Actually Intel is one of the few companies to implement the full IEEE FP standard, most other processors do not. The 80 bit is for internal datapaths for the multiplier and stuff, still you get 64bit doubles out....:)
Well the general idea for RISC according to Paterson came from having to program the microcode in the VAXen, so at some point they figured out that the translation overhead from the CISC instructions was unecesary. CISC came to be as a sort of "compresion" mechanism in which one instruciton could be fetched from memory and then it can be mapped into a sequence of micro-ops inside the CPU. Thus the intruction memory space and bandwidth requirements were substatial (afterall memory was a premium back then). People figured out that memory may get larger, so why not just fetch the microops themselves, so it was a trade off between memory footprint and decoding complexity.
Of course early designs like the CDC 6600 were pretty RISC, although they didn't know it. And the CISC came out as a solution to the memory footprint problems that such architectures poised (in scaling down from supers to minis where cost was a priority).
However it is really eye opening to know the actual translation overhead, control related with the translator in the x86 family is significant. And almost half of the dynamic transistor budget (not SRAM) goes into the translation/scheduling portion not the functional. So the overhead is much more significan than...
And yes their approach is microprogramming, except that RISC-core is more buzzword compliant:).
They were late? As in offering one of the first lines of RISC workstations on the market is being "late" jeez! (The DECStations came out in '88, which I believe places them almost earlier than the PA-RISC, or even SPARCStations).
When it came out Alpha was faster than any other CPU period, and not just faster but significantly more powerful than anything else in the market. Obviously you must have a very selective memory if you don't remember the early 90's and everyone lusting after Alpha big time. And that was a clear advantage, however DEC could not market themselves even if they were a free whore in a port.
" Decoding x86 is a major problem. See how many different solutions have been employed to do it since the PPro and K5. * x86 chips are designed to the core to run x86. The fact that it breaks x86 into smaller instructions doesn't make it an Alpha beind a x86 decoder. It's like saying PPC970 is an Alpha behind a PowerPC decoder."
I have newsflash for you, x86 instructions have never been executed natively (most of them anyways), it is all microcode. This whole shebang of x86 "risc" core is from people who don't get it that all that the P6 is doing is using a leaner microcode that can be pipelined and superscaled.
The influence of the Alpha in the x86 comes from the fact that Intel did indeed buy out the Alpha technology after they settled their lawsuit with DEC, and a shitload of the Alpha team ended up developing for intel. Thus a lot of the internal pipeline technology ended up in the PIV et al... it doesn't make it an alpha, but some of the design principles are there.
Actually Sequent (now part of IBM) has been offering large SMP x86 machines from the mid 80's. I remember seeing a 16 processor 386 based machine in the late 80s. Also their new NUMA-Q architecture used Xeon's and scaled up to 64 processors. Also Data General moved their Aviion line from motorola 88K's to Xeons and could also scale up to 32 processors using their own NUMA architecture. Compaq has also offered 8-way x86 machines for a long time too...
The R in RISC is for complexity not number.. in the opposite as the C in CISC. Has nothing to do with how many Opcodes are present in your ISA.
RISC just implies that each instruction is easy to decode and can be geared towards achieving an effective single clock execution rate. Whether you have 1 or 2000 instructions in the ISA is irrelevant as long as those instructions are pretty simple.
This whole CISC being RISCy is not that innovative though, all that Intel and AMD have done is pipeline their microcode. In fact RISC can be viewed as the microcode for the CISC machines of yore, which is in fact how it came to be.
Sorry Out of Order execution has been done for ages before Intel implemented it. As usual different name, same concept. But the same ideas behind RISC, superscalar, out-of-order, pipelined, etc.. have been around (and implemented) since the 60's.
Not even in micros, as I believe Metaflow and other vendors had out-of-order CPUS out there way before Intel released the P6 microarchitecture.
Actually the firs series of big Alpha systems (7000) were a VAX with an alpha CPU. So the parent wasn't that retarded afterall.... so believe it or not your apple ][ analogy wasn't a joke but a real description of what the early alpha servers were./EOL
The point being that just because your guy believes strongly in issues doesn't necessarily make him the right person for the job. But you like him, and by all means vote for him. Such is the nature of a republic/democracy whatever the system it is that we have.
Again you want a dictator not a president. Like a much wiser guy than me once said: "To err is human, to truly correct is divine." A person that is asked if he can think of any mistake he has done while in office, and can not think of anything he's done wrong is not the right person for the job. That is why Bush is fatally flawed as a president. You however see that as a strength, which is why I am f$%king scared of this country right now.
