performance of hardware dynamic scheduling
|
From: Ken Hagan on 3 Apr 2008 06:40 On Wed, 02 Apr 2008 09:45:16 +0100, Nick Maclaren <nmm1(a)cus.cam.ac.uk> wrote: > The other thing that I and my colleagues would dearly LOVE is to be > able to buy commodity systems (and I mean systems, not CPUs), at the > same price as the current ranges, with half the performance and one > quarter the power consumption. It could be done, and would enable > cool, quiet, small, more reliable office machines - plus eliminating > many of our machine-room problems. > > Intel and others are missing a trick. Such 'green line' systems do > exist from specialist vendors, but are typically VERY pricey, and > often somewhat weird. A lot of people want what I have described, > though the marketdroids don't seem to realise it. This may change quite soon. I've just bought a new system at home. It sits in the lounge. When sizing up the options, anything with a fan was ruled out. Over the years, I've met many people (in fact, I think they are the majority) who object to mysterious whistles, hums and such like from the boxes that sit under their TV. If the industry wants everyone to put a PC in that place, it will *have* to start offering *silent* devices at *consumer* prices. (Happily, this has started to happen. Google for fanless PVR. All we need now is for the idea to spread to media PCs and from there to PCs in general.)
From: Nick Maclaren on 3 Apr 2008 06:49 In article <op.t806mi1ass38k4(a)khagan.ttx>, "Ken Hagan" <K.Hagan(a)thermoteknix.com> writes: |> |> Over the years, I've met many people (in fact, I think they are the |> majority) who object to mysterious whistles, hums and such like from |> the boxes that sit under their TV. If the industry wants everyone to |> put a PC in that place, it will *have* to start offering *silent* |> devices at *consumer* prices. (Happily, this has started to happen. |> Google for fanless PVR. All we need now is for the idea to spread to |> media PCs and from there to PCs in general.) Even with my hearing loss (c. 60 dB, and much more in the higher ranges), my Sun V20Z at work sounds like a miniature jet aircraft warming up. Apparently, the more recent models have improved that .... Fanless isn't critical, provided that you can have one large, slow fan per box. Those can be made VERY quiet. But the sort of fans that go on top of CPUs can't be. Regards, Nick Maclaren.
From: Ken Hagan on 3 Apr 2008 11:07 On Thu, 03 Apr 2008 13:44:16 +0100, Bill Todd <billtodd(a)metrocast.net> wrote: > Intel's IA64 failure had nothing whatsoever to do with not understanding > what the market wanted: it was all about failure to execute technically > (or perhaps more specifically failure to be realistic in their technical > - and schedule - projections) plus the hubris to believe that they could > foist the resulting multi-billion-dollar-yawner on the market anyway > (rather than quietly letting it die when it came up so abysmally short > of its grandiose hype). So the vendor has to guess what will be implementable just as much as they have to guess what will be desirable. They can be wrong on either front. In any case, we're not talking about a completely free market here. There's a considerable barrier to entry and only a handful of current players. If they are all making similarly over-cautious decisions, how would we know? >> The markets aren't infallible either. > > They don't have to be: they *define* demand (though are certainly often > able to be influenced in that definition). But they *define* demand only several years after the vendors have committed to what they will supply. At the time the supply decisions are being made, the punter isn't putting up any money at all. In addition, both vendor and market are probably fairly conservative. It makes little sense for vendors to bet billions of dollars on something *very* different from tried and tested, and uninformed punters will almost always ask for "more of what I have now". > I guess I must have missed the clamor, then: could you direct me to > evidence that the market is demanding that trade-off for *existing* > products (rather than anticipating that it will make possible some > interesting new ones)? That's not a reasonable request. The fact that new products are possible is part of the clamor. If you restrict your survey to those you are already selling to, then once again you'll get the answer "more of the same". For example, people looking for a fanless PVR probably don't even consider themselves as part of the market for x86 chips.
