Jihad needs scientists
in [Design]
|
From: krw on 13 Mar 2007 13:38 In article <45F644B7.6A6487DD(a)hotmail.com>, rabbitsfriendsandrelations(a)hotmail.com says... > > > "nonsense(a)unsettled.com" wrote: > > > MassiveProng wrote: > > >kensmith(a)green.rahul.net Gave us: > > > > > > Tell us, oh masterTARD, what would the maximum clock be on a 40 volt > > > logic swing. > > > > > > Do you even know what slew rate is? > > > > > > The reason it was 5 volts is because it was a reasonable voltage > > > that could be slewed to at a decent rate. > > > > > > NOW, we are at 3.3 volts and even 1.2V. The reason is slew rate, > > > and the fact that we can transition much faster at those swings than > > > we ever could at 5V. > > > > Here's a clue for you. High clock rates and complex > > high density chips have a significant problem with > > heat, the main reason for the ever lowering voltages > > in CPU's. > > *One* of the reasons. Clearly slew rate is implicated equally. Wrong again, dumb donkey. CMOS slews faster at higher voltages. Power is squared (at least) as voltage increases. The reason voltage is being forced down is precisely because of power. The reason power supply voltages aren't falling faster is because speed is important too. -- Keith
From: krw on 13 Mar 2007 13:38 In article <t5khc4-tr.ln1(a)sirius.tg00suus7038.net>, ewill(a)sirius.tg00suus7038.net says... > In sci.physics, MassiveProng > <MassiveProng(a)thebarattheendoftheuniverse.org> > wrote > on Mon, 12 Mar 2007 16:52:21 -0700 > <3mpbv2h1ts0jed1imfmdt5h52ukk3bcskj(a)4ax.com>: > > On Mon, 12 Mar 2007 14:34:47 +0000 (UTC), kensmith(a)green.rahul.net > > (Ken Smith) Gave us: > > > >>In article <45F4CF7A.BD6428AD(a)hotmail.com>, > >>Eeyore <rabbitsfriendsandrelations(a)hotmail.com> wrote: > >>[....] > >>>So Mr Expert. Why isn't TTL made on a 40 Volt process ? > >> > >>Thats obvious. Its so there is a market for MOSFET drivers. I still want > >>a PIC made with Supertex's HV CMOS. > >> > > > > > > Tell us, oh masterTARD, what would the maximum clock be on a 40 volt > > logic swing. > > > > Do you even know what slew rate is? > > > > The reason it was 5 volts is because it was a reasonable voltage > > that could be slewed to at a decent rate. > > > > NOW, we are at 3.3 volts and even 1.2V. The reason is slew rate, > > and the fact that we can transition much faster at those swings than > > we ever could at 5V. > > > > There would be no GHz+ Pentiums if we were still at 5 Volt logic > > levels. > > > > Getteth thyself a clue. > > Well, I for one was under the opinion that the reason E > = 1.2V is because P = E^2/R and also proportional to the > switching frequency of the transistors (since each switch > requires a small pulse of current; the more pulses, the > higher the power required to switch), and also proportional > to the total transistor area, which AFAIK has been largely > constant even as we pack more transistors per die. Dynamic power is proportional to CFV^2 (R doesn't figure into the equation given that the circuit switches). You're missing static power though. Static power (leakage) gets significant with smaller geometries. > Assuming one can run a chip at both 1.2V and 5V at the > same clock rate, the 5V running will run far hotter -- > about 17.4 x more power, in fact. At the same clock, yes (5^2/1.2^2). > Then again, there's a fair number of factors here, not the > least of which is process control. :-) A chip, like an > airplane, is a bunch of compromises, one of them being > clock speed versus power dissipation. Logic delay vs. power dissipation, really. Clock speed gets into microarchitectural issues as well. For example, a P4 will run at a higher clock speed than a P3 in the same technology. The P3 will likely outperform it at the same power though. > The x86 series > is an excellent example of a total bodge-up because > it compromised so many things (for example, it's still > source-code compatible with the 8080A and probably with > the Z80 as well!). Source code compatibility means nothing here. > But it still works. > > Now, assuming anyone wants an HV CMOS 40V logic swing, I > for one would think that the best method of achieving such > is through a 1.2V-to-40V swing converter, which presumably > would be a carefully-adjusted standard CMOS transistor-pair > inverter which flips at about 0.6 volts to ground. The rest > of the circuitry can run at 1.2 V without trouble. Not sure why you'd want 40V, but you're not going to get it on the same hunk of silicon. > If necessary the level shifter can feed another, more standard inverter. > > There might be better solutions; it's been a long time since I've been > anywhere near semiconductor designs, and I only really worked with > slow-speed digital stuff at most. -- Keith
