So if in binary you have 1 bit, 2 bits = 1 nibble, 2 nibbles = 1 byte, does that mean in terary you have 1 trit, 3 trits = 1 tribble, 3 tribbles = 1 tribe?
Only because you don't know the proper statutes and regulations to make your position legitimate, but instead choose to perjure yourself everytime you sign a 1040.
"Hydrogen combining with oxygen is an *EXOTERMIC* reaction. It also makes a big explosion."
Not always. At hell, I mean, Intel Fabs, we flow O2, heat it up with an SiC lamp to 650 C, then flow H2 and end up with steam for one of the processes in our diffusion furnaces.
Hate to brake this to you, but not only is that NOT Common Law (which the SC has said does not exist in the US in any event), but precedents aren't even legal. The Constitution, USC (and USCS and USCA) as well as the CFR all say that precendents are not to be used.
". Current Silicon technology uses 12" wafers"
Well, sort of. Almost. D1B is Intel's only Fab that's using 12", all the others are 8". Fab22 is designed for 12", but we're still stacking it for 8" and won't make the switch up to 12" for a couple years.
AFAIK, TSMC is the only other foundary moving to 12". Everyone else seems to be taking a "wait and see" attitude. We'll get there eventually.
Anyway, the point is that even though 12" is possible and even in extremely limited production, it's just that, experimental right now. 8" is standard.
IBM came to my company asking us to build them a furnace that could handle Arsenic. We said no, specifically because of the hazards involved in working with Arsenic-contaminated equipment.
Um, the point of an aptitude test is to see what you know, not see whether or not you know how to use a calculator or an encyclopedia. Of course someone outfitted with a computer would do better on the SAT than someone without. It's called MOTO, KAWG: Master of the obvious, knows answer when given.
For the last time, Amiga, under C= never had any intention of going PPC.
It's nice of you to praise Phase5, since they were instrumental in getting Amiga into its current sorry state, due to the Kernal War with H&P. Pretty impressive that that little incident cost them a couple hundred thousand sales of PPC cards.
And lets not forget the German liquidators who stonewalled QuickPak for over a year and ended up selling to GW. If it wasn't for that little maneuver, we would have DEC Alpha (I refuse to say Compaq or Intel) Amigas now, as was there well documented plan.
Re: "Only 833MHz"?... Itanium runs at 800MHz...
on
Alpha Up For Grabs?
·
· Score: 1
I couldn't remember what speed Itanium was running at. I assumed (thus making an ass of myself) that it was in the 1.5GHz range.
Also remember, we're talking about a 5 year old chip here vs something "brand new."
Try this on for size. Take the EV6 core, remove the 128KB conventional L1 cache (9 mil trans), now we're down to 6 mil trans. Have MoSys add 128KB 1T-SRAM (1.1 mil trans) L1 and 1MB (8.8 mil trans) L2. Now we're back up to 15.8 mil (.6 mil higher than we started with) and have gobs of super hi speed, low latency cache. Now we fab it in IBMs CMOS-9S in a 9-layer process at.13um. We've got an 82mm2 chip now, probably running at ~1.6GHz.
The EV6 Bus scales to 400MHz. It's sitting at 133 right now. So make the jump to 400MHz. Alpha is used to having 128bit or 256bit memory subsystems, right? So you make a new memory standard and screw JEDEC. Samsung just released 300MHz DDR SDRAM, they'll hit 400MHz in the next 1-2 years. This stuff is normally used for video cards, right? Ah, but it's intended for 128- or 256bit systems and has a hi-speed serial connect, perfect for our needs. All you need to do is design a new memory controller (which a friend of mine is doing) for this.
That'd make for a pretty bloomin' fast h/w setup. But it's not h/w alone that make a computer tic, you still need s/w. So pick an OS, like AtheOS (which would need to be ported, but it's a pretty small OS still) that has all the features you want and go to town.
But as someone else was kind enough to point out, I'm not in charge.
Don't be stupid. As a friendly Amiga person myself, I'm telling you this in a nice way.
Alpha is not a company, it is a product line. They therefore could not have sold themselves. Compaq, a company, bought DEC (and if you knew anything about corporate law, you would know that it would've been very difficult for DEC to say no). Compaq is now selling it off to Intel, because Compaq doesn't have any balls and refuses to market Alpha. Instead, they strive for maximum profits/sale instead of striving for market proliferation. That's the fist step towards killing a product line.
You're kidding, right? The original Alpha team was awesome. The Series 7 bunch that Compaq had to hire because everyone else left for AMD after the DEC buyout hasn't put out a chip in 5 years. That's pathetic.
Don't get me wrong, I love Alphas and this is bad news. But the guys they have there now are incompetent.
No. RISC is a philosophy. It is Reduced Instruction Set COMPLEXITY. You reduce the complexity of a given instruction, even if it means adding more instructions over all. Hense the terms Load/Store vs MOV. The only thing RISCy about x86 is adding L2 cache and FPU to the core, and that's only due to advances in fab technology. "Modern" CISC has nothing else in common with RISC.
