RISC actually refers to a bundle of tricks and optimizations. Most modern "RISC" processors are not all that reduced in instruction set size. Some of these tricks are:
Regular instruction set - all of the instructions are the same number of bits in length. CISC processors often have some instruction that are 2 bytes long, others are 4 bytes long, etc. Decoding this in a timely fashion was a problem back in the late 80's, early 90's. Today, there are lots and lots of transistors on the CPU to throw at this.
Single cycle execution - Originally RISC instructions were designed to be executed in a single cycle. CISC processors often had highly complex instructions that would take multiple cycles to execute (sometimes 100's or 1000's in the case of the VAX). Today, processors are pipelined but the idea is still the same - pump out instructions once per clock cycle. CISC processors tend to do this now as well.
Lots of registers - Registers are fast access memory inside the CPU itself. CISC processors varied in the number of registers available. Some, like the VAX and 68000 had lots of registers while others like the 80x86 have smaller numbers. RISC chips typically have large number of registers (unless you consider the 6502 as a RISC chip:-)). More registers make for faster code in general. I think that the new x64 extensions have increased the number of registers in the Intel architecture and, in any case, fast on-chip caches make the point almost moot.
Uncomplicated instruction set - The "R" in RISC usually really meant that all of the complicated instructions were thrown out and instead the compiler would just emit code to do them. Most of these terribly complex instructions (the VAX had a classic example which was the "evaluate polynomial" instruction) were actually microcoded (microcode is software that is a level below assembler, typically hardwired into the CPU) so they were not much faster than simply coding the routine using assembler. Easier to use if you're coding assembler by hand but who does that?
So, RISC was a lot more things than moving complex instructions from hardware to software AND the complex instructions that were moved from hardware to software were typically slow "in hardware" because they weren't really hardware but microcode. Most of the advantages of pure RISC are no longer as large as they were in the early '90's because the sheer number of transistors available (more transistors means you can make more difficult things run faster) is so large now that those optimizations don't buy you much. The battle now really comes down to how much money can you throw at the chip in order to produce new versions of it that run faster. When you look at the sheer amount of money invested in the Intel architecture vs PowerPC or SPARC or Alpha you will begin to have a lot more respect for what the RISC camp accomplished with a much smaller budget.
It wasn't just the model that was asinine
on
DECnet Isn't Dead
·
· Score: 1
The 7 layer model was bad enough (the bottom layers weren't bad, but no one could ever figure out the difference between "Application" and "Presentation as I recall or something like that)but the actual protocols blew donkeys.
For a long time there was no connectionless (UDP equivalent) protocol. We used to have UltraNet as an option on our mini-supers and I was the guy in charge of porting the drivers, etc. UltraNet, in 1990, ran at up to 1Gbps. Very cool stuff. Unfortunately they chose to make their core protocol OSI and it was connection based only (in order to get things running at 1Gbps they offloaded most of the protocol processing into hardware). This was fine in certain supercomputing environments where people wanted to to just blast files around with FTP but as soon as you got out into the working world people wanted to run NFS over it (this was pre-TCP NFS) and the performance just stank.
At another point I want to a Usenix and went to a seminar on the OSI RPC protocol. I had been digging around in the guts of SunRPC and NFS quite a bit at that time and so I went up after the class was over and asked the instructor about how they had made tradeoffs, etc. This guy was one of the people who wrote the spec and he said to me "Well, I've never really looked at SunRPC". And boy, did it show. The protocol was all bit and byte stuffed for maximum bit efficiency just as we were moving to faster networks (so not so much need for bit efficiency) and RISC processors that gave you huge penalties for accessing non-aligned data (so processing the protocol would be damn slown). That convinced me that the whole thing was a crock. When GOSIP (this was the government push to have OSI as their standard networking protocols) died everyone ran away from OSI as fast as they could holding their noses.
I hit the submit button before I finished posting. One more thing...
Most business are running on some kind of PBX. Either the PBX has backup power (in which case VOIP will probably keep going) or they can't make phone calls anyway.