"On the one hand, we have Bush; a guy who (most of the time) has a clear agenda, strong beliefs, and tends to stick to them (whether we like them or not). The end result is that he may do some things that we don't like, simply because he believes in them."
Well, Hitler too did things because he had "clear" beliefs... and he stuck to them. So what is your point? He is not the president to stick to his beliefs, but to work for the American people. According to you then, a lemming would be the ideal democratic leader, right? I have newsflash for you a democracy means that there are many different beliefs, a president has to be able to accomodate most of them, w/o having to impose his own. You want a dictator, not a president...
And please don't come up with the whole "moral clarity vs. flip/flopping" crap, not like Dubya has ever flip/flopped, no siree bob:
1. Social Security Surplus
BUSH PLEDGES NOT TO TOUCH SOCIAL SECURITY SURPLUS... "We're going to keep the promise of Social Security and keep the government from raiding the Social Security surplus." [President Bush, 3/3/01]...BUSH SPENDS SOCIAL SECURITY SURPLUS The New York Times reported that "the president's new budget uses Social Security surpluses to pay for other programs every year through 2013, ultimately diverting more than $1.4 trillion in Social Security funds to other purposes." [The New York Times, 2/6/02]
2. Patient's Right to Sue
GOVERNOR BUSH VETOES PATIENTS' RIGHT TO SUE... "Despite his campaign rhetoric in favor of a patients' bill of rights, Bush fought such a bill tooth and nail as Texas governor, vetoing a bill coauthored by Republican state Rep. John Smithee in 1995. He... constantly opposed a patient's right to sue an HMO over coverage denied that resulted in adverse health effects." [Salon, 2/7/01]...CANDIDATE BUSH PRAISES TEXAS PATIENTS' RIGHT TO SUE... "We're one of the first states that said you can sue an HMO for denying you proper coverage... It's time for our nation to come together and do what's right for the people. And I think this is right for the people. You know, I support a national patients' bill of rights, Mr. Vice President. And I want all people covered. I don't want the law to supersede good law like we've got in Texas." [Governor Bush, 10/17/00]...PRESIDENT BUSH'S ADMINISTRATION ARGUES AGAINST RIGHT TO SUE "To let two Texas consumers, Juan Davila and Ruby R. Calad, sue their managed-care companies for wrongful denials of medical benefits 'would be to completely undermine' federal law regulating employee benefits, Assistant Solicitor General James A. Feldman said at oral argument March 23. Moreover, the administration's brief attacked the policy rationale for Texas's law, which is similar to statutes on the books in nine other states." [Washington Post, 4/5/04]
3. Tobacco Buyout
BUSH SUPPORTS CURRENT TOBACCO FARMERS' QUOTA SYSTEM... "They've got the quota system in place -- the allotment system -- and I don't think that needs to be changed." [President Bush, 5/04]...BUSH ADMINISTRATION WILL SUPPORT FEDERAL BUYOUT OF TOBACCO QUOTAS "The administration is open to a buyout." [White House spokeswoman Jeanie Mamo, 6/18/04]
4. North Korea
BUSH WILL NOT OFFER NUCLEAR NORTH KOREA INCENTIVES TO DISARM... "We developed a bold approach under which, if the North addressed our long-standing concerns, the United States was prepared to take important steps that would have significantly improved the lives of the North Korean people. Now that North Korea's covert nuclear weapons program has come to light, we are unable to pursue this approach." [President's Statement, 11/15/02]...BUSH ADMINISTRATION OFFERS NORTH KOREA INCENTIVES TO DISARM"Well, we will work to take steps to ease their political and economic isolation. So there would be -- what you would see would be some provisional or temporary proposals that would only lead to lasting benefit after North Korea dismantles
"Better to have never served at all, than to quit like a coward after four months."
The truth is that hundreds of civilians and many a soldier from Spain have died because we supported the US, however not a single US solider has ever died for supporting Spain. And with "thankful" comments like the ones uttered by the previous poster, I sincerelly hope not a single drop of Spanish blood is ever wasted for America ever again.
For a preson coming from a country than one hundred years ago declared war on us, and than fifty years ago decided to starve us to death you certainly have some brass balls asking our soldiers to die for your comfy little behind.
Next time you go and pick up the gun and you get your butt over there... is this what the "land of the brave, home of the free" has de-evolved into? "better to have never served at all" LOL!
Nope French Cartoonist relates to the procedence of the author, not his/her works. If he was a cartoonist in French, then your half assed retort would have had some basis:)
True, the 601 was a hybrid of sorts. It implemented most of the 32-bit PowerPC ISA, plus most of POWER. The main difference was that the 601 did not implement the POWER cache control instructions.