Really, what did you expect? The series 7 ('364) is designed by people that have no clue WTF they're doing. It's a '264 core with a Rambus controller on it, not much more. The '264 came out in '96, so 5 years later and they still can't release a new product? I smell incompetence.
Ahh, the series 8 ('464). Now *that's* a pretty chip. 8 instructions/cycle issued, something like 256 in flight at once, multi-threaded core. Tick for tick it has twice the performance of the '264, and it runs at twice the clock. Too bad they won't fab it in state of the art processes.
Yes, Alpha still knocks the pants off of Itanic. The embarassing part about it for Intel is that the Alpha is doing it at "just" 833MHz. How "fast" was the Itanic going again?
"Incidently (I know it was mentioned in passing be another poster) Intel bought the alpha FAB from DEC, all current aplha chips are manufactured by Intel, the latest Alphas were largley designed by Intel engineers, it makes sense for Intel to own the archirecture outright."
Uhh, no. Alpha chips are currently fabbed by Intel, IBM and Samsung. Intel stole design secrets in the current (EV6) Alpha, not developed them.
Hunh?! PowerPC is derived from Power, not the other way around. One of the reasons IBM didn't want to jump on the AltiVec bandwagon is because Power already has a phenomenally powerful FPU. They sell Power in the high end (as you said, AS/400s and RS/6000s) and PowerPC in the low end, where they don't need FPU strength as much. Adding AV was counter-productive as far as they were concerned.
It's amusing that you mention Alpha for a NG console. DEC was actually working on an Alpha based console/arcade box (think of it as being similar to the DC/NAOMI relationship). Things were going well and even had titles lined up. See the game Dronez by Zetha Games? Originally designed for this Alpha platform. Then Compaq bought out DEC. Yet another chance for Alpha to be in the home destroyed.
Alphas are power hungry only because of the tech level used to fab them,.25um. If they fabbed it in.13um (not even using copper or anything else special) a 500MHz Alpha would consume 21W. It would also cost significantly less to fab than the EE does.
StrongARM (or XScale, it's successor) would be a much worse choice than Alpha. All it has going for it is power consumpiton efficiency. Alpha crushes it in interger perfomance, and StrongARM doesn't even have an FPU.
Slashdot Mirror
So if in binary you have 1 bit, 2 bits = 1 nibble, 2 nibbles = 1 byte, does that mean in terary you have 1 trit, 3 trits = 1 tribble, 3 tribbles = 1 tribe?
Only because you don't know the proper statutes and regulations to make your position legitimate, but instead choose to perjure yourself everytime you sign a 1040.
"Hydrogen combining with oxygen is an *EXOTERMIC* reaction. It also makes a big explosion."
Not always. At hell, I mean, Intel Fabs, we flow O2, heat it up with an SiC lamp to 650 C, then flow H2 and end up with steam for one of the processes in our diffusion furnaces.
Hate to brake this to you, but not only is that NOT Common Law (which the SC has said does not exist in the US in any event), but precedents aren't even legal. The Constitution, USC (and USCS and USCA) as well as the CFR all say that precendents are not to be used.
Title 26, United States Code, Section 861, Subchapter N.
Quit being ignorant and take the time to educate yourself.
(knowing that you're not talking monetary)
Did you just say that free software is a natural right? You're kidding, right?
http://www.larryelder.com/Gore/goredubiousrecord.h tm
". Current Silicon technology uses 12" wafers" Well, sort of. Almost. D1B is Intel's only Fab that's using 12", all the others are 8". Fab22 is designed for 12", but we're still stacking it for 8" and won't make the switch up to 12" for a couple years. AFAIK, TSMC is the only other foundary moving to 12". Everyone else seems to be taking a "wait and see" attitude. We'll get there eventually. Anyway, the point is that even though 12" is possible and even in extremely limited production, it's just that, experimental right now. 8" is standard.
IBM came to my company asking us to build them a furnace that could handle Arsenic. We said no, specifically because of the hazards involved in working with Arsenic-contaminated equipment.
Um, the point of an aptitude test is to see what you know, not see whether or not you know how to use a calculator or an encyclopedia. Of course someone outfitted with a computer would do better on the SAT than someone without. It's called MOTO, KAWG: Master of the obvious, knows answer when given.
That's right, not only is this a class warfare issue, but it's a racism one as well. Sheesh.
The "900MHz" is referring to USIII, not Itanium.
I'd rather have a couple RD5000 RenderDrives from www.art.co.uk myself, even if they are a little pricy.
For the last time, Amiga, under C= never had any intention of going PPC.
It's nice of you to praise Phase5, since they were instrumental in getting Amiga into its current sorry state, due to the Kernal War with H&P. Pretty impressive that that little incident cost them a couple hundred thousand sales of PPC cards.
And lets not forget the German liquidators who stonewalled QuickPak for over a year and ended up selling to GW. If it wasn't for that little maneuver, we would have DEC Alpha (I refuse to say Compaq or Intel) Amigas now, as was there well documented plan.