For me, power outages are a minor annoyance. But for companies? How can they deal with virtually all lines to their associates being cut?
If the power goes out to a business you have a lot more troubles than just the phone. Most business rely on computers - oops, they're down. Many newer office building have a lot of interior space that doesn't get natural lighting - oops, everyone's in the dark.
Hello! 1950 is calling and they want their business environment back.
In small companies and large budgets are constantly being revised. Publicly traded companies are just as bad if not worse than small companies as they are managed to their quarterly earnings, not to a yearly budget. Within departments projects get approved, disapproved and juggled regularly. Some other department goes over budget and your department pays. The company has a "hiring freeze" or, worse, layoffs.
Get real. Yah, everyone does a yearly budget but it is not a cast in stone contract and it's subject to revision at any time. Head count is constantly tweaked. I've worked, and managed, in large (yes, publicly traded) and small companies and I've seen it again and again and again.
Well, what typically happens is that Engineering is told "You can go hire X heads" (employees). Engineering looks around, tries to find good people, it takes some time, and by the time they are ready to hire someone, upper management goes "Oh, no, you can't have those heads anymore."
Chip's mistake, from what I can tell, is not consulting with a lawyer before he did anything
Yah, everybody's mistake is not consulting with a lawyer before tying their shoes or opening their mouths. I think that everyone should have a lawyer with them at all times to make sure that they don't do or say anything actionable.
I say it's not a choice because it's what their counsel tells them.
Maybe the problem is that lawyers tend to see every problem as requiring legal action and more billable hours.
The reason for having a patent system is "To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries" (that's what it says in the Constitution).
Allowing entities to patent large swaths of ideas and then expect them to be "good stewards" of them is not what the patent system is for. Why should we have to rely on Microsoft's good graces to use what are obviously trivial ideas.
Overly broad patents should not be granted. AT&T had a long history of being good about licensing patents widely and not trying to extort money for them. SBC, the inheiritor of that patent from Ameritech, went fishing for money with it. What guarantees are there that Microsoft will not be taken over by someone else (since you believe they are such wonderful folks) who would decide to enforce all of their dodgy patents?
No, solutions that rely on corporate self-restraint are fundamentally flawed. Furthermore, having Microsoft be the owner of the patent is just luck. The system needs to stop granting these things. In fact, the system needs to start punishing people who apply for this crap. Patent lawyers will apply for just about anything because the risks associated with not researching prior art enough are completely negligible.
Do you know how to read? Here's my second paragraph:
Japan got what they had coming to them. Looking at the effects of the atomic bombings in isolation and going "Oh, how awful" is worthless. You have to look at the whole war and take actions like the atomic bombings in the context of the time.
Now how is this blaming the U.S. for the war?
As far as vets from WWII or any conflict I have a lot of respect for them and I don't denigrate them for what they did, it was necessary. However what they did and what was done to them by the war was horrible, within the bounds of the Geneva Convention or not. War is sometimes necessary and the results are sometimes noble but the process is always brutal.
You always have a say in the matter. If the people of Japan had known where WWII would lead them they probably would have declined to follow.
Do I blame civilian Iraqis for the gassing of the Kurds? Yes, I do. One man cannot rule a country. It takes a huge infrastructure of complicit people to handle it. Make no mistake, Sadaam remained in power for as long as he did because a large number of Iraqis wanted him to. And I blame the American electorate for supporting Dubya in invading Iraq even though they were lied to.
How many people did Pol Pot slay by his own hand? The killing fields of Cambodia, like the death camps of Nazi Germany and the gulags of the USSR were operated by people who were complicit. They may not have been the majority, but the majority believed, for whatever reason, that they were better off not rebelling but instead allowed these people to perpetrate their atrocities. It may come down to thinking that you have a better chance of surviving or that your kids may have a better chance of surviving if you don't fight but the choice is there and it is made.
Showing the effects of just about any weapon of war will sicken a normal person. And yet, somehow, we keep on managing to figure out ways to dehumanize opponents enough to justify in our minds waging war on them.