From: Nick Maclaren on 3 Apr 2008 11:29 In article <op.t81ixz1gss38k4(a)khagan.ttx>, "Ken Hagan" <K.Hagan(a)thermoteknix.com> writes: |> |> So the vendor has to guess what will be implementable just as much as they |> have to guess what will be desirable. They can be wrong on either front. And often are, especially on the former. |> In any case, we're not talking about a completely free market here. There's |> a considerable barrier to entry and only a handful of current players. If |> they are all making similarly over-cautious decisions, how would we know? We're here because we're here because we're here .... Yes. See below :-( |> >> The markets aren't infallible either. |> > |> > They don't have to be: they *define* demand (though are certainly often |> > able to be influenced in that definition). |> |> But they *define* demand only several years after the vendors have |> committed to what they will supply. At the time the supply decisions |> are being made, the punter isn't putting up any money at all. Right. But it's not true that the market defines demand, anyway, for the reasons you give above. If a category of product isn't practically available (for any reason, including excessive price), the markets will say there is no demand whether or not the customers are screaming for it. As an example, in the late 1970s, many people said that there was a major demand for workstations for program development and similar work in research environments - I was attacked quite viciously for saying so, from both the "mainframes rule" camp and the "IBM PC" one. I said then that the key price was 5,000 pounds and the key functionality was that of a simple but tolerably reliable Unix system. For over 5 years, people pointed out that the "market" said I was wrong - and then the functionality and price was delivered. The rest is history :-) |> In addition, both vendor and market are probably fairly conservative. |> It makes little sense for vendors to bet billions of dollars on something |> *very* different from tried and tested, and uninformed punters will almost |> always ask for "more of what I have now". And uninformed and unimaginative 'experts' :-( :-( :-( |> > I guess I must have missed the clamor, then: could you direct me to |> > evidence that the market is demanding that trade-off for *existing* |> > products (rather than anticipating that it will make possible some |> > interesting new ones)? |> |> That's not a reasonable request. The fact that new products are possible |> is part of the clamor. If you restrict your survey to those you are already |> selling to, then once again you'll get the answer "more of the same". |> For example, people looking for a fanless PVR probably don't even consider |> themselves as part of the market for x86 chips. In any case, he could find it if he could be bothered to look. I was perhaps the first person to use the measure performance/watt in an HPC procurement, but it isn't all that rare now. You have to read the actual procurement documents to see, as the number of terms used is legion. Regards, Nick Maclaren.
From: David Kanter on 3 Apr 2008 12:58
On Apr 3, 12:21 am, n...(a)cus.cam.ac.uk (Nick Maclaren) wrote: > In article <91142bbc-6d1d-4b9e-9d0e-e98a06ad7...(a)i36g2000prf.googlegroups.com>,David Kanter <dkan...(a)gmail.com> writes: > > |> > |> Nick, do you even know what a PLL is? It's not architecturally > |> visible... > > http://www.faqs.org/rfcs/rfc1305.html > > I don't care whether you think that RFC is abusing the term "phase-locked > loop" or not. A phase locked loop typically refers to a piece of analog or mixed signal circuitry. As far as I know, hardware PLLs (which I'll distinguish from SW PLLs - i.e. whatever NTP is using) are wholly unaffected by SMT and are not architecturally visible. > The fact is that NTP is used for almost all serious time > synchronisation on modern computers, it uses that technique, it's done in > software, and SMT plays Old Hob with it. Indeed, as a moderate expert > on the topic, I doubt that anyone could get it to work properly without > temporarily disabling that feature by one means or another. Alright, so how do you get NTP working on any of the latest PPC servers? Those ALL use SMT, clearly its a solvable problem. > And may I suggest that you take your own advice? In particular, before > posting further nonsense on this topic, find out my NTP-related record. I'm more than happy to learn about NTP, as I've never claimed to be an expert. However, in my experience I have never heard of anyone use the term PLL for anything in SW. Even wikipedia really mentions it only in relation to hardware. http://en.wikipedia.org/wiki/Phase-locked_loop David |