The POWER instructions which are not part of the PowerPC instruction set are:
The POWER MQ register is used in 64 bit multiplication and division: these instructions write to two destination registers, and this feature is not supported in PowerPC. The 601 chip, however, supports all of the above instructions (except cache control) and it even provides the MQ register.
PowerPC is a subset of POWER, in the sense that PowerPC lacks some of the POWER instructions, not the other way around. And no instructions are not emulated using microcode since both are RISC machines and have thus no microcode!
IBM uses PowerPC on some AIX boxen, and are able to run POWER binaries by emulating some of the instructions via OS calls. Pretty simple stuff no need for microcode at all... just translate a POWER instruction into n equivalent PowerPC instructions. And voila...
Newsflash for you: there ain't much friction out there in Space:) so the solar sail will not slow down much. Particles may hit it, but the density is not enough to counteract the momentum the sail will gather at the beginning of the trip.
That could have been a VAX (probably an 11-780 for the vintage you described). The VAX was able to run PDP-11 code, in fact the name of the VAX was VAX-11 as in a virtual address extension of the PDP-11.
I believe Tandem's Non-stop OS and Stratus' FTX had the five 9s thing going for them. Although I am unaware that there was an actual certification process.
SGI, CRAY or whatever offers you a single memory image, which is a VERY different thing than a separate image cluster of commodity parts. Two very different things.
Oh, and not every cluster out there uses MPI BTW.
But by all means, you know everything there is to about computer architecture, so I guess we should just constrain ourselves to clusters of Opterons... Jeez, I guess we should have to make due with a dual 486 and call it a day. Afterall, what is the point... right?
Slashdot Mirror
Obviously you are new to the ways of the woman genre
Actually Intel is one of the few companies to implement the full IEEE FP standard, most other processors do not. The 80 bit is for internal datapaths for the multiplier and stuff, still you get 64bit doubles out.... :)
Not really, the P6 OO was pretty limited, Alphas, POWER3s, and the R10K were much much much more aggressive on the out of order execution.
Oh, and supers had pretty aggresive modified OO and Tomasulo's like units way before the P6, and they did it FAST too.
So what was your point?
Not really, endianess has nothing to do really... the only thing wrong with x86 is just one thing: REGISTERS!!!!
And if by RISC you meant microcode...
Well the general idea for RISC according to Paterson came from having to program the microcode in the VAXen, so at some point they figured out that the translation overhead from the CISC instructions was unecesary. CISC came to be as a sort of "compresion" mechanism in which one instruciton could be fetched from memory and then it can be mapped into a sequence of micro-ops inside the CPU. Thus the intruction memory space and bandwidth requirements were substatial (afterall memory was a premium back then). People figured out that memory may get larger, so why not just fetch the microops themselves, so it was a trade off between memory footprint and decoding complexity.
:).
Of course early designs like the CDC 6600 were pretty RISC, although they didn't know it. And the CISC came out as a solution to the memory footprint problems that such architectures poised (in scaling down from supers to minis where cost was a priority).
However it is really eye opening to know the actual translation overhead, control related with the translator in the x86 family is significant. And almost half of the dynamic transistor budget (not SRAM) goes into the translation/scheduling portion not the functional. So the overhead is much more significan than...
And yes their approach is microprogramming, except that RISC-core is more buzzword compliant
They were late? As in offering one of the first lines of RISC workstations on the market is being "late" jeez! (The DECStations came out in '88, which I believe places them almost earlier than the PA-RISC, or even SPARCStations).
When it came out Alpha was faster than any other CPU period, and not just faster but significantly more powerful than anything else in the market. Obviously you must have a very selective memory if you don't remember the early 90's and everyone lusting after Alpha big time. And that was a clear advantage, however DEC could not market themselves even if they were a free whore in a port.
" Decoding x86 is a major problem. See how many different solutions have been employed to do it since the PPro and K5.
* x86 chips are designed to the core to run x86. The fact that it breaks x86 into smaller instructions doesn't make it an Alpha beind a x86 decoder. It's like saying PPC970 is an Alpha behind a PowerPC decoder."
I have newsflash for you, x86 instructions have never been executed natively (most of them anyways), it is all microcode. This whole shebang of x86 "risc" core is from people who don't get it that all that the P6 is doing is using a leaner microcode that can be pipelined and superscaled.