Japan got what they had coming to them. Looking at the effects of the atomic bombings in isolation and going "Oh, how awful" is worthless. You have to look at the whole war and take actions like the atomic bombings in the context of the time.
I live in Japan currently, my wife is Japanese and my children are half-Japanese (I am American). I enjoy Japan and I like the Japanese people. It's hard to imagine now how a war like WWII could have been fought by them.
My landlord, at 80+, was in the Army and served during WWII. He's a nice old man who likes to garden and play with my kids. I've never had a conversation with him about what he did during the war though it wouldn't surprise me if he had been running around with a bayonet through Nanking or poking POWs along the Bataan trail. It was what you did at that time and somehow there is a collective insanity that sweeps men up and gives them license to run amok.
My grandfather drove landing boats in the Pacific during WWII. He never talked much about it, but my grandmother told me he used to wake up in cold sweats in the middle of the night after he got back. I knew other men from his generation who had been to war and must have been through and done terrible things. Yet they came back and went back to normal lives and did normal things and we sat and ate dinner with them. And we, as a society, condoned what they had done and dreamed up ways to kill more people faster and easier while still being concerned about what kind of car to drive and what kind of school the kids should go to.
Death comes to us one at a time. Each life lost is a tragedy. Atomic weapons changes these tragedies into statistics but make no mistake, each death is still a tragedy. And each life lost to a bullet is just as much a tragedy as one lost to a nuke. War is terrible and destructive and to be avoided. Let's not pretend that some ways of making war are better than others.
No, he's just pulling the Nyquist limit out and saying that's really the sample rate, or some such. I think what he means is that the maximum frequency recorded by a CD is 22 KHz which is true since the sampling rate is 44.1 KHz.
Actually, electronic communications ARE protected by the 4th Amendment, at least the Supreme Court decided so in Katz vs United States (incidentally overturning a previous Supreme Court ruling that they were not). There's a nice little summary here.
Well, Lucas, I looked through some of your other posts and noticed that your have encryption turned on on your wireless network. Why? Do you have something to hide?
I assume that you have encryption turned on to keep bad people from hacking into your network and reading your PRIVATE data. Now, how good a job do you think your ISP is going to do of securing all of the logs of all of your activity?
And since your Commercial ISP is doing Business in the Several States and engaging in "Interstate Commerce" by contracting with Commercial out-of-state network providers, Big Brother, under the MIGHTY ALL-ENCOMPASSING INTERSTATE COMMERCE CLAUSE, has the "Constitutional" authority to require your Commercial ISP to keep logs of all your mother fucking e-mails and online activities
Whilst the Interstate Commerce Clause has been abused mightily your private conversations are not commerce and for the Feds to requires logs of them to be kept based on it seem like a bit of stretch to me. It might be conceivable to ask that what IP addresses you connect to be logged, but even there, the IP addresses you connect to are not relevant to commerce because the ISP does not charge you based on where you are connecting to (as opposed to traditional long distance service).
Brokerages communications with and about their customers are pretty obviously directly relevant to commerce.
BTW, I'm more of a libertarian fucktard. Let's try to keep our ad hominem attacks accurate, shall we?
I RTFA and, again, "child pornography" is being trotted our as the excuse for violating everyone's rights. Does anyone have any idea how much kiddie porn is really out there? I'd go look but I don't want anything hanging around in my browser cache.
Emails aren't "papers and effects"? The Constitution could not, of course, cover every technological and social innovation but this seems covered without stretching the interpretation at all.
You said the right words - don't you think that this is an unlawful search and seizure?
Amendment IV - Search and seizure. Ratified 12/15/1791.
The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.
Amendment IV - Search and seizure. Ratified 12/15/1791.
The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.
Slashdot Mirror
RISC actually refers to a bundle of tricks and optimizations. Most modern "RISC" processors are not all that reduced in instruction set size. Some of these tricks are:
:-)). More registers make for faster code in general. I think that the new x64 extensions have increased the number of registers in the Intel architecture and, in any case, fast on-chip caches make the point almost moot.
