Take ALL the dust inside a typical pc that needs cleaning, and turn it sticky. There will be a lot of parts you will never be able to get clean, a lot of fans that will never spin again. Trust me, it's horrible.
I've been mucking about with the idea of building a cpu from logic gates for a while now, and I might have to use a "single instruction" architecture. The registers and inputs and outputs of small function units would all be addressable memory. Technically you could argue that some addresses have instruction meanings as a result, but what the hell. A program would be along the lines of: Move from constant #000 (could be memory loc #000) to increment input 0 (say memory loc #002) Move from memory loc #100 (a real memory location) to add unit input 0 (say memory loc #006) Move from memory loc #101 to add unit input 1 (say memory loc #007) Move from add unit output 0 (say memory loc #008) to memory loc #101 Move from increment output 0 (say memory loc #003) to increment input 0 (loc #002) and then a conditional jump to the top on the increment output reaching a value, which I figure would have to be handled by a subtract and a "conditional unit", with the conditional unit returning one input or another depending on yet another input (fed from the subtract), which could then be moved into the PC. Yes, this technically means it does a non-conditional jump to one of two addresses depending on a condition, not a "conditional jump", but that's close enough.
It would be easy to build due to being so incredibly modular. You could even make it execute two "instructions" per clock or more to make some more traditional instructions take only a single clock cycle, instead of a cycle per input and output.
The PPE is almost exactly the same as a single core of the 360's chip. The SPUs are each about the same as well, but their power is limited by insanely small local memory and huge latencies to the main memory (to the point where they can't read it directly, they have to issue DMA transfers).
This gives the PS3's cell theoretically 2.3x the performance of the 360's cpu (1 PPE + 6 SPUs in the PS3 vs essentially 3 PPEs in the 360), but in practice less than 2x. When running games the OS reserves one SPE, and one is disabled to improve manufacturing yield, which is why I say 6 SPUs. I don't know if the 7th is available to use when the PS3 is running Linux, but I doubt it.
Folding's own PS3 FAQ says that "The GPU client is still the fastest", blowing your claim of "a single PS3 outputiing 10x what a GPU based algorythm is kicking out". In fact, the stats page shows GPUs contributing more TFLOPS worth of work units than PS3s, with fewer active clients, suggesting that GPUs are on average 3-4x as powerful as PS3s.
Lastly, I have been a PS3 and 360 developer for a few years now, so I think I might have some clue about their relative performance.
The "double-precision version" I was talking about wasn't for its double-precision capabilities, more for the fact that it's a newer and more powerful version of the cell chip in general. My apologies for not being clearer.
I did find that the Intel Core i7 has a theoretical of 70 double-precision gflops. The single-precision number should be much higher, as the Pentium 4 apparently managed 70 single-precision gflops.
Still, compared to the top-end gfx chips' over 1 TFlop of power, the cell is weedy.
The Cell (at least the usable portion) is less than twice as powerful as the xbox 360's tri-core cpu. The gpu is weaker than the 360's, and it is slightly more powerful than the cell, but even harder to program for. Overall, both consoles have a similar theoretical performance.
The cpu+gpu put together in either one are still outclassed by just your 8800GT, let alone a modern gpu (the GTX 285 is single-chip and readily available, and 3-4x as powerful as your 8800). This is all working in single-precision, and I can't find any single-precision performance numbers for a modern cpu, but I'd bet that they easily outclass PS3s too.
Though the article makes it sound like they chose PS3s for their performance/cost ratio, so the fact that it doesn't have top-end outright performance is perhaps irrelevant to them. I still think they should have got a cell-chip-based blade server, using the double-precision version of the cell chip (which is not the one that's in the PS3), and probably would have access to two more SPUs (the PS3 reserves one for OS and has one disabled for yeild) per cell chip. Knowing reporting these days, that's probably what they did get.
OTOH, the nvidia card in my media center refuses to read the EDID of my projector.
Linux nVidia driver only reads EDID 1.x, no 2.0. If the manufacturer fails to put a useful backup EDID 1.x as required by EDID 2.0 spec then [...]
It works fine with the old version of the driver (about a year old)
Read the whole thing before you comment. Are you suggesting that the older driver could read EDID 2.0, but the new one can't? That seems backwards.
It seems more likely to me that the drivers DO now support EDID 2.0, and the 2.0 EDID the projector is sending is garbage or out-of-spec in some way. The older drivers would have used the EDID 1.x backup, which would be ok.
The exploit is for an OUTBOUND connection via a malicious link in an email etc. If it was an inbound connection then it really would come under the "yawn" category and be blocked by everyone's firewalls. Very few "home" grade routers / firewalls will block outgoing connections at all, let alone by default, because that is such a pain to manage with the plethora of online games about, all using different outgoing ports.
