Lets say a typical computer running BOINC contributes 1 GFlop at 100W (1e2W). So at 2e6 GFlops, tats 2e8W or 2e5 kW.
According to the energy department, we can assume that 1.4 pounds of CO2 per KWh, so that says BOINC is at ~3e5 pounds/hour of CO2, or about 140 tons/hour of CO2.
I get a very similar number if I back of the envelope what a coal plant should be based on ~500 tons/1 GW.
Both are very mainline FPGAs, both have full devkits, references designs, include the tools, linux support on Xilinx at least (not sure on Altera), and are both at your price point.
I don't want to name names, but Netalyzr showed that several major ISPs already do this, and allows you to check for yourself what the behavior is on your network.
Comcast is following the lead of other major ISPs which have been doing this for some time now.
You can get a good standalone nav system for $150 for your car, which doesn't end up costing ungodly-per-month.
For those who have a smartphone, yeah, the standalone nav system is dead. But for those unwilling to pay the $40/month more that something like an iPhone costs, the standalone nav systems have an excellent role still to play.
Not suprising, and it comes down to two factors: Piracy, and cheating.
For piracy, by only supporting multiplayer through BattleNet, it is BattleNet, not the software itself, that validates the CD keys/antipiracy measures. So no more LAN-arcades composed of a single cracked copy.
As important, without a client-server game model, cheating would be a huge problem. BattleNet forces the user to play in a client-server mode. If they wanted to support LAN only play, they'd either need to include the server as well or have a separate network architecture for peer-to-peer play.
"Piracy wars are not solved by solving the halting problem. Piracy wars are solved by making the other poor bastard solve the halting problem..."
This is actually a really clever and somewhat unexpected approach that the BluRay DRM folks have hit on. Rather than doing DRM, have a program and basically force those who are cracking the disks to crack every title differently. Its basically force those who want to develop ripping software to do AV style analysis on every new disk that comes out.
Yes, the DRM on any individual disk will always fall eventually because all the data must be on the disk and recoverable from the disk by the player. But it makes it very VERY annoying for those writing the unauthorized decryption software.
The iPhone password input doesn't mask the most recent inputed character in a password dialog, but masks all the older ones and masks the input one after 2-3 seconds.
Thus you get the masking mostly, but the feedback to prevent errors (which are considerably more annoying on the iPhone/iPod touch keyboard arrangement when typing blind).
One of the things Apple learned well by observing others was the Osborne Effect. And its true: Would you buy a "new" iPhone if you were told a better one was 6 months away, and all the cool features it would have eventually?
The good websites don't pull inappropriate tricks with their pages, the mediocre sites would eventually figure out that they aren't getting indexed by search engines, and improve, and the terrible sites would remain in obscurity, partying with geocities.
Sorry, this is just plain untrue. Have you looked at the source for the FRONT PAGE of Google lately?
The head is 2 script blobs and a style sheet blob.
The body has onload loading of images, an iframe with a bunch of onload crap, etc...
Even the slashdot front page has javascript which is using document.write().
The only way to really index the web these days is to be javascript aware and actually render it.
The spammers WILL get your email address. Be it web trawling, google searchers, or stealing email address off of compromised computers, the spammers will get, and then resell, you email address.
Trying to keep the spammers from getting your email address is a lost cause, and not a battle worth fighting.
Yes, the censorship can be annoying and ridiculous (eg, GTA series: its OK to murder hundreds in a random crime spree, but god forbid there be hidden, unaccessable content of still-underwear-clad figures bumpin-boots).
But I think it has very LITTLE effect on art in games. EG, what effect did censorship have on something like Braid?
The game developers which are actually serious about doing ART are not interested in building "Sex-laden-splatter-fest-3000".
While it's difficult to second guess the decisions a trial lawyer makes, it is hard for me to understand why defendant's lawyers gave the plaintiffs a free pass on the MediaSentry/Jacobson nonsense, and didn't even call their own expert. It is likewise difficult to understand why the jurors weren't instructed as to what the plaintiffs had to prove in order to establish a "distribution", or why it was assumed that they were entitled to recover statutory damages (as opposed to actual damages) at all, there having been no questions or instructions relating to the essential elements of that.
Probably because the realization that the expert witness (I read the written testimony, and I know Yongdae Kim as a colleague: he is an excellent and honest researcher and professor) can't really contribute to the defense.
This is a civil trial, in terms of probabilities. The probability of misidentification of the computer goes way down when you remove the wireless, password protect the computer, and go from there.
The problem is, Media Sentry's evidence is pretty compelling: Identify her IP, identify her commonly used username, identify the songs, and she wasn't offering up a defense of being a poisoner/spreading bogus files.
As much as you'd like to believe it is nonsense, an honest expert witness for the defense would be forced to acknowledge all of this.