I couldn't remember what speed Itanium was running at. I assumed (thus making an ass of myself) that it was in the 1.5GHz range.
.13um. We've got an 82mm2 chip now, probably running at ~1.6GHz.
Also remember, we're talking about a 5 year old chip here vs something "brand new."
Try this on for size. Take the EV6 core, remove the 128KB conventional L1 cache (9 mil trans), now we're down to 6 mil trans. Have MoSys add 128KB 1T-SRAM (1.1 mil trans) L1 and 1MB (8.8 mil trans) L2. Now we're back up to 15.8 mil (.6 mil higher than we started with) and have gobs of super hi speed, low latency cache. Now we fab it in IBMs CMOS-9S in a 9-layer process at
The EV6 Bus scales to 400MHz. It's sitting at 133 right now. So make the jump to 400MHz. Alpha is used to having 128bit or 256bit memory subsystems, right? So you make a new memory standard and screw JEDEC. Samsung just released 300MHz DDR SDRAM, they'll hit 400MHz in the next 1-2 years. This stuff is normally used for video cards, right? Ah, but it's intended for 128- or 256bit systems and has a hi-speed serial connect, perfect for our needs. All you need to do is design a new memory controller (which a friend of mine is doing) for this.
That'd make for a pretty bloomin' fast h/w setup. But it's not h/w alone that make a computer tic, you still need s/w. So pick an OS, like AtheOS (which would need to be ported, but it's a pretty small OS still) that has all the features you want and go to town.
But as someone else was kind enough to point out, I'm not in charge.
yes, and the concept of the micro-op is the antithesis of the entire RISC philosophy.
Don't be stupid. As a friendly Amiga person myself, I'm telling you this in a nice way. Alpha is not a company, it is a product line. They therefore could not have sold themselves. Compaq, a company, bought DEC (and if you knew anything about corporate law, you would know that it would've been very difficult for DEC to say no). Compaq is now selling it off to Intel, because Compaq doesn't have any balls and refuses to market Alpha. Instead, they strive for maximum profits/sale instead of striving for market proliferation. That's the fist step towards killing a product line.
"some of the best in the industry"
You're kidding, right? The original Alpha team was awesome. The Series 7 bunch that Compaq had to hire because everyone else left for AMD after the DEC buyout hasn't put out a chip in 5 years. That's pathetic.
Don't get me wrong, I love Alphas and this is bad news. But the guys they have there now are incompetent.
No. RISC is a philosophy. It is Reduced Instruction Set COMPLEXITY. You reduce the complexity of a given instruction, even if it means adding more instructions over all. Hense the terms Load/Store vs MOV. The only thing RISCy about x86 is adding L2 cache and FPU to the core, and that's only due to advances in fab technology. "Modern" CISC has nothing else in common with RISC.
Really, what did you expect? The series 7 ('364) is designed by people that have no clue WTF they're doing. It's a '264 core with a Rambus controller on it, not much more. The '264 came out in '96, so 5 years later and they still can't release a new product? I smell incompetence.
Ahh, the series 8 ('464). Now *that's* a pretty chip. 8 instructions/cycle issued, something like 256 in flight at once, multi-threaded core. Tick for tick it has twice the performance of the '264, and it runs at twice the clock. Too bad they won't fab it in state of the art processes.
Yes, Alpha still knocks the pants off of Itanic. The embarassing part about it for Intel is that the Alpha is doing it at "just" 833MHz. How "fast" was the Itanic going again?
"Incidently (I know it was mentioned in passing be another poster) Intel bought the alpha FAB from DEC, all current aplha chips are manufactured by Intel, the latest Alphas were largley designed by Intel engineers, it makes sense for Intel to own the archirecture outright."
Uhh, no. Alpha chips are currently fabbed by Intel, IBM and Samsung. Intel stole design secrets in the current (EV6) Alpha, not developed them.
Hunh?! PowerPC is derived from Power, not the other way around. One of the reasons IBM didn't want to jump on the AltiVec bandwagon is because Power already has a phenomenally powerful FPU. They sell Power in the high end (as you said, AS/400s and RS/6000s) and PowerPC in the low end, where they don't need FPU strength as much. Adding AV was counter-productive as far as they were concerned.
It's amusing that you mention Alpha for a NG console. DEC was actually working on an Alpha based console/arcade box (think of it as being similar to the DC/NAOMI relationship). Things were going well and even had titles lined up. See the game Dronez by Zetha Games? Originally designed for this Alpha platform. Then Compaq bought out DEC. Yet another chance for Alpha to be in the home destroyed.
Alphas are power hungry only because of the tech level used to fab them, .25um. If they fabbed it in .13um (not even using copper or anything else special) a 500MHz Alpha would consume 21W. It would also cost significantly less to fab than the EE does.
StrongARM (or XScale, it's successor) would be a much worse choice than Alpha. All it has going for it is power consumpiton efficiency. Alpha crushes it in interger perfomance, and StrongARM doesn't even have an FPU.