...and is then followed by the html for the google search results page for telnet http. In other words, it's perfectly possible to make a request to a web server using telnet, including performing a query (in this case a google search). If that query was an SQL injection, someone would have an exploited database on their hands.
See: Left 4 Dead. Great concept, but almost impossible to get dedicated servers running for it. Or you can look at the recently released-for-PC game Borderlands - what a clusterfuck; the community eventually figured out what ports to unblock on their firewall, but even now people are having problems getting people to connect to their game/server.
We have an INX dedicated server we can switch between Left 4 Dead and the Left 4 Dead 2 demo. Actually using it is a pain in the arse though. We haven't used the feature of associating it to our steam group since they added it, because it didn't support grouping up in a lobby and choosing gamemode, level, characters etc before playing. You had to restart the server to change gamemodes! Setting a search key and force_dedicated_servers list seem to work though, so we've been using that.
For Borderlands, only the one of us with a public IP (actually multiple static IPs and a router that supports multiple DMZs with different IPs, so he's technically still behind NAT) has ever hosted a game successfully. Even better, it never works first time, he always has to restart Borderlands before we can actually connect. The game itself though (apart from a few minor pre-first-patch bugs) is awesome. Who doesn't want a rocket launcher that fires a spread of five rockets that set everything on fire when they explode?
It also assumes that the farm haven't been over-fertilising their field for years, and that as a result there is enough fertiliser left in the soil to grow crops to the normal "fertilised" yield without adding more fertiliser this year. Anyone who tests this should also test a control section of field that was in the same starting state (fertilised or not the years before, grew the same crops the years before, etc) and doesn't get fertilised and doesn't get this process.
If firewire licensed the usb port, they could pull the same trick. Release a "FireUSB" which is compatible with ordinary USB and also better in firewire ways. 'course, they'd probably never get the license.
I would happily pay an honest price for (on-demand) movies and series if only it was as convenient as buying app-store apps and if it would actually be available over here. For example: the new Stargate series, it'll be years before it's on TV here, and they'll probably mess up the order (I have no clue why they do this, but they can's seem to ever show any series in the correct order over here), stop halfway through a season, broadcast it at random times, etc. It's almost as if they don't want people to follow the series.
Indeed. They're getting better with some series in the UK actually being broadcast only a week or so behind the first broadcast in the US, but I still remember people saying: "Have you seen last night's new SG1 episode?" and me replying "Two years ago.".
It's getting good with music, there are a lot of places to legally download mp3s (i.e. with no copy-restrictions) now. Legal TV/movies for download? Not so much.
Indeed. When I first built my pc it was almost all pirated software and games, with the remainder being free.
Now, my pc has no pirated software or games on it at all (and with quite an extensive games collection). TV and movies on the other hand, would be a lot more expensive to buy all I have downloaded. I'm working on it though, slowly.
Personally I am most bothered by Vista/7's Big Brother moves (protected video path)
Misunderstood. The protected video path is only for HDCP support, and is entirely unused normally.
stupid interface changes
In Windows 7 Paint has a ribbon interface. Take that however you want.
and that brain dead "upgrade" of file moving/copying.
What? You don't like being able to hit "retry" when a transfer fails instead of ending up part-way through a move/copy? What about the option to "keep both files (rename the new one)" when you copy a file to a location that already contains one with the same name? How about actually being able to copy a folder over another and get asked what to do about naming conflicts inside, instead of only being able to choose "cancel transfer" vs "overwrite everything"? (they fixed the ludicrously slow network file transfer bug years ago)
Wait, I misread. You are saying you choose the real or dummy receipt, at the pollbooth. That's not bad, but your vote could be changed because you can never prove whether you took the real or dummy receipt. So much for watching over your vote by taking the real receipt.
Slashdot Mirror
Thin layer?
Take ALL the dust inside a typical pc that needs cleaning, and turn it sticky. There will be a lot of parts you will never be able to get clean, a lot of fans that will never spin again. Trust me, it's horrible.
I've been mucking about with the idea of building a cpu from logic gates for a while now, and I might have to use a "single instruction" architecture. The registers and inputs and outputs of small function units would all be addressable memory. Technically you could argue that some addresses have instruction meanings as a result, but what the hell. A program would be along the lines of:
Move from constant #000 (could be memory loc #000) to increment input 0 (say memory loc #002)
Move from memory loc #100 (a real memory location) to add unit input 0 (say memory loc #006)
Move from memory loc #101 to add unit input 1 (say memory loc #007)
Move from add unit output 0 (say memory loc #008) to memory loc #101
Move from increment output 0 (say memory loc #003) to increment input 0 (loc #002)
and then a conditional jump to the top on the increment output reaching a value, which I figure would have to be handled by a subtract and a "conditional unit", with the conditional unit returning one input or another depending on yet another input (fed from the subtract), which could then be moved into the PC. Yes, this technically means it does a non-conditional jump to one of two addresses depending on a condition, not a "conditional jump", but that's close enough.