Yongdae Kim's written testimony mostly covered cases which could not have occured (no wireless, etc) and which were specifically ruled out by the judge for being irrelevant possibilities, or which would be exceedingly unlikely (a trojan on the system soley for KaZaA, IP address hijacking which, if Ms Thomas's computer was on, might result in RST storms from unexpected data, dropping the hijacker, by a hijacker who anyway was trying specifically to frame Ms Thomas), and on the stand he'd have to say so.
This is likely why Dr Kim was not put on the stand: during mock cross examination, Ms Thomas's laywer realized just how damaging Dr Kim's testimony would be in the hands of a plantiff's attorney.
We have observed some firewalls that, regardless of settings, block random packets over UDP port 53 but do allow real DNS. And it may also be NATs as well. Nats do weird things, and you are behind a NAT.
The test works by first sending a random UDP packet that our server echos back (it tries 5 times). IF that fails, it then tries to send a request thats a legitimate DNS request to the same server (again, it tries 5 times).
It also tries to send a legitimate DNS request that will produce a large (~1800B) EDNS-present response. Again, this is attempted 5 times.
Looking at the transcript (the "transcript" link on the results page), the non-DNS request over port 53 was blocked, the large-response was not received, but the real request was allowed to pass unmodified.
QED, something in the network: A host-based firewall, a network firewall, a NAT, etc, parses DNS and only allows valid DNS requests through on UDP port 53.
Which NAT are you using?
But as a comcast customer myself, all my packets on port 53 are unaffected, whether or not they are valid DNS requests. As are most other Comcast customers. So I don't believe you are seeing a Comcast issue.
That is probably your NAT or Firewall. We have observed in the big flashcrowd that there are scattered individuals (not from any particular ISP) who have NATs or firewalls that will only allow real DNS requests through.
This is probably your NAT. We see such behavior among random visitors, but not those restricted to Comcast, and only a few Comcast-based visitors show this behavior.
Slashdot Mirror
Lets say a typical computer running BOINC contributes 1 GFlop at 100W (1e2W). So at 2e6 GFlops, tats 2e8W or 2e5 kW.
According to the energy department, we can assume that 1.4 pounds of CO2 per KWh, so that says BOINC is at ~3e5 pounds/hour of CO2, or about 140 tons/hour of CO2.
I get a very similar number if I back of the envelope what a coal plant should be based on ~500 tons/1 GW.
http://www.eia.doe.gov/cneaf/electricity/page/co2_report/co2report.html
The researcher found the computer that was used as the entry point for commands into the botnet.
This has nothing to do with who is responsbile for the attack.
Xilinx: http://www.xilinx.com/products/devkits/HW-SPAR3AN-SK-UNI-G.htm
Altera: http://www.altera.com/products/devkits/altera/kit-cyc3-starter.html
Both are very mainline FPGAs, both have full devkits, references designs, include the tools, linux support on Xilinx at least (not sure on Altera), and are both at your price point.
Comcast's change should have NO effect on any other ISP, and additionally, DNS issues should not cause packet loss.
I don't want to name names, but Netalyzr showed that several major ISPs already do this, and allows you to check for yourself what the behavior is on your network.
Comcast is following the lead of other major ISPs which have been doing this for some time now.
You can get a good standalone nav system for $150 for your car, which doesn't end up costing ungodly-per-month.
For those who have a smartphone, yeah, the standalone nav system is dead. But for those unwilling to pay the $40/month more that something like an iPhone costs, the standalone nav systems have an excellent role still to play.
Not suprising, and it comes down to two factors: Piracy, and cheating.
For piracy, by only supporting multiplayer through BattleNet, it is BattleNet, not the software itself, that validates the CD keys/antipiracy measures. So no more LAN-arcades composed of a single cracked copy.
As important, without a client-server game model, cheating would be a huge problem. BattleNet forces the user to play in a client-server mode. If they wanted to support LAN only play, they'd either need to include the server as well or have a separate network architecture for peer-to-peer play.
"Piracy wars are not solved by solving the halting problem. Piracy wars are solved by making the other poor bastard solve the halting problem..."
This is actually a really clever and somewhat unexpected approach that the BluRay DRM folks have hit on. Rather than doing DRM, have a program and basically force those who are cracking the disks to crack every title differently. Its basically force those who want to develop ripping software to do AV style analysis on every new disk that comes out.
Yes, the DRM on any individual disk will always fall eventually because all the data must be on the disk and recoverable from the disk by the player. But it makes it very VERY annoying for those writing the unauthorized decryption software.
MiniUSB is rated for 1000 connect/disconnect cycles
MicroUSB is rated for 10,000 connect/disconnect cycles, and is also thinner by about 1.5mm (critical on modern thin devices).
Given the power consumption on some smartphones, having the more durable connector is IMO, essential.
As much as Stallman would like to say otherwise, Linux is not GNU/Linux, and GNU is not all free software.
