Android DOES support dual-SIM phones, so don't place the blame there. Just Google it, you'll find Android-based dual-SIM phones. Just not sold by AT&T, or TMobile.
I'm not sure why AT&T doesn't carry any, and maybe they soon will, now that they're using Enterproid. There's no reason to say that your two SIM cards won't both be locked to AT&T. You pay for two plans, but only carry one phone. Seems like a win for them.
It may have been the intent to have separate lines for Clear customers, but I know that wasn't the case at the Albany, NY airport. There, Clear customers just got the skip to the front of the line, just like airline / airport workers. They still went through the same check lines as everybody else; just didn't need to wait to show their ID.
That works only if there is a predictable "conversation" between client and server. If the client sends an encrypted key that changes each time, to which the server responds uniquely, you won't be able to simply replay the packet stream or design your own server. Not to say that this can't be done, but it will be a much more challenging proposition.
Yes, IBM did pretty much all the work, the STIDC is in Austin, on an IBM site, if I remember correctly. However, in terms of business purposes, very little of IBM had any plans for the Cell. They, Sony and Toshiba put in $400 Mill to develop the chip but IBM didn't have a roadmap for further development until the past year or so. From what I've seen, the STI guys were heads-down Cell development, and weren't paying too much attention to other potential applications until the thing finally went to fab. At that point, people at IBM started looking around and saying: "Hey, this is actually a pretty sweet chip. Maybe we can sell this to somebody." Mercury ( http://www.mc.com/cell/ ) was on board for actually making a product before IBM got around to slapping them into a blade-center chassis. Just because the initial announcements didn't say PS3 doesn't mean that it wasn't for the PS3 that it was originally developed. Most companies play things like that a bit close to the vest until a certain level of risk has been abated.
<sarcasm>But you are right, I "don't know anything about the Cell". The 2 blades (and 2 Cells per blade) that I've got at work are just collecting dust, and we never actually TALK to people at IBM.</sarcasm>
And not only that... The Cell *WAS* primarily designed for the PS3. Toshiba was involved partly due to being involved in the PS2, and IBM brings the Fab Brians + other skills. IBM didn't even think about using the Cell for anything until quite recently.
I can pretty much guarantee that it is Single Precision Floating Point or fixed point. DP drops the cell down to around 26 GFLOP from 256 GFLOP peak that can be reached (theoretically) with SP or Fixed-Point.
Actually it isn't too bad. We're 15 minutes from downtown Albany, 20 from Saratoga Springs. Also AMD is putting in their new plant pretty close by.
As for GE GRC itself and "pure research," a lot has been done there in the past, and some is still going on today, but these days there is a stronger requirement for GE business tie-in, from what I understand. Of course, I've only been there about a year or so.
Also, GE has Research Centers in Shanghai, Bangalore and Munich.
Here's the problem I have with Sony's public explaination as to why they are only using 7 SPUs. They can't control *which* SPU might have a fault during fab. Therefore, either: 1) They are just hoping that most faults will lie in SPU7 (or whichever they are turning off), or 2) They'll just turn off whichever SPU happens to have a serious fault.
The problem with #2 is this: The EIB. The Element Interconnect Bus is a ring structure, (2 16byte rings in each direction). You get maximum throughput when you talk to your neighbor. If you end up trying to talk to an SPU or the PPU or IOIFs that is half-way across the chip, it takes a lot longer to get there. Therefore, to get maximum performance out of a the CBE, you're going to want to be able to specify what programs are running on what SPU. If you can't know apriori which SPU is where, and what SPUs are even available, this gets significantly more difficult.
Sony understands that Data is King now, not MHz (though the Cell has been clocked to 4GHz), hence why they have the 200+GB/s EIB. It would be kinda silly to hamper its usage by making it nondeterministic which SPU would be gone.
Any game developers out there that can say if Sony has specified exactly which SPU is disabled, or if it is random?
Interesting... VERY interesting... I also have Safari 1.2.3, v125.9. When I hover my mouse over the link, it shows www.microsoft.com in the status bar. If I click the link, I go to google, but if I r-click and choose "Open Link in New Tab" (or new window) I go to www.microsoft.com.
Odd. Very odd. Hopefully Apple will arrange for some consistency in operation soon.