It would be easy to build due to being so incredibly modular. You could even make it execute two "instructions" per clock or more to make some more traditional instructions take only a single clock cycle, instead of a cycle per input and output.
"What is it and where is it?"
Are you sure didn't mean G Spot? :)
Which is the name of a program for identifying the codecs used in .avi files, so you know what to get to be able to play them.
Seriously.
And before someone mentions it, I was talking in FLOPS because it's easier to find those numbers than integer ops numbers.
The PPE is almost exactly the same as a single core of the 360's chip. The SPUs are each about the same as well, but their power is limited by insanely small local memory and huge latencies to the main memory (to the point where they can't read it directly, they have to issue DMA transfers).
This gives the PS3's cell theoretically 2.3x the performance of the 360's cpu (1 PPE + 6 SPUs in the PS3 vs essentially 3 PPEs in the 360), but in practice less than 2x. When running games the OS reserves one SPE, and one is disabled to improve manufacturing yield, which is why I say 6 SPUs. I don't know if the 7th is available to use when the PS3 is running Linux, but I doubt it.
Folding's own PS3 FAQ says that "The GPU client is still the fastest", blowing your claim of "a single PS3 outputiing 10x what a GPU based algorythm is kicking out". In fact, the stats page shows GPUs contributing more TFLOPS worth of work units than PS3s, with fewer active clients, suggesting that GPUs are on average 3-4x as powerful as PS3s.
Lastly, I have been a PS3 and 360 developer for a few years now, so I think I might have some clue about their relative performance.
The "double-precision version" I was talking about wasn't for its double-precision capabilities, more for the fact that it's a newer and more powerful version of the cell chip in general. My apologies for not being clearer.
I did find that the Intel Core i7 has a theoretical of 70 double-precision gflops. The single-precision number should be much higher, as the Pentium 4 apparently managed 70 single-precision gflops.
Still, compared to the top-end gfx chips' over 1 TFlop of power, the cell is weedy.
The Cell (at least the usable portion) is less than twice as powerful as the xbox 360's tri-core cpu. The gpu is weaker than the 360's, and it is slightly more powerful than the cell, but even harder to program for. Overall, both consoles have a similar theoretical performance.
The cpu+gpu put together in either one are still outclassed by just your 8800GT, let alone a modern gpu (the GTX 285 is single-chip and readily available, and 3-4x as powerful as your 8800). This is all working in single-precision, and I can't find any single-precision performance numbers for a modern cpu, but I'd bet that they easily outclass PS3s too.
Though the article makes it sound like they chose PS3s for their performance/cost ratio, so the fact that it doesn't have top-end outright performance is perhaps irrelevant to them. I still think they should have got a cell-chip-based blade server, using the double-precision version of the cell chip (which is not the one that's in the PS3), and probably would have access to two more SPUs (the PS3 reserves one for OS and has one disabled for yeild) per cell chip. Knowing reporting these days, that's probably what they did get.
OTOH, the nvidia card in my media center refuses to read the EDID of my projector.
Linux nVidia driver only reads EDID 1.x, no 2.0. If the manufacturer fails to put a useful backup EDID 1.x as required by EDID 2.0 spec then [...]
It works fine with the old version of the driver (about a year old)
Read the whole thing before you comment. Are you suggesting that the older driver could read EDID 2.0, but the new one can't? That seems backwards.
It seems more likely to me that the drivers DO now support EDID 2.0, and the 2.0 EDID the projector is sending is garbage or out-of-spec in some way. The older drivers would have used the EDID 1.x backup, which would be ok.
Try Vista / 7's DPI setting, it scales any legacy programs if necessary, or lets clever programs handle it themselves.
The exploit is for an OUTBOUND connection via a malicious link in an email etc. If it was an inbound connection then it really would come under the "yawn" category and be blocked by everyone's firewalls. Very few "home" grade routers / firewalls will block outgoing connections at all, let alone by default, because that is such a pain to manage with the plethora of online games about, all using different outgoing ports.
Anyone who hasn't seen Blackadder, I recommend you find a way to do so.
Your GET request and your host header are the wrong way round. It's request first, then headers (including host).
Try this: /search?q=telnet+http HTTP/1.0
telnet www.google.com 80
GET
Host: www.google.com
(ending in a blank line)
You should get a html response from google, which begins like this:
HTTP/1.0 200 OK
Cache-Control: private, max-age=0
Date: ##
Expires: -1
Content-Type: text/html; charset=ISO-8859-1
Set-Cookie: SS=Q0=##; path=/search
Set-Cookie: PREF=ID=##:TM=##:LM=##:S=##; expires=##; path=/; domain=.google.com
Set-Cookie: NID=28=##-##-##; expires=##; path=/; domain=.google.com; HttpOnly
Server: gws
X-XSS-Protection: 0
...and is then followed by the html for the google search results page for telnet http. In other words, it's perfectly possible to make a request to a web server using telnet, including performing a query (in this case a google search). If that query was an SQL injection, someone would have an exploited database on their hands.
See: Left 4 Dead. Great concept, but almost impossible to get dedicated servers running for it. Or you can look at the recently released-for-PC game Borderlands - what a clusterfuck; the community eventually figured out what ports to unblock on their firewall, but even now people are having problems getting people to connect to their game/server.
We have an INX dedicated server we can switch between Left 4 Dead and the Left 4 Dead 2 demo. Actually using it is a pain in the arse though. We haven't used the feature of associating it to our steam group since they added it, because it didn't support grouping up in a lobby and choosing gamemode, level, characters etc before playing. You had to restart the server to change gamemodes! Setting a search key and force_dedicated_servers list seem to work though, so we've been using that.
For Borderlands, only the one of us with a public IP (actually multiple static IPs and a router that supports multiple DMZs with different IPs, so he's technically still behind NAT) has ever hosted a game successfully. Even better, it never works first time, he always has to restart Borderlands before we can actually connect.
The game itself though (apart from a few minor pre-first-patch bugs) is awesome. Who doesn't want a rocket launcher that fires a spread of five rockets that set everything on fire when they explode?
It also assumes that the farm haven't been over-fertilising their field for years, and that as a result there is enough fertiliser left in the soil to grow crops to the normal "fertilised" yield without adding more fertiliser this year. Anyone who tests this should also test a control section of field that was in the same starting state (fertilised or not the years before, grew the same crops the years before, etc) and doesn't get fertilised and doesn't get this process.
If firewire licensed the usb port, they could pull the same trick. Release a "FireUSB" which is compatible with ordinary USB and also better in firewire ways.
'course, they'd probably never get the license.
Hey this might actually become my first 5-star posting *hopes for the best*
I'm afraid you're at the bottom of the page, and most people run out of mod points before they get here.
I mean I'd be a little bonkers to drive some of the tiny cars that are popular in Europe on an interstate
I don't see why, most of them are faster than US cars.
I would happily pay an honest price for (on-demand) movies and series if only it was as convenient as buying app-store apps and if it would actually be available over here. For example: the new Stargate series, it'll be years before it's on TV here, and they'll probably mess up the order (I have no clue why they do this, but they can's seem to ever show any series in the correct order over here), stop halfway through a season, broadcast it at random times, etc. It's almost as if they don't want people to follow the series.
Indeed. They're getting better with some series in the UK actually being broadcast only a week or so behind the first broadcast in the US, but I still remember people saying: "Have you seen last night's new SG1 episode?" and me replying "Two years ago.".
It's getting good with music, there are a lot of places to legally download mp3s (i.e. with no copy-restrictions) now. Legal TV/movies for download? Not so much.
Indeed. When I first built my pc it was almost all pirated software and games, with the remainder being free.
Now, my pc has no pirated software or games on it at all (and with quite an extensive games collection). TV and movies on the other hand, would be a lot more expensive to buy all I have downloaded. I'm working on it though, slowly.
Personally I am most bothered by Vista/7's Big Brother moves (protected video path)
Misunderstood. The protected video path is only for HDCP support, and is entirely unused normally.
stupid interface changes
In Windows 7 Paint has a ribbon interface. Take that however you want.
and that brain dead "upgrade" of file moving/copying.
What? You don't like being able to hit "retry" when a transfer fails instead of ending up part-way through a move/copy? What about the option to "keep both files (rename the new one)" when you copy a file to a location that already contains one with the same name? How about actually being able to copy a folder over another and get asked what to do about naming conflicts inside, instead of only being able to choose "cancel transfer" vs "overwrite everything"?
(they fixed the ludicrously slow network file transfer bug years ago)
I say "wouldn't touch that with a barge pole", which may be a step back from actually giving a length...
It still seems easier to just not take a receipt at all if you are being coerced. No need for a fake.
Much simpler.
Wait, I misread. You are saying you choose the real or dummy receipt, at the pollbooth. That's not bad, but your vote could be changed because you can never prove whether you took the real or dummy receipt. So much for watching over your vote by taking the real receipt.