Regular instruction set - all of the instructions are the same number of bits in length. CISC processors often have some instruction that are 2 bytes long, others are 4 bytes long, etc. Decoding this in a timely fashion was a problem back in the late 80's, early 90's. Today, there are lots and lots of transistors on the CPU to throw at this.
Single cycle execution - Originally RISC instructions were designed to be executed in a single cycle. CISC processors often had highly complex instructions that would take multiple cycles to execute (sometimes 100's or 1000's in the case of the VAX). Today, processors are pipelined but the idea is still the same - pump out instructions once per clock cycle. CISC processors tend to do this now as well.
Lots of registers - Registers are fast access memory inside the CPU itself. CISC processors varied in the number of registers available. Some, like the VAX and 68000 had lots of registers while others like the 80x86 have smaller numbers. RISC chips typically have large number of registers (unless you consider the 6502 as a RISC chip
Uncomplicated instruction set - The "R" in RISC usually really meant that all of the complicated instructions were thrown out and instead the compiler would just emit code to do them. Most of these terribly complex instructions (the VAX had a classic example which was the "evaluate polynomial" instruction) were actually microcoded (microcode is software that is a level below assembler, typically hardwired into the CPU) so they were not much faster than simply coding the routine using assembler. Easier to use if you're coding assembler by hand but who does that?
So, RISC was a lot more things than moving complex instructions from hardware to software AND the complex instructions that were moved from hardware to software were typically slow "in hardware" because they weren't really hardware but microcode. Most of the advantages of pure RISC are no longer as large as they were in the early '90's because the sheer number of transistors available (more transistors means you can make more difficult things run faster) is so large now that those optimizations don't buy you much. The battle now really comes down to how much money can you throw at the chip in order to produce new versions of it that run faster. When you look at the sheer amount of money invested in the Intel architecture vs PowerPC or SPARC or Alpha you will begin to have a lot more respect for what the RISC camp accomplished with a much smaller budget.
The 7 layer model was bad enough (the bottom layers weren't bad, but no one could ever figure out the difference between "Application" and "Presentation as I recall or something like that)but the actual protocols blew donkeys.
For a long time there was no connectionless (UDP equivalent) protocol. We used to have UltraNet as an option on our mini-supers and I was the guy in charge of porting the drivers, etc. UltraNet, in 1990, ran at up to 1Gbps. Very cool stuff. Unfortunately they chose to make their core protocol OSI and it was connection based only (in order to get things running at 1Gbps they offloaded most of the protocol processing into hardware). This was fine in certain supercomputing environments where people wanted to to just blast files around with FTP but as soon as you got out into the working world people wanted to run NFS over it (this was pre-TCP NFS) and the performance just stank.
At another point I want to a Usenix and went to a seminar on the OSI RPC protocol. I had been digging around in the guts of SunRPC and NFS quite a bit at that time and so I went up after the class was over and asked the instructor about how they had made tradeoffs, etc. This guy was one of the people who wrote the spec and he said to me "Well, I've never really looked at SunRPC". And boy, did it show. The protocol was all bit and byte stuffed for maximum bit efficiency just as we were moving to faster networks (so not so much need for bit efficiency) and RISC processors that gave you huge penalties for accessing non-aligned data (so processing the protocol would be damn slown). That convinced me that the whole thing was a crock. When GOSIP (this was the government push to have OSI as their standard networking protocols) died everyone ran away from OSI as fast as they could holding their noses.
I hit the submit button before I finished posting. One more thing...
Most business are running on some kind of PBX. Either the PBX has backup power (in which case VOIP will probably keep going) or they can't make phone calls anyway.
For me, power outages are a minor annoyance. But for companies? How can they deal with virtually all lines to their associates being cut?
If the power goes out to a business you have a lot more troubles than just the phone. Most business rely on computers - oops, they're down. Many newer office building have a lot of interior space that doesn't get natural lighting - oops, everyone's in the dark.
Hello! 1950 is calling and they want their business environment back.