In addition to these, there is a paper coming out in the next ISPASS conference from some researchers at Notre Dame which looks at the effects of increasing the pipeline depth on the memory subsystem. It turns out that as you crank up the pipeline depth, you decrease the amount of "work" that can be done in a single cycle (obviously). The papers from ISCA fail to fully take the memory subsystem into consideration.
Now, for the most part, Comp. Sci and Eng. majors assume L1-caches to have 1-cycle latenies. Most current "real" processors do NOT have 1-cycle latencies, because it takes too long to access a cache of any useful size. As the pipeline depth increases, it gets much more difficult to have large L1 (or L2) caches.
Using the cache design simulator Cacti, we were able to get data on the approximate maximum sized cache, based off of pipeline depth (yes, this is fab-tech independant, check the paper for details). For example, if you consider a 5-cycle L1 delay (this is for a hit, not a miss), the maximum cache size you can get for a 10-stage is 512KB (as a 256K Instruction and 256K Data), for a 15-stage is 128, for a 20-stage is 32K, and a 25-stage would be 4K!
We simulated up to a 50-stage pipeline (the Intel paper above claims that a 50-stage pipeline is best), and the fastest cache we could simulate at that speed takes 8 cycles to read from the L1. This is for a 4K cache! (2K instruction, 2K data).
As anybody who has studied Computer Architecture before knows, caches need size to be effective. There are going to be some serious memory issues with these deeply-pipelined processors!
To a large part: Universities. Many of the mirrors are run by colleges / college LUGS. Of course, that means that the money is coming from 1) Students 2) Government, and therefore all of us through our taxes. Just remember, when paying your taxes, you are buying a faster apt-get update && apt-get dist-upgrade
Believe it or not, eventually you do learn how to read them without much trouble.
This is quite true. For whatever reason, my eyes very quickly bounced from "conversion from `_List_iterator" to "to non-scalar type `list".
That basically tells me all I need to know, along with the line 109 of readply.cpp. Maybe it is just mental training, but the STL errors that g++ generates are workable messages. You just need to get used to finding the relevant information.
Of course, I try to stay away from C++ anyway, and stick to C, but that's just personal preference for the most part. The real smart thing is to use the tool (language) that fits the problem best.
So you have 2 USB ports, but only 1 works. Now add a USB keyboard to your mix. SCREWED! Why? Well, I don't remember the details, but from what I remember, USB keyboards need to be directly connected to the computer, and cannot go through a USB hub.
Of course that could just be my particular keyboard. (Sun Type6 USB, on my x86 Linux Box)
I could see how this *might* work if you bought the product directly from Borland. If you go through a 3rd party at all, then your check is not going to Borland, but the 3rd party. What could a contract with a place like E.B. or Best Buy do to counter your Borland contract?
IANAL (I just did a little research into these bills): There are 2 types of wiretapping "Title III", which allows the recording of conversations, and "Pen Register/Trap Trace". "Title III" warrants require that the government (or other agency) convince a judge that there is "probable cause" that a wiretap would produce evidence of a crime. The "PR/TT" warrants are MUCH easier to obtain, (I think that judges cannot even really turn them down) but only allow the recording of telephone numbers. NO CONTENT.
These bills would modify the PR/TT wiretaps to allow the recording of "routing" and "addressing" information on all electronic communication, but cannot contain any content. Now I don't know about you, but www.google.com/search?q=George+Bush sure seems to me like it gives away content. The bills don't define exactly what "content" is, and so it is up to the enforcement agencies to determine what is and what isn't content. Basically, we then allow agencies to read everything, for free. Judges have effectively been taken out of the picture.
Also, the constitution forbids "blank" warrants. You must describe exactly where the tapping is to take place. The USA and PATRIOT acts would allow nationwide taps. This completely disregards the jurisdiction of a judge (except some higher courts).
Change your 'grep XXXXXXXXXXXXXXX' to 'grep default.ida' That way, you can get all the different variants. The 'X's are used by Code Red II, not the initial one.
This is the way to "defend" against software piracy. Defeat the hackers in a struggle through technology. Litigation in the courts is just not the way to stop people in the end. I have no problem with people wanting to have their customers pay for their product. I like how DirectTV responded to the piracy. Corporations (RIAA, MPAA, etc): BEAT US TECHNICALLY, NOT IN COURT! It means SO much more.
