I can't fathom, though, why nVidia -- a graphics chipset maker which has nothing to do with the ray tracing you're describing -- would be interested in this.
The bandwidth between the CPU and the graphics chipset is a frequent bottleneck. This is why the graphics adapter often has a special slot (VLB vs. ISA; AGP vs. PCI; PCI-e x16 vs. PCI-e x1) and why we're starting to see the marriage of the CPU and graphics chipset (AMD buying ATI, nVidia talking about making their own CPU, Intel making graphics chips, etc.). Put this on the cloud, and your bandwidth is shot, and suddenly latency is a huge issue.
In general, you're right. But we're talking about ISPs here. It's their network -- of course they can get it layer 2!
Well, Comcast's network is a massive hodge podge (to put it politely). I'm not even sure the head end I (seem to be) connecting to can get my MAC address from the packet (at least if traceroute is to be believed). There's definitely at least one switch between my neighborhood's segment (in the boondocks of Puget Sound) and the rest of the service area (which goes down with alarming frequency).
Back in Pittsburgh, the architecture was quite different. They purchased the system from AT&T, who purchased it from Times Mirror, etc. Heck, I originally had two separate cable wires coming in to the cable box (switched on a 12V DC signal from the box itself); apparently, TM's switching equipment couldn't handle all the channels they wanted to offer. Not surprisingly, this had to be redone before cable modem service was offered. I had DSL service (from Bell Atlantic) years before cable modems were even offered in the area.
Argh, I messed up. They're actually using your endpoint's MAC (a DD-WRT-ified Linksys router, in my case), not the cable modem's. That makes a bit more sense, doesn't it?:-) Sorry for the confusion.
Yeah, the only stack I can see on the modem itself is a management interface on 192.168.100.1. Of course, there's the DOCSIS stuff on the cable side, but I don't have the tools (or time or inclination) to view what's going on there.
At least Comcast got the opt-out implementation right. It's done by the cable modem's MAC address, which means that all DNS lookup traffic will start getting NXDOMAIN queries. Oddly, their instructions indicate that this only takes effect when your modem does its next DHCP client lease. My guess is they've blocked off a range of IPs as "opt out," and just assign your MAC to get a lease from the out out range.
I'd greatly prefer it if Comcast had just left things alone, of course; at least, though, they didn't fall into the old "The Web is the Internet" fallacy like Bell Canada.
Understood. I was just amused at the thought of a GNU project being filed under BSD. Not quite like bringing antimatter and matter together (or emacs/vi), but the fallout would be amusing nonetheless.
Re-read it. I said, "Until they fix these issues, I'll hold on to my incandescents and _carbon arc lamps_." Nobody would seriously use carbon arc lamps these days -- that's my tongue-in-cheek way of saying, "Git off my lawn." It's humor. Apparently too subtle for/.
Where the hell did I say I wouldn't switch? My god, you folks are all humor-impaired! I said I wouldn't switch from my "incandescents and carbon arc lamps." Because, y'know, everyone is still using carbon arc lamps to shine the lights inside of their barns so they can hitch up their buggys...
You folks are just so amazingly clueless and unwilling to think, it's sad.
No, 28 VA == 28 W at 1.0 pf. Do you not understand that VA is the same units as a watt, and the notational difference is used to distinguish between real and apparent power?
And, no, your other appliances will not use the energy. You're thinking of bypass capacitors used to smooth out ripples. Appliances can't "steal" current on different phases. Where are you getting this line of reasoning?
Did the poster only ask people who haven't tried a new lamp in the last 5 years?
That'd be me. It's mixed, depending on the installation location and brand. Outdoors, in 20-30ÂF weather: 2-3 minutes to become acceptable, 5 minutes to stop becoming brighter. Indoors, 40 seconds or so for GEs, 20 seconds for Home Depot cheap-o brand (but they tend to go within a year or so). Aside from the outdoors, I've kept all the CFLs indoors or added a mix of CFL and incandescent.
Oh, and that list was not meant to be an, "OMG, CFL bad and here's why" list. More of a, "There's no such thing as a free lunch; here's what you need to consider" list. But it would seem that CFLs are a surprisingly political topic on/.