In small companies and large budgets are constantly being revised. Publicly traded companies are just as bad if not worse than small companies as they are managed to their quarterly earnings, not to a yearly budget. Within departments projects get approved, disapproved and juggled regularly. Some other department goes over budget and your department pays. The company has a "hiring freeze" or, worse, layoffs.
Get real. Yah, everyone does a yearly budget but it is not a cast in stone contract and it's subject to revision at any time. Head count is constantly tweaked. I've worked, and managed, in large (yes, publicly traded) and small companies and I've seen it again and again and again.
Well, what typically happens is that Engineering is told "You can go hire X heads" (employees). Engineering looks around, tries to find good people, it takes some time, and by the time they are ready to hire someone, upper management goes "Oh, no, you can't have those heads anymore."
It sucks to be on either end, trust me.
Chip's mistake, from what I can tell, is not consulting with a lawyer before he did anything
Yah, everybody's mistake is not consulting with a lawyer before tying their shoes or opening their mouths. I think that everyone should have a lawyer with them at all times to make sure that they don't do or say anything actionable.
I say it's not a choice because it's what their counsel tells them.
Maybe the problem is that lawyers tend to see every problem as requiring legal action and more billable hours.
I get up and go to the bathroom during ads on television.
The reason for having a patent system is "To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries" (that's what it says in the Constitution).
Allowing entities to patent large swaths of ideas and then expect them to be "good stewards" of them is not what the patent system is for. Why should we have to rely on Microsoft's good graces to use what are obviously trivial ideas.
Overly broad patents should not be granted. AT&T had a long history of being good about licensing patents widely and not trying to extort money for them. SBC, the inheiritor of that patent from Ameritech, went fishing for money with it. What guarantees are there that Microsoft will not be taken over by someone else (since you believe they are such wonderful folks) who would decide to enforce all of their dodgy patents?
No, solutions that rely on corporate self-restraint are fundamentally flawed. Furthermore, having Microsoft be the owner of the patent is just luck. The system needs to stop granting these things. In fact, the system needs to start punishing people who apply for this crap. Patent lawyers will apply for just about anything because the risks associated with not researching prior art enough are completely negligible.
They're going to show up at the senator's door with a dump truck full of money?
Do you know how to read? Here's my second paragraph:
Japan got what they had coming to them. Looking at the effects of the atomic bombings in isolation and going "Oh, how awful" is worthless. You have to look at the whole war and take actions like the atomic bombings in the context of the time.
Now how is this blaming the U.S. for the war?
As far as vets from WWII or any conflict I have a lot of respect for them and I don't denigrate them for what they did, it was necessary. However what they did and what was done to them by the war was horrible, within the bounds of the Geneva Convention or not. War is sometimes necessary and the results are sometimes noble but the process is always brutal.
You always have a say in the matter. If the people of Japan had known where WWII would lead them they probably would have declined to follow.
Do I blame civilian Iraqis for the gassing of the Kurds? Yes, I do. One man cannot rule a country. It takes a huge infrastructure of complicit people to handle it. Make no mistake, Sadaam remained in power for as long as he did because a large number of Iraqis wanted him to. And I blame the American electorate for supporting Dubya in invading Iraq even though they were lied to.
How many people did Pol Pot slay by his own hand? The killing fields of Cambodia, like the death camps of Nazi Germany and the gulags of the USSR were operated by people who were complicit. They may not have been the majority, but the majority believed, for whatever reason, that they were better off not rebelling but instead allowed these people to perpetrate their atrocities. It may come down to thinking that you have a better chance of surviving or that your kids may have a better chance of surviving if you don't fight but the choice is there and it is made.
Showing the effects of just about any weapon of war will sicken a normal person. And yet, somehow, we keep on managing to figure out ways to dehumanize opponents enough to justify in our minds waging war on them.
Japan got what they had coming to them. Looking at the effects of the atomic bombings in isolation and going "Oh, how awful" is worthless. You have to look at the whole war and take actions like the atomic bombings in the context of the time.