One quick change. "www.ndlug.nd.com" doesn't exist... Instead, it is "ftp.ndlug.nd.edu" However, it is currently down. Why, oh why, do these things happen over break??? The NDLUG Czar will be back to Notre Dame around the 8th of Jan, or so to work on ftp.ndlug.nd.edu, until then, try other mirrors.
In my High-school experience, (I'm now a Junior CSE major in college), our CS classes were really not focused as well as they should. Not to say that the teacher was bad, but he didn't really know WHAT he should be teaching. By calling the class "Computer Science," he should have been teaching, not only how to use a language, but also how to write programs.
We had CS1, CS2, CS3. CS1 == QBASIC, CS2 == Pascal, CS3 == whatever language the student wanted (but students had to teach themselves the language). We used IBM 286 machines, 1 of them had a 10MB HDD, the others had not HDD... Network boot. Our most difficult program in both CS1 and CS2 were such that they took me about 30 minutes to do, including goofing off, and he gave use a good 3 class periods (50 min. periods) to do it.
Our teacher was good that he recognized those of us in the class who could actually program, and he gave us more challenging assignments, etc, but he still didn't stress philosophy of programming, or even style much. I remember talking to him one time, and he told me that he was going to much the CS1 people use gotos to jump around the functions!!
It wasn't until College that I really learned to program, and although I still have more experience than a lot of people in my classes, those that began really writing code in high school, or before, have quite an advantage. I moved to Linux from Windows towards the end of my freshman year, when we first got exposed to unix, and have never looked back. It has helped me enormously.
Overall, I say that they should either re-structure the "CS" classes, at my school, and others whose "CS" classes are similar, or change the name to something more reflective of the class, such as "How to use BASIC", or "Free A"
Slashdot Mirror
Android DOES support dual-SIM phones, so don't place the blame there. Just Google it, you'll find Android-based dual-SIM phones. Just not sold by AT&T, or TMobile.
I'm not sure why AT&T doesn't carry any, and maybe they soon will, now that they're using Enterproid. There's no reason to say that your two SIM cards won't both be locked to AT&T. You pay for two plans, but only carry one phone. Seems like a win for them.
It may have been the intent to have separate lines for Clear customers, but I know that wasn't the case at the Albany, NY airport. There, Clear customers just got the skip to the front of the line, just like airline / airport workers. They still went through the same check lines as everybody else; just didn't need to wait to show their ID.
Well, I first read the tag 'typoinsummary' as "Typo In Space".
I guess you never quite get what you want. I could use a good lager right now though. Have to wait until after work, sadly.
Not including those deaths that occurred in Gary, Indiana, how many were there?
That works only if there is a predictable "conversation" between client and server. If the client sends an encrypted key that changes each time, to which the server responds uniquely, you won't be able to simply replay the packet stream or design your own server. Not to say that this can't be done, but it will be a much more challenging proposition.
Yes, IBM did pretty much all the work, the STIDC is in Austin, on an IBM site, if I remember correctly. However, in terms of business purposes, very little of IBM had any plans for the Cell. They, Sony and Toshiba put in $400 Mill to develop the chip but IBM didn't have a roadmap for further development until the past year or so. From what I've seen, the STI guys were heads-down Cell development, and weren't paying too much attention to other potential applications until the thing finally went to fab. At that point, people at IBM started looking around and saying: "Hey, this is actually a pretty sweet chip. Maybe we can sell this to somebody." Mercury ( http://www.mc.com/cell/ ) was on board for actually making a product before IBM got around to slapping them into a blade-center chassis. Just because the initial announcements didn't say PS3 doesn't mean that it wasn't for the PS3 that it was originally developed. Most companies play things like that a bit close to the vest until a certain level of risk has been abated.
<sarcasm>But you are right, I "don't know anything about the Cell". The 2 blades (and 2 Cells per blade) that I've got at work are just collecting dust, and we never actually TALK to people at IBM.</sarcasm>
And not only that... The Cell *WAS* primarily designed for the PS3. Toshiba was involved partly due to being involved in the PS2, and IBM brings the Fab Brians + other skills. IBM didn't even think about using the Cell for anything until quite recently.
I can pretty much guarantee that it is Single Precision Floating Point or fixed point. DP drops the cell down to around 26 GFLOP from 256 GFLOP peak that can be reached (theoretically) with SP or Fixed-Point.
