hp laptop would host all the virtual network described by the devices inside the circle.
One of the virtual machines would possess 2 nics , one for internal networking and another for external.
I am going to let it be bridged nic, so that i can use ip aliasing to get all the ips i need to simulate pooling of ips.
One machine on the virtual network will act as a dns server, and another as a world wide web server.
When the setup is complete , the expected result is I would try to access a www server of virtual domain from my hp laptop. Hopefully , it should return a different ip each time i try to hit the website with a different ip.
And this is how , ladies and gentlemen , I propose you could do awfull network simulations with minimal hardware.
The only downlfall is this: I cannot just move my laptop to another location and start doing experimentation. This is because, my virtual router (one machine with 2 nics) would require the same changes that my pc would, and I am not sure till now how can I do that in case of virtual box.
One awwwwwwwfully lame experience is installing operating systems using virtual box. It takes billion years to install, and you need to get a good distro so that your machine wont be overloaded. I used two candidates damn small linux, and puppy linux.
DSL is faster ,based on knoppix and all, but it sucks at package mgmt.
puppy is a bit slower than DSL, but apparently its better at packet mgmt.
I choose puppy cause need to install stuff for admins eg. dhcp scripts and all...
If your a new guy in masters, i suggest you to go for a big processor and memory, you will need it because you will more or less have to use virtualization.
Bye till later
Subscribe to:
Posts (Atom)