And lets face it, Debian has a choice:
Either not include a useful application for the sake of "purity", or include a useful runtime and applications which use it.
IS the goal to create a useful system or a pure system?
The iPhone password input doesn't mask the most recent inputed character in a password dialog, but masks all the older ones and masks the input one after 2-3 seconds.
Thus you get the masking mostly, but the feedback to prevent errors (which are considerably more annoying on the iPhone/iPod touch keyboard arrangement when typing blind).
One of the things Apple learned well by observing others was the Osborne Effect. And its true: Would you buy a "new" iPhone if you were told a better one was 6 months away, and all the cool features it would have eventually?
Sorry, this is just plain untrue. Have you looked at the source for the FRONT PAGE of Google lately?
The head is 2 script blobs and a style sheet blob.
The body has onload loading of images, an iframe with a bunch of onload crap, etc...
Even the slashdot front page has javascript which is using document.write().
The only way to really index the web these days is to be javascript aware and actually render it.
History. I haven't updated my front page in years.
The spammers WILL get your email address. Be it web trawling, google searchers, or stealing email address off of compromised computers, the spammers will get, and then resell, you email address.
Trying to keep the spammers from getting your email address is a lost cause, and not a battle worth fighting.
Yes, the censorship can be annoying and ridiculous (eg, GTA series: its OK to murder hundreds in a random crime spree, but god forbid there be hidden, unaccessable content of still-underwear-clad figures bumpin-boots).
But I think it has very LITTLE effect on art in games. EG, what effect did censorship have on something like Braid?
The game developers which are actually serious about doing ART are not interested in building "Sex-laden-splatter-fest-3000".
Probably because the realization that the expert witness (I read the written testimony, and I know Yongdae Kim as a colleague: he is an excellent and honest researcher and professor) can't really contribute to the defense.
This is a civil trial, in terms of probabilities. The probability of misidentification of the computer goes way down when you remove the wireless, password protect the computer, and go from there.
The problem is, Media Sentry's evidence is pretty compelling: Identify her IP, identify her commonly used username, identify the songs, and she wasn't offering up a defense of being a poisoner/spreading bogus files.
As much as you'd like to believe it is nonsense, an honest expert witness for the defense would be forced to acknowledge all of this.
Yongdae Kim's written testimony mostly covered cases which could not have occured (no wireless, etc) and which were specifically ruled out by the judge for being irrelevant possibilities, or which would be exceedingly unlikely (a trojan on the system soley for KaZaA, IP address hijacking which, if Ms Thomas's computer was on, might result in RST storms from unexpected data, dropping the hijacker, by a hijacker who anyway was trying specifically to frame Ms Thomas), and on the stand he'd have to say so.
This is likely why Dr Kim was not put on the stand: during mock cross examination, Ms Thomas's laywer realized just how damaging Dr Kim's testimony would be in the hands of a plantiff's attorney.
The RIAA's evidence is compelling.
The IP address was hers, with no WiFi, no NAT, and a password-protected Windows box
The username chosen was the one she's used online traditionally for 16 years.
The problem is, the nearly inevitible $100K verdict will not be justice, even if it is legally correct.
Are you a Wow Way customer? IF so, please email netalyzr-help@icsi.berkeley.edu
We would like to look into this in more detail.
thank you.
We did do a story submission, but not as an ask-slashdot. The article did not get accepted.
We will probably wait until we are out of beta before we attempt to submit a story to slashdot again ourselves.
We have observed some firewalls that, regardless of settings, block random packets over UDP port 53 but do allow real DNS. And it may also be NATs as well. Nats do weird things, and you are behind a NAT.
The test works by first sending a random UDP packet that our server echos back (it tries 5 times). IF that fails, it then tries to send a request thats a legitimate DNS request to the same server (again, it tries 5 times).
It also tries to send a legitimate DNS request that will produce a large (~1800B) EDNS-present response. Again, this is attempted 5 times.
Looking at the transcript (the "transcript" link on the results page), the non-DNS request over port 53 was blocked, the large-response was not received, but the real request was allowed to pass unmodified.
QED, something in the network: A host-based firewall, a network firewall, a NAT, etc, parses DNS and only allows valid DNS requests through on UDP port 53.
Which NAT are you using?
But as a comcast customer myself, all my packets on port 53 are unaffected, whether or not they are valid DNS requests. As are most other Comcast customers. So I don't believe you are seeing a Comcast issue.
That is probably your NAT or Firewall. We have observed in the big flashcrowd that there are scattered individuals (not from any particular ISP) who have NATs or firewalls that will only allow real DNS requests through.
A colleague who knew about our launch told us we just got slashdotted.
We actually WANT to get slashdotted, because that helps us measure the network.
This is probably your NAT. We see such behavior among random visitors, but not those restricted to Comcast, and only a few Comcast-based visitors show this behavior.
Your netalyzr results show no DNS issues in the link you posted, using a Comcast DNS server: