While it its old, it adresses the issues associated with current system design. One thing people are overlooking is that the basic architecture of pc's (non supercomputers) has remained relatively static for quite some time. Perhaps in order to see a radical change in operating system design we also need to rethink the hardware architectures that we have used for soo long.
among other things, designing next generation graphics cards is a serious exercise in computer architecture, vlsi design, and algorithm development; these people arent just system integrators or product engineers...next generation stuff has to come from somewhere other than a reference design...these people are absolutely scientists.
you dont need a beaker and a lab coat to be considered a scientist.
you are right. i apologize, i was too focused on two independant variables that pointed to the same data that i never bothered to consider the case of passing the same variable to both arguments of the function.
now that i think about it, not only is your swap function fundamentally flawed, but your understanding of pointer casting and type representation is as well.
you are casting a pointer to an int (allocated 4 bytes) as a pointer to an unsigned long (allocated 8 bytes) without considering any type casting that must occur. you then, through your swap funciton perform an operation on the casted int pointer as if it were allocated as an unsigned long, no wonder your numerical answer does not even work out correctly, the bottom 4 bytes of the unsigned long are undefined.
hrm, from this snip; it appears that you are passing by reference then performing the swap operation directly on the data and not the pointers to the data.
try this:
void swap(unsigned long **a, unsinged long **b){
*a = *a^*b;
*b = *a^*b;
*c = *a^*b; }
i was of the impression that the variable swap was an operation on pointers to objects, so it is an integer operation that in no way manipulates the data referenced.
its a simple 3 instruction operation that is done competely within the registers of the processor, i fail to see how it would not work under any situation assuming you are, in fact, manipulating pointers.
hehe, i was in error replying to an AC post. i am fully aware of the associativity of the xor operator. the (score: 0) makes it look like my post is attempting to disprove the grandparent...confusing...quite.
You also need to make sure that A and B are distinct, not pointers to the same value, otherwise you just get zero out.
it was in response to this that i posted a simple counter example.
i'm confused, if you are so self aware then you must realize that publicly declaring your superiority is pure masturbation...what good is being superior if nobody is willing to acknowledge you as being superior.
there are an infinite number of ways of going about life. your way is no better than mine anybody elses.
just be glad you are here to experinece it any way you choose and leave it at that.
video processor (gpu) die sizes are huge producing a lower yield from each wafer; that yield is then lowered even more after discarding the defective dies. then look at the total volume in market, while its a large number, its much smaller in comparison to total cpu volumes.
throw in design and manufacturing and additional components and your price for realizing a profit are rather high.
then factor in the niche demand and the knowledge that people within this group have, in the past, proven that they will buy it at an insane price initially...wa-la! high prices.
i'm not gonna try and argure your point only answer you on the CAD software questions...
IC Layout: MAGIC Simulation: BSIM Technology models and libraries: MOSIS Verilog: Icarus Verilog
It is possible given these tools to head down the right path to homebrewing a cpu given the level of intelligence and passion is there...however, without corperate backing there are two huge hurdles to overcome...a free(beer) program to translate an HDL(hardware description language) to a gate level description, and an auto place and route program to help layout the synthesized gate level description. But with a sound design, working verilog and spice simulations of various blocks, and the energy to show off and shop your design around, you could easily find financial backing to meet you halfway and finish the project.
maybe she didnt have all the tools, but you can definately learn ic design on your own...you just have to be highly motivated, very patient, and intelligent.
You raise a good point; assuming your assumption that the company that screws its employees actually does manage to become more competitive.
Most people really do believe that working harder really is better than working smarter; however after spending the last 2 years working for a Japanese automotive supplier where the average work day is around 25 to 40% longer than that of most German suppliers, I fail to see the vast disparity between our quality/price and theirs.
Simply put; they stay competitive, meaning they know how to make more out of an 8 hour day than we do during a 10-12 hour day.
People who are passionate about their work and love what they do aren't working, they're doing what they love. It just goes to show that while money can motivate people, passion for the work is a better motivator.
You are absolutely right passion for work does motivate better and innovation will most certainly follow...but, as much as I'd love to live in an idealistic bubble, the sad truth is that there are simply not enough people in this world that are wired that way...through simple evolution the majority of the people in the world are passionate about procreation, consumption, and 300 channels of mind numbing crap.
So until the 97% out there bumps their collective heads on the toilet and have some life altering epiphany...the few will have to work their asses off to lead the many...monitary motivation is just one of the easier methods.
I dont want to sound like I know what I'm talking about, but aren't what you guys are refering to as analog clocks really digital? I mean, dont all clocks that dont use a crystal oscillator use gears to quantize the movement of a winding or weights? Cause that sounds digital to me.
Good god!!! Its about time...I thought I was going to get all the way to the bottom of these postings and not have seen at least one Monty Python reference...c'mon geeks...you're better than that!
"MiraVision can even be both things at once. Anyone wanting to watch TV while brushing their hair or fixing their make-up can switch to "picture in mirror" mode, which displays a small image in the corner of the frame..."
Slashdot Mirror
While it its old, it adresses the issues associated with current system design. One thing people are overlooking is that the basic architecture of pc's (non supercomputers) has remained relatively static for quite some time. Perhaps in order to see a radical change in operating system design we also need to rethink the hardware architectures that we have used for soo long.
dude.
among other things, designing next generation graphics cards is a serious exercise in computer architecture, vlsi design, and algorithm development; these people arent just system integrators or product engineers...next generation stuff has to come from somewhere other than a reference design...these people are absolutely scientists.
you dont need a beaker and a lab coat to be considered a scientist.
dude.
"i built a shelf"
dude.
you are right. i apologize, i was too focused on two independant variables that pointed to the same data that i never bothered to consider the case of passing the same variable to both arguments of the function.
dude.
ok,nevermind what i just said, i understand what you are doing...i'm still not convinced you are right but your casting does make sense to me.
dude.
now that i think about it, not only is your swap function fundamentally flawed, but your understanding of pointer casting and type representation is as well.
you are casting a pointer to an int (allocated 4 bytes) as a pointer to an unsigned long (allocated 8 bytes) without considering any type casting that must occur. you then, through your swap funciton perform an operation on the casted int pointer as if it were allocated as an unsigned long, no wonder your numerical answer does not even work out correctly, the bottom 4 bytes of the unsigned long are undefined.
dude.
shit...the *c should have been a *a.
void swap(unsigned long **a, unsinged long **b){
*a = *a^*b;
*b = *a^*b;
*a = *a^*b;
}
hrm, from this snip; it appears that you are passing by reference then performing the swap operation directly on the data and not the pointers to the data.
try this:
void swap(unsigned long **a, unsinged long **b){
*a = *a^*b;
*b = *a^*b;
*c = *a^*b;
}
dude.
i was of the impression that the variable swap was an operation on pointers to objects, so it is an integer operation that in no way manipulates the data referenced.
its a simple 3 instruction operation that is done competely within the registers of the processor, i fail to see how it would not work under any situation assuming you are, in fact, manipulating pointers.
dude.
hehe, i was in error replying to an AC post. i am fully aware of the associativity of the xor operator. the (score: 0) makes it look like my post is attempting to disprove the grandparent...confusing...quite.