Really? Huh. Admittedly, I haven't measured it, but I had always heard that you shouldn't use fluorescents (standard or CFLs) in places where you expect the light to be on less than 30 minutes on average. I can see how this wouldn't be true for CFLs, though.
What about lifetime? Does more frequent power cycling affect their useful life?
Wow, I had no idea this would've been taken as an anti-CFL rant. Apparently neither my viewpoint nor the article's came through in the summary, which is more that there's no such thing as a free lunch (i.e. CFLs have downsides, too.). I think the headline was regrettably chosen, though, which I took from the original article and reworded to fit in/.'s character limits.
Bruce, you make a lot of good points. Yes, the mercury output is less than if you're on coal electricity (we're mostly hydro here). And unless lighting is the large majority of your electric bill (which it isn't for most households), the power factor of those CFLs aren't going to matter. I was surprised, however, to find that the power factor is as low as it is. I'll be happy once we have fewer integrated ballasts (which are produced as cheaply as possible). Spending an extra dollar on the ballast could improve the power factor and other issues significantly.
However, there's one claim that is problematic:
I only have one incandescent lamp left in my home, and that one is going to leave someday soon too. We are a healthy, happy family, and we're spending less and hurting the environment less. That's the reality for CFL users.
If you're using CFLs indiscriminately, you're applying the technology suboptimally. A rarely and briefly-used hallway light, for example, would be better served by an incandescent. The initial power surge to kick off the light will eat into your usage and savings -- very minutely, but still not the benefit you think you're getting. Which was really the point of submitting this article: we can't blindly use any technology -- CFLs, hybrid cars, wind power, etc. -- thinking that it's the perfect solution. Some thought has to be applied, and that requires information.
The original text was written for a different audience, though I thought it would work for/. as well. Guess not.:-)
To power the 13 W light bulb with a 0.45 pf, you need to generate 28 VA (plus change) of power. This same electrical output could also power a 28 W light bulb with a 1.0 pf. (All assuming negligible resistive transmission losses.) I know it's not consumed by the end user, but it is energy being put out at the plant nonetheless.
Heh, no rant to unleash. Just that CFLs aren't the panacea that they're often made out to be. (Note that I live in an area where you are often considered an $EVIL_DEITY if you don't use CFLs, shop with reusable hemp sacks, install a cistern to store rain water in the driveway underneath your hybrid biodiesel-electric-flower car, etc.)
That said, I did go out and buy a bunch of CFLs. I've found:
The generic Home Depot brand had a failure rate of ~50% within the first 6 months, regardless of installation location. (Wish I had kept the receipt.) I've only lost GE brand bulb I purchased from Ace over the past two years, and that was because I dropped it and it shattered.
Warm-up times have been a problem, moreso with the GE brand. It takes between 70-120 seconds for most of the bulbs to reach full brightness. This makes it a non-candidate in many locations (are you going to wait 2 minutes before walking down the stairs to the basement?).
Mercury disposal has been a hassle. Between the (privately-owned) garbage collection company and the various municipalities, it's difficult to get a consistent answer. Muni says it's ok to throw out in the normal trash; disposal company wants nothing to do with them (for fear of future lawsuits, I'm guessing).
The low power factor finding was surprising to me. I had naively assumed they were reasonably close to unity, say 0.8 or so. The finding that they can be as low as 0.4 was surprising.
Anyway, I thought that the claim that I'd be sticking to my carbon arc lamps would be clear that I'm making this tongue-in-cheek (it's like claiming that I commute to work in a steam powered horseless buggy); I still use CFLs, just a bit more judiciously, and like having this kind of information. Guess the average/. reader isn't as savvy as I had thought. Ah, well.
MySQL strives to provide RDBMS and ACID semantics, though its quality of service (QoS) may fall short. By contrast, these "slacker" databases don't even try to support RDBMS or ACID; even if they operated perfectly, they won't provide RDBMS/ACID.
I work for one of the companies in question (no, I don't speak for them). We rely heavily on a combination of these "slacker" dbs, Berkeley dbs, memcached, Oracle, flat files, and tape backups. Each fills a niche. I wish these articles would quit trying to create a false dichotomy.
Slashdot Mirror
I can't fathom, though, why nVidia -- a graphics chipset maker which has nothing to do with the ray tracing you're describing -- would be interested in this.