Actually it isn't too bad. We're 15 minutes from downtown Albany, 20 from Saratoga Springs. Also AMD is putting in their new plant pretty close by.
As for GE GRC itself and "pure research," a lot has been done there in the past, and some is still going on today, but these days there is a stronger requirement for GE business tie-in, from what I understand. Of course, I've only been there about a year or so.
Also, GE has Research Centers in Shanghai, Bangalore and Munich.
That's why Cedega exists.
Here's the problem I have with Sony's public explaination as to why they are only using 7 SPUs. They can't control *which* SPU might have a fault during fab. Therefore, either: 1) They are just hoping that most faults will lie in SPU7 (or whichever they are turning off), or 2) They'll just turn off whichever SPU happens to have a serious fault.
The problem with #2 is this: The EIB. The Element Interconnect Bus is a ring structure, (2 16byte rings in each direction). You get maximum throughput when you talk to your neighbor. If you end up trying to talk to an SPU or the PPU or IOIFs that is half-way across the chip, it takes a lot longer to get there. Therefore, to get maximum performance out of a the CBE, you're going to want to be able to specify what programs are running on what SPU. If you can't know apriori which SPU is where, and what SPUs are even available, this gets significantly more difficult.
Sony understands that Data is King now, not MHz (though the Cell has been clocked to 4GHz), hence why they have the 200+GB/s EIB. It would be kinda silly to hamper its usage by making it nondeterministic which SPU would be gone.
Any game developers out there that can say if Sony has specified exactly which SPU is disabled, or if it is random?
Somebody who wants 8 SPUs? The CBE comes with 8 SPUs, but for some reason, Sony is disabling 1 of them on the version they are putting in the PS3.
Will you also be releasing the OSX version of Mambo/SystemSim, or just the Linux/PPC version?
Interesting... VERY interesting... I also have Safari 1.2.3, v125.9. When I hover my mouse over the link, it shows www.microsoft.com in the status bar. If I click the link, I go to google, but if I r-click and choose "Open Link in New Tab" (or new window) I go to www.microsoft.com.
Odd. Very odd. Hopefully Apple will arrange for some consistency in operation soon.
In addition to these, there is a paper coming out in the next ISPASS conference from some researchers at Notre Dame which looks at the effects of increasing the pipeline depth on the memory subsystem. It turns out that as you crank up the pipeline depth, you decrease the amount of "work" that can be done in a single cycle (obviously). The papers from ISCA fail to fully take the memory subsystem into consideration.
Now, for the most part, Comp. Sci and Eng. majors assume L1-caches to have 1-cycle latenies. Most current "real" processors do NOT have 1-cycle latencies, because it takes too long to access a cache of any useful size. As the pipeline depth increases, it gets much more difficult to have large L1 (or L2) caches.
Using the cache design simulator Cacti, we were able to get data on the approximate maximum sized cache, based off of pipeline depth (yes, this is fab-tech independant, check the paper for details). For example, if you consider a 5-cycle L1 delay (this is for a hit, not a miss), the maximum cache size you can get for a 10-stage is 512KB (as a 256K Instruction and 256K Data), for a 15-stage is 128, for a 20-stage is 32K, and a 25-stage would be 4K!
We simulated up to a 50-stage pipeline (the Intel paper above claims that a 50-stage pipeline is best), and the fastest cache we could simulate at that speed takes 8 cycles to read from the L1. This is for a 4K cache! (2K instruction, 2K data).
As anybody who has studied Computer Architecture before knows, caches need size to be effective. There are going to be some serious memory issues with these deeply-pipelined processors!
To a large part: Universities. Many of the mirrors are run by colleges / college LUGS. Of course, that means that the money is coming from 1) Students 2) Government, and therefore all of us through our taxes. Just remember, when paying your taxes, you are buying a faster apt-get update && apt-get dist-upgrade
Believe it or not, eventually you do learn how to read them without much trouble.
This is quite true. For whatever reason, my eyes very quickly bounced from "conversion from `_List_iterator" to "to non-scalar type `list".
That basically tells me all I need to know, along with the line 109 of readply.cpp. Maybe it is just mental training, but the STL errors that g++ generates are workable messages. You just need to get used to finding the relevant information.
Of course, I try to stay away from C++ anyway, and stick to C, but that's just personal preference for the most part. The real smart thing is to use the tool (language) that fits the problem best.
So you have 2 USB ports, but only 1 works. Now add a USB keyboard to your mix. SCREWED! Why? Well, I don't remember the details, but from what I remember, USB keyboards need to be directly connected to the computer, and cannot go through a USB hub.
Of course that could just be my particular keyboard. (Sun Type6 USB, on my x86 Linux Box)
Don't even joke about something like this! I can just see it now:
Boss: Are you sure we need all this obtrusive advertising? What if it makes people upset at us?
Marketing: Just look at this posting from slashdot. Geeks love this stuff.
Boss: Well, if it works for them... Do it!
I could see how this *might* work if you bought the product directly from Borland. If you go through a 3rd party at all, then your check is not going to Borland, but the 3rd party. What could a contract with a place like E.B. or Best Buy do to counter your Borland contract?
IANAL (I just did a little research into these bills): There are 2 types of wiretapping "Title III", which allows the recording of conversations, and "Pen Register/Trap Trace". "Title III" warrants require that the government (or other agency) convince a judge that there is "probable cause" that a wiretap would produce evidence of a crime. The "PR/TT" warrants are MUCH easier to obtain, (I think that judges cannot even really turn them down) but only allow the recording of telephone numbers. NO CONTENT.
These bills would modify the PR/TT wiretaps to allow the recording of "routing" and "addressing" information on all electronic communication, but cannot contain any content. Now I don't know about you, but www.google.com/search?q=George+Bush sure seems to me like it gives away content. The bills don't define exactly what "content" is, and so it is up to the enforcement agencies to determine what is and what isn't content. Basically, we then allow agencies to read everything, for free. Judges have effectively been taken out of the picture.
Also, the constitution forbids "blank" warrants. You must describe exactly where the tapping is to take place. The USA and PATRIOT acts would allow nationwide taps. This completely disregards the jurisdiction of a judge (except some higher courts).
Change your 'grep XXXXXXXXXXXXXXX' to 'grep default.ida' That way, you can get all the different variants. The 'X's are used by Code Red II, not the initial one.
This is the way to "defend" against software piracy. Defeat the hackers in a struggle through technology. Litigation in the courts is just not the way to stop people in the end. I have no problem with people wanting to have their customers pay for their product. I like how DirectTV responded to the piracy. Corporations (RIAA, MPAA, etc): BEAT US TECHNICALLY, NOT IN COURT! It means SO much more.
One quick change. "www.ndlug.nd.com" doesn't exist... Instead, it is "ftp.ndlug.nd.edu" However, it is currently down. Why, oh why, do these things happen over break??? The NDLUG Czar will be back to Notre Dame around the 8th of Jan, or so to work on ftp.ndlug.nd.edu, until then, try other mirrors.
Branden
Treasurer, NDLUG
In my High-school experience, (I'm now a Junior CSE major in college), our CS classes were really not focused as well as they should. Not to say that the teacher was bad, but he didn't really know WHAT he should be teaching. By calling the class "Computer Science," he should have been teaching, not only how to use a language, but also how to write programs.
We had CS1, CS2, CS3. CS1 == QBASIC, CS2 == Pascal, CS3 == whatever language the student wanted (but students had to teach themselves the language). We used IBM 286 machines, 1 of them had a 10MB HDD, the others had not HDD... Network boot. Our most difficult program in both CS1 and CS2 were such that they took me about 30 minutes to do, including goofing off, and he gave use a good 3 class periods (50 min. periods) to do it.
Our teacher was good that he recognized those of us in the class who could actually program, and he gave us more challenging assignments, etc, but he still didn't stress philosophy of programming, or even style much. I remember talking to him one time, and he told me that he was going to much the CS1 people use gotos to jump around the functions!!
It wasn't until College that I really learned to program, and although I still have more experience than a lot of people in my classes, those that began really writing code in high school, or before, have quite an advantage. I moved to Linux from Windows towards the end of my freshman year, when we first got exposed to unix, and have never looked back. It has helped me enormously.
Overall, I say that they should either re-structure the "CS" classes, at my school, and others whose "CS" classes are similar, or change the name to something more reflective of the class, such as "How to use BASIC", or "Free A"
-- Just my thoughts...
bmoore