I live in Japan currently, my wife is Japanese and my children are half-Japanese (I am American). I enjoy Japan and I like the Japanese people. It's hard to imagine now how a war like WWII could have been fought by them.
My landlord, at 80+, was in the Army and served during WWII. He's a nice old man who likes to garden and play with my kids. I've never had a conversation with him about what he did during the war though it wouldn't surprise me if he had been running around with a bayonet through Nanking or poking POWs along the Bataan trail. It was what you did at that time and somehow there is a collective insanity that sweeps men up and gives them license to run amok.
My grandfather drove landing boats in the Pacific during WWII. He never talked much about it, but my grandmother told me he used to wake up in cold sweats in the middle of the night after he got back. I knew other men from his generation who had been to war and must have been through and done terrible things. Yet they came back and went back to normal lives and did normal things and we sat and ate dinner with them. And we, as a society, condoned what they had done and dreamed up ways to kill more people faster and easier while still being concerned about what kind of car to drive and what kind of school the kids should go to.
Death comes to us one at a time. Each life lost is a tragedy. Atomic weapons changes these tragedies into statistics but make no mistake, each death is still a tragedy. And each life lost to a bullet is just as much a tragedy as one lost to a nuke. War is terrible and destructive and to be avoided. Let's not pretend that some ways of making war are better than others.
No, he's just pulling the Nyquist limit out and saying that's really the sample rate, or some such. I think what he means is that the maximum frequency recorded by a CD is 22 KHz which is true since the sampling rate is 44.1 KHz.
Hey, the joke's no good if you don't really do it :-). Besides, what *good* is the foe list anyway?
OK...you're a butthead.
:-)
And you're on my foe list.
Happy now?
I see that we did not Read The Fine Article. The beginning of the second paragraph:
Data retention rules could permit police to obtain records of e-mail chatter, Web browsing or chat-room activity
Records of "E-mail chatter and chat-room activity" certainly sound like YOUR EFFECTs to me.
Actually, electronic communications ARE protected by the 4th Amendment, at least the Supreme Court decided so in Katz vs United States (incidentally overturning a previous Supreme Court ruling that they were not). There's a nice little summary here.
Well, Lucas, I looked through some of your other posts and noticed that your have encryption turned on on your wireless network. Why? Do you have something to hide?
I assume that you have encryption turned on to keep bad people from hacking into your network and reading your PRIVATE data. Now, how good a job do you think your ISP is going to do of securing all of the logs of all of your activity?
And since your Commercial ISP is doing Business in the Several States and engaging in "Interstate Commerce" by contracting with Commercial out-of-state network providers, Big Brother, under the MIGHTY ALL-ENCOMPASSING INTERSTATE COMMERCE CLAUSE, has the "Constitutional" authority to require your Commercial ISP to keep logs of all your mother fucking e-mails and online activities
Whilst the Interstate Commerce Clause has been abused mightily your private conversations are not commerce and for the Feds to requires logs of them to be kept based on it seem like a bit of stretch to me. It might be conceivable to ask that what IP addresses you connect to be logged, but even there, the IP addresses you connect to are not relevant to commerce because the ISP does not charge you based on where you are connecting to (as opposed to traditional long distance service).
Brokerages communications with and about their customers are pretty obviously directly relevant to commerce.
BTW, I'm more of a libertarian fucktard. Let's try to keep our ad hominem attacks accurate, shall we?
I RTFA and, again, "child pornography" is being trotted our as the excuse for violating everyone's rights. Does anyone have any idea how much kiddie porn is really out there? I'd go look but I don't want anything hanging around in my browser cache.
Emails aren't "papers and effects"? The Constitution could not, of course, cover every technological and social innovation but this seems covered without stretching the interpretation at all.
You said the right words - don't you think that this is an unlawful search and seizure?
Amendment IV - Search and seizure. Ratified 12/15/1791.
The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.
US Constitution
Amendment IV - Search and seizure. Ratified 12/15/1791.
The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.
U.S. Constitution
Section 8
The Congress shall have Power
Clause 3: To regulate Commerce with foreign Nations, and among the several States, and with the Indian Tribes;