The bandwidth between the CPU and the graphics chipset is a frequent bottleneck. This is why the graphics adapter often has a special slot (VLB vs. ISA; AGP vs. PCI; PCI-e x16 vs. PCI-e x1) and why we're starting to see the marriage of the CPU and graphics chipset (AMD buying ATI, nVidia talking about making their own CPU, Intel making graphics chips, etc.). Put this on the cloud, and your bandwidth is shot, and suddenly latency is a huge issue.
Sorry; I was attempting to mod you insightful, but my trackpad decided to hit "Troll" instead. This reply exists only to undo my mismoderation.
In general, you're right. But we're talking about ISPs here. It's their network -- of course they can get it layer 2!
Well, Comcast's network is a massive hodge podge (to put it politely). I'm not even sure the head end I (seem to be) connecting to can get my MAC address from the packet (at least if traceroute is to be believed). There's definitely at least one switch between my neighborhood's segment (in the boondocks of Puget Sound) and the rest of the service area (which goes down with alarming frequency).
Back in Pittsburgh, the architecture was quite different. They purchased the system from AT&T, who purchased it from Times Mirror, etc. Heck, I originally had two separate cable wires coming in to the cable box (switched on a 12V DC signal from the box itself); apparently, TM's switching equipment couldn't handle all the channels they wanted to offer. Not surprisingly, this had to be redone before cable modem service was offered. I had DSL service (from Bell Atlantic) years before cable modems were even offered in the area.
Argh, I messed up. They're actually using your endpoint's MAC (a DD-WRT-ified Linksys router, in my case), not the cable modem's. That makes a bit more sense, doesn't it? :-) Sorry for the confusion.
Yeah, the only stack I can see on the modem itself is a management interface on 192.168.100.1. Of course, there's the DOCSIS stuff on the cable side, but I don't have the tools (or time or inclination) to view what's going on there.
At least Comcast got the opt-out implementation right. It's done by the cable modem's MAC address, which means that all DNS lookup traffic will start getting NXDOMAIN queries. Oddly, their instructions indicate that this only takes effect when your modem does its next DHCP client lease. My guess is they've blocked off a range of IPs as "opt out," and just assign your MAC to get a lease from the out out range.
I'd greatly prefer it if Comcast had just left things alone, of course; at least, though, they didn't fall into the old "The Web is the Internet" fallacy like Bell Canada.
Understood. I was just amused at the thought of a GNU project being filed under BSD. Not quite like bringing antimatter and matter together (or emacs/vi), but the fallout would be amusing nonetheless.
And gcc is widely used as the compiler for BSD, but I doubt it'll be filed there. :-)
SIGINT isn't the right tool for tracking terrorist cells anyway.
Absolutely. It's SIGUSR1 if you want it to print statistics out. Then you can send a SIGHUP, SIGTERM, or SIGKILL, as necessary.
Re-read it. I said, "Until they fix these issues, I'll hold on to my incandescents and _carbon arc lamps_." Nobody would seriously use carbon arc lamps these days -- that's my tongue-in-cheek way of saying, "Git off my lawn." It's humor. Apparently too subtle for /.
Enjoy your little conspiracy enclave!
Where the hell did I say I wouldn't switch? My god, you folks are all humor-impaired! I said I wouldn't switch from my "incandescents and carbon arc lamps." Because, y'know, everyone is still using carbon arc lamps to shine the lights inside of their barns so they can hitch up their buggys...
You folks are just so amazingly clueless and unwilling to think, it's sad.
Whoosh!
No, 28 VA == 28 W at 1.0 pf. Do you not understand that VA is the same units as a watt, and the notational difference is used to distinguish between real and apparent power?
And, no, your other appliances will not use the energy. You're thinking of bypass capacitors used to smooth out ripples. Appliances can't "steal" current on different phases. Where are you getting this line of reasoning?
Absolutely. But we can still do better. I'm just amazed at how bad some of these cheap designs are.
Ahh! Yes, I even watched that episode, now that you mention it. I can't believe I fell for the old myth even having had it busted.
Thanks for the LBL paper. Good source.
Did the poster only ask people who haven't tried a new lamp in the last 5 years?