The influence of the Alpha in the x86 comes from the fact that Intel did indeed buy out the Alpha technology after they settled their lawsuit with DEC, and a shitload of the Alpha team ended up developing for intel. Thus a lot of the internal pipeline technology ended up in the PIV et al... it doesn't make it an alpha, but some of the design principles are there.
Actually Sequent (now part of IBM) has been offering large SMP x86 machines from the mid 80's. I remember seeing a 16 processor 386 based machine in the late 80s. Also their new NUMA-Q architecture used Xeon's and scaled up to 64 processors. Also Data General moved their Aviion line from motorola 88K's to Xeons and could also scale up to 32 processors using their own NUMA architecture. Compaq has also offered 8-way x86 machines for a long time too...
Actually it is VAXen, as in FOXen...
The R in RISC is for complexity not number.. in the opposite as the C in CISC. Has nothing to do with how many Opcodes are present in your ISA.
RISC just implies that each instruction is easy to decode and can be geared towards achieving an effective single clock execution rate. Whether you have 1 or 2000 instructions in the ISA is irrelevant as long as those instructions are pretty simple.
This whole CISC being RISCy is not that innovative though, all that Intel and AMD have done is pipeline their microcode. In fact RISC can be viewed as the microcode for the CISC machines of yore, which is in fact how it came to be.
Sorry Out of Order execution has been done for ages before Intel implemented it. As usual different name, same concept. But the same ideas behind RISC, superscalar, out-of-order, pipelined, etc.. have been around (and implemented) since the 60's.
Not even in micros, as I believe Metaflow and other vendors had out-of-order CPUS out there way before Intel released the P6 microarchitecture.
Intel's FP has been 64bit IEEE for ages btw...
Actually the firs series of big Alpha systems (7000) were a VAX with an alpha CPU. So the parent wasn't that retarded afterall.... so believe it or not your apple ][ analogy wasn't a joke but a real description of what the early alpha servers were. /EOL
The point being that just because your guy believes strongly in issues doesn't necessarily make him the right person for the job. But you like him, and by all means vote for him. Such is the nature of a republic/democracy whatever the system it is that we have.
Again you want a dictator not a president. Like a much wiser guy than me once said: "To err is human, to truly correct is divine." A person that is asked if he can think of any mistake he has done while in office, and can not think of anything he's done wrong is not the right person for the job. That is why Bush is fatally flawed as a president. You however see that as a strength, which is why I am f$%king scared of this country right now.
"On the one hand, we have Bush; a guy who (most of the time) has a clear agenda, strong beliefs, and tends to stick to them (whether we like them or not). The end result is that he may do some things that we don't like, simply because he believes in them."
...BUSH SPENDS SOCIAL SECURITY SURPLUS The New York Times reported that "the president's new budget uses Social Security surpluses to pay for other programs every year through 2013, ultimately diverting more than $1.4 trillion in Social Security funds to other purposes." [The New York Times, 2/6/02]
...CANDIDATE BUSH PRAISES TEXAS PATIENTS' RIGHT TO SUE... "We're one of the first states that said you can sue an HMO for denying you proper coverage... It's time for our nation to come together and do what's right for the people. And I think this is right for the people. You know, I support a national patients' bill of rights, Mr. Vice President. And I want all people covered. I don't want the law to supersede good law like we've got in Texas." [Governor Bush, 10/17/00] ...PRESIDENT BUSH'S ADMINISTRATION ARGUES AGAINST RIGHT TO SUE "To let two Texas consumers, Juan Davila and Ruby R. Calad, sue their managed-care companies for wrongful denials of medical benefits 'would be to completely undermine' federal law regulating employee benefits, Assistant Solicitor General James A. Feldman said at oral argument March 23. Moreover, the administration's brief attacked the policy rationale for Texas's law, which is similar to statutes on the books in nine other states." [Washington Post, 4/5/04]
...BUSH ADMINISTRATION WILL SUPPORT FEDERAL BUYOUT OF TOBACCO QUOTAS "The administration is open to a buyout." [White House spokeswoman Jeanie Mamo, 6/18/04]
...BUSH ADMINISTRATION OFFERS NORTH KOREA INCENTIVES TO DISARM"Well, we will work to take steps to ease their political and economic isolation. So there would be -- what you would see would be some provisional or temporary proposals that would only lead to lasting benefit after North Korea dismantles
Well, Hitler too did things because he had "clear" beliefs... and he stuck to them. So what is your point? He is not the president to stick to his beliefs, but to work for the American people.
According to you then, a lemming would be the ideal democratic leader, right? I have newsflash for you a democracy means that there are many different beliefs, a president has to be able to accomodate most of them, w/o having to impose his own. You want a dictator, not a president...