You also need to make sure that A and B are distinct, not pointers to the same value, otherwise you just get zero out.
it was in response to this that i posted a simple counter example.
dude.
a=0110
b=0110
a= 0110 xor 0110 = 0000
b= 0000 xor 0110 = 0110
a= 0000 xor 0110 = 0110
maybe you should rethink your assumption.
dude.
i'm confused, if you are so self aware then you must realize that publicly declaring your superiority is pure masturbation...what good is being superior if nobody is willing to acknowledge you as being superior.
there are an infinite number of ways of going about life. your way is no better than mine anybody elses.
just be glad you are here to experinece it any way you choose and leave it at that.
dude.
video processor (gpu) die sizes are huge producing a lower yield from each wafer; that yield is then lowered even more after discarding the defective dies. then look at the total volume in market, while its a large number, its much smaller in comparison to total cpu volumes.
throw in design and manufacturing and additional components and your price for realizing a profit are rather high.
then factor in the niche demand and the knowledge that people within this group have, in the past, proven that they will buy it at an insane price initially...wa-la! high prices.
i suck at business, but thats my watered down 2c.
dude.
i'm not gonna try and argure your point only answer you on the CAD software questions...
IC Layout: MAGIC
Simulation: BSIM
Technology models and libraries: MOSIS
Verilog: Icarus Verilog
It is possible given these tools to head down the right path to homebrewing a cpu given the level of intelligence and passion is there...however, without corperate backing there are two huge hurdles to overcome...a free(beer) program to translate an HDL(hardware description language) to a gate level description, and an auto place and route program to help layout the synthesized gate level description. But with a sound design, working verilog and spice simulations of various blocks, and the energy to show off and shop your design around, you could easily find financial backing to meet you halfway and finish the project.
maybe she didnt have all the tools, but you can definately learn ic design on your own...you just have to be highly motivated, very patient, and intelligent.
dude.
You raise a good point; assuming your assumption that the company that screws its employees actually does manage to become more competitive.
Most people really do believe that working harder really is better than working smarter; however after spending the last 2 years working for a Japanese automotive supplier where the average work day is around 25 to 40% longer than that of most German suppliers, I fail to see the vast disparity between our quality/price and theirs.
Simply put; they stay competitive, meaning they know how to make more out of an 8 hour day than we do during a 10-12 hour day.
dude.
"it seems to me you went as "Guy with Giant iPod around his neck."
more like the "Guy with the super ultra shitty facsimile of a Guy with [a] Giant iPod around his neck."
the effort is there...but it just looks like crap.
dude.
People who are passionate about their work and love what they do aren't working, they're doing what they love.
It just goes to show that while money can motivate people, passion for the work is a better motivator.
You are absolutely right passion for work does motivate better and innovation will most certainly follow...but, as much as I'd love to live in an idealistic bubble, the sad truth is that there are simply not enough people in this world that are wired that way...through simple evolution the majority of the people in the world are passionate about procreation, consumption, and 300 channels of mind numbing crap.
So until the 97% out there bumps their collective heads on the toilet and have some life altering epiphany...the few will have to work their asses off to lead the many...monitary motivation is just one of the easier methods.
dude.
I dont want to sound like I know what I'm talking about, but aren't what you guys are refering to as analog clocks really digital? I mean, dont all clocks that dont use a crystal oscillator use gears to quantize the movement of a winding or weights? Cause that sounds digital to me.
dude.
Come on...where's the love? :-P
dude.
Remember: bigger storage is better, as is smaller size
In related news...working fewer hours and making more money kicks ass.
dude.
Good god!!! Its about time...I thought I was going to get all the way to the bottom of these postings and not have seen at least one Monty Python reference...c'mon geeks...you're better than that!
dude.
well the calls were routed to 311, so the calls got answered, just maybe not as quickly. Yes this sucks, but the calls didn't go nowhere.
I aint go no reason to doubt what you say.
dude.
aparently, i can read...but cant spell.
damn...
dude.
doesnt anybody actually RFTA???
"MiraVision can even be both things at once. Anyone wanting to watch TV while brushing their hair or fixing their make-up can switch to "picture in mirror" mode, which displays a small image in the corner of the frame..."
in rebuttal to this i submit that if you have the money to buy a ferrari...you either:
a.) can afford to also have a larger penis
b.) already have a large penis
c.) had to choose between the two and chose the car
d.) are a unic.
dude.