That'd be me. It's mixed, depending on the installation location and brand. Outdoors, in 20-30ÂF weather: 2-3 minutes to become acceptable, 5 minutes to stop becoming brighter. Indoors, 40 seconds or so for GEs, 20 seconds for Home Depot cheap-o brand (but they tend to go within a year or so). Aside from the outdoors, I've kept all the CFLs indoors or added a mix of CFL and incandescent.
Oh, and that list was not meant to be an, "OMG, CFL bad and here's why" list. More of a, "There's no such thing as a free lunch; here's what you need to consider" list. But it would seem that CFLs are a surprisingly political topic on /.
You don't think that the difference in cost would be pretty negligible?
Sure, but using an incandescent is also a negligible (end-user) cost. Summed across the total populace, this could be significant.
But also see the sibling reply from quenda, who states that my original point is actually a myth.
Really? Huh. Admittedly, I haven't measured it, but I had always heard that you shouldn't use fluorescents (standard or CFLs) in places where you expect the light to be on less than 30 minutes on average. I can see how this wouldn't be true for CFLs, though.
What about lifetime? Does more frequent power cycling affect their useful life?
Think incandescent is hard, try getting a carbon arc lamp. (This last line was meant as a joke, btw... :-)
Wow, I had no idea this would've been taken as an anti-CFL rant. Apparently neither my viewpoint nor the article's came through in the summary, which is more that there's no such thing as a free lunch (i.e. CFLs have downsides, too.). I think the headline was regrettably chosen, though, which I took from the original article and reworded to fit in /.'s character limits.
Bruce, you make a lot of good points. Yes, the mercury output is less than if you're on coal electricity (we're mostly hydro here). And unless lighting is the large majority of your electric bill (which it isn't for most households), the power factor of those CFLs aren't going to matter. I was surprised, however, to find that the power factor is as low as it is. I'll be happy once we have fewer integrated ballasts (which are produced as cheaply as possible). Spending an extra dollar on the ballast could improve the power factor and other issues significantly.
However, there's one claim that is problematic:
If you're using CFLs indiscriminately, you're applying the technology suboptimally. A rarely and briefly-used hallway light, for example, would be better served by an incandescent. The initial power surge to kick off the light will eat into your usage and savings -- very minutely, but still not the benefit you think you're getting. Which was really the point of submitting this article: we can't blindly use any technology -- CFLs, hybrid cars, wind power, etc. -- thinking that it's the perfect solution. Some thought has to be applied, and that requires information.
If the CFL players aren't hitting as hard (real power) as it looks on TV (apparent power), then they're also below unity power factor.
Whether your power company cares, though... :-)
The original text was written for a different audience, though I thought it would work for /. as well. Guess not. :-)
To power the 13 W light bulb with a 0.45 pf, you need to generate 28 VA (plus change) of power. This same electrical output could also power a 28 W light bulb with a 1.0 pf. (All assuming negligible resistive transmission losses.) I know it's not consumed by the end user, but it is energy being put out at the plant nonetheless.
Heh, no rant to unleash. Just that CFLs aren't the panacea that they're often made out to be. (Note that I live in an area where you are often considered an $EVIL_DEITY if you don't use CFLs, shop with reusable hemp sacks, install a cistern to store rain water in the driveway underneath your hybrid biodiesel-electric-flower car, etc.)
That said, I did go out and buy a bunch of CFLs. I've found:
Anyway, I thought that the claim that I'd be sticking to my carbon arc lamps would be clear that I'm making this tongue-in-cheek (it's like claiming that I commute to work in a steam powered horseless buggy); I still use CFLs, just a bit more judiciously, and like having this kind of information. Guess the average /. reader isn't as savvy as I had thought. Ah, well.
MySQL strives to provide RDBMS and ACID semantics, though its quality of service (QoS) may fall short. By contrast, these "slacker" databases don't even try to support RDBMS or ACID; even if they operated perfectly, they won't provide RDBMS/ACID.
I work for one of the companies in question (no, I don't speak for them). We rely heavily on a combination of these "slacker" dbs, Berkeley dbs, memcached, Oracle, flat files, and tape backups. Each fills a niche. I wish these articles would quit trying to create a false dichotomy.
How is this like google maps?
They both use Javascript.