And please don't come up with the whole "moral clarity vs. flip/flopping" crap, not like Dubya has ever flip/flopped, no siree bob:
1. Social Security Surplus
BUSH PLEDGES NOT TO TOUCH SOCIAL SECURITY SURPLUS... "We're going to keep the promise of Social Security and keep the government from raiding the Social Security surplus." [President Bush, 3/3/01]
2. Patient's Right to Sue
GOVERNOR BUSH VETOES PATIENTS' RIGHT TO SUE... "Despite his campaign rhetoric in favor of a patients' bill of rights, Bush fought such a bill tooth and nail as Texas governor, vetoing a bill coauthored by Republican state Rep. John Smithee in 1995. He... constantly opposed a patient's right to sue an HMO over coverage denied that resulted in adverse health effects." [Salon, 2/7/01]
3. Tobacco Buyout
BUSH SUPPORTS CURRENT TOBACCO FARMERS' QUOTA SYSTEM... "They've got the quota system in place -- the allotment system -- and I don't think that needs to be changed." [President Bush, 5/04]
4. North Korea
BUSH WILL NOT OFFER NUCLEAR NORTH KOREA INCENTIVES TO DISARM... "We developed a bold approach under which, if the North addressed our long-standing concerns, the United States was prepared to take important steps that would have significantly improved the lives of the North Korean people. Now that North Korea's covert nuclear weapons program has come to light, we are unable to pursue this approach." [President's Statement, 11/15/02]
"Better to have never served at all, than to quit like a coward after four months."
The truth is that hundreds of civilians and many a soldier from Spain have died because we supported the US, however not a single US solider has ever died for supporting Spain. And with "thankful" comments like the ones uttered by the previous poster, I sincerelly hope not a single drop of Spanish blood is ever wasted for America ever again.
For a preson coming from a country than one hundred years ago declared war on us, and than fifty years ago decided to starve us to death you certainly have some brass balls asking our soldiers to die for your comfy little behind.
Next time you go and pick up the gun and you get your butt over there... is this what the "land of the brave, home of the free" has de-evolved into? "better to have never served at all" LOL!
Nope French Cartoonist relates to the procedence of the author, not his/her works. If he was a cartoonist in French, then your half assed retort would have had some basis :)
Where can you get a left hand drive skyline my friend?
True, the 601 was a hybrid of sorts. It implemented most of the 32-bit PowerPC ISA, plus most of POWER. The main difference was that the 601 did not implement the POWER cache control instructions.
The POWER instructions which are not part of the PowerPC instruction set are:
cache control: clcs, clf, cli, dclst
arithmetic: abs, doz, dozi, nabs
MQ register: div, divs, mul, sleq, sliq, slliq, sllq, slq, sraiq, sraq, sreq, sriq, srilq,srlq
others: lscbx, maskg, maskir, mfsri, rac, rlmi, rrib, sle, sre
The POWER MQ register is used in 64 bit multiplication and division: these instructions write to two destination registers, and this feature is not supported in PowerPC. The 601 chip, however, supports all of the above instructions (except cache control) and it even provides the MQ register.
No... no... no...
PowerPC is a subset of POWER, in the sense that PowerPC lacks some of the POWER instructions, not the other way around. And no instructions are not emulated using microcode since both are RISC machines and have thus no microcode!
IBM uses PowerPC on some AIX boxen, and are able to run POWER binaries by emulating some of the instructions via OS calls. Pretty simple stuff no need for microcode at all... just translate a POWER instruction into n equivalent PowerPC instructions. And voila...
Newsflash for you: there ain't much friction out there in Space :) so the solar sail will not slow down much. Particles may hit it, but the density is not enough to counteract the momentum the sail will gather at the beginning of the trip.
That could have been a VAX (probably an 11-780 for the vintage you described). The VAX was able to run PDP-11 code, in fact the name of the VAX was VAX-11 as in a virtual address extension of the PDP-11.
I believe Tandem's Non-stop OS and Stratus' FTX had the five 9s thing going for them. Although I am unaware that there was an actual certification process.
SGI, CRAY or whatever offers you a single memory image, which is a VERY different thing than a separate image cluster of commodity parts. Two very different things.
Oh, and not every cluster out there uses MPI BTW.
But by all means, you know everything there is to about computer architecture, so I guess we should just constrain ourselves to clusters of Opterons... Jeez, I guess we should have to make due with a dual 486 and call it a day. Afterall, what is the point... right?
.....crickets........
... which is why Yahoo is using Google as their search engine? :)