From: krw on 13 Mar 2007 13:38 In article <et5uie$8ss_001(a)s887.apx1.sbo.ma.dialup.rcn.com>, jmfbahciv(a)aol.com says... > In article <3f1c7$45f554cb$4fe7735$10594(a)DIALUPUSA.NET>, > "nonsense(a)unsettled.com" <nonsense(a)unsettled.com> wrote: > >jmfbahciv(a)aol.com wrote: > <snip> > > >> You have noted that he stripped my post to make it appear that > >> I was agreeing with his factoid. I was talking about something > >> completely different. > > > >Happens all the time. > > > >These are the people who when asked "What is pi" will argue to > >death that it is 3.14. > > > ># the ratio of the circumference to the diameter of a circle; > >approximately equal to 3.14159265358979323846... > ># private detective: someone who can be employed as a detective to > >collect information > ># principal investigator: the scientist in charge of an experiment or > >research project > ># the 16th letter of the Greek alphabet > ># protease inhibitor: an antiviral drug used against HIV; interrupts HIV > >replication by binding and blocking HIV protease; often used in > >combination with other drugs > > > >wordnet.princeton.edu/perl/webwn > > In my corner of the biz, it would have been read as priority interrupt. "Primary Input". ;-) -- Keith
From: krw on 13 Mar 2007 14:12 In article <et68m3$t53$4(a)blue.rahul.net>, kensmith(a)green.rahul.net says... > In article <et5vdg$8qk_002(a)s887.apx1.sbo.ma.dialup.rcn.com>, > <jmfbahciv(a)aol.com> wrote: > >In article <et3o1m$rad$2(a)blue.rahul.net>, > > kensmith(a)green.rahul.net (Ken Smith) wrote: > [....] > >>There need not be any increase in the risk. Its all a matter of starting > >>with a reasonable OS and not adding buffer over runs. > > > >There will always be buffer overruns. > > Only in badly written software using languages without run time checking > does it happen. We now have more than enough CPU speed that buffer > overruns should be a thing of the past. What percentage of software over the last decade has been written in C/C++? How much in, say, Ada? > > The risk of a single system > >house site configuration is the lack of redundancy. That is > >what makes it dangerous. The danger is not just viral infections > >but with users typing at a system which also controls vital > >functions within the house. > > I would never network my PC to my pacemaker. I wouldn't connect mine to my toaster. Toast is simply too important to trust to a PeeCee and M$ (can't drink coffee anymore, so toast will have to do here). > That said, you are assuming that the user runs as root/superuser. I very > rarely ever do on my home system. I do it a little more often on the work > system but that is because I need to directly fiddle a few hardware things > from time to time. M$ almost insists that users run as admin. Unix and variants are the opposite. What's the relative population of each group? -- Keith
From: Ken Smith on 13 Mar 2007 21:59
In article <MPG.2060b389becb9ca598a106(a)news.individual.net>, krw <krw(a)att.bizzzz> wrote: >In article <et68m3$t53$4(a)blue.rahul.net>, kensmith(a)green.rahul.net >says... >> In article <et5vdg$8qk_002(a)s887.apx1.sbo.ma.dialup.rcn.com>, >> <jmfbahciv(a)aol.com> wrote: >> >In article <et3o1m$rad$2(a)blue.rahul.net>, >> > kensmith(a)green.rahul.net (Ken Smith) wrote: >> [....] >> >>There need not be any increase in the risk. Its all a matter of starting >> >>with a reasonable OS and not adding buffer over runs. >> > >> >There will always be buffer overruns. >> >> Only in badly written software using languages without run time checking >> does it happen. We now have more than enough CPU speed that buffer >> overruns should be a thing of the past. > >What percentage of software over the last decade has been written in >C/C++? How much in, say, Ada? How mush is "badly written"? Most software is truly horrid. I still have a hope that this may change. It may take the deaths of a hundred cute puppies live on TV, due to a bug, to get the point. >> I would never network my PC to my pacemaker. > >I wouldn't connect mine to my toaster. Toast is simply too important Yes, and with real peanut butter, it is even better. >to trust to a PeeCee and M$ (can't drink coffee anymore, so toast >will have to do here). I wouldn't trust M$ for anything. >> That said, you are assuming that the user runs as root/superuser. I very >> rarely ever do on my home system. I do it a little more often on the work >> system but that is because I need to directly fiddle a few hardware things >> from time to time. > >M$ almost insists that users run as admin. Unix and variants are the >opposite. What's the relative population of each group? With any luck, Vista will be the death of M$ and every home user will switch to Apples. Apple should get the customers just for those great ads. -- -- kensmith(a)rahul.net forging knowledge |