The BT tracker does not know jack about how much you uploaded to / download from everyone...... all it knows is how much your client claimed it uploaded and downloaded.
Here is all that a peer sends to a tracker while doing tracker updates:
GET [InfoHash]?peer_id=[PeerID]&port=[Port]&uploaded=[ Amount Uploaded]&downloaded=[Amount Downloaded]&left=[Amount Left]&key=[Private ID]
The rest is the usual generic HTTP header stuff such as application name, encoding and compression options.
Since I did write a BT client in late 2003 (and currently am in the middle of rewriting it), I probably know what I am writing about.
Normal BT trackers only know how much a peer claimed to have uploaded and downloaded. The only real way to detect leeches would be to get feedback from specially written BT clients about actual peer behavior and report to the tracker. Generic BT clients do not do any of this so anyone who knows a tiny bit of python or Java could modify a BT client to report 10X as much upload as actual and be virtually freed from upload/download ratios.
I do not think this design would fit in the smallest FPGAs... but a Spartan3-1500 (XC3S1500) does cost only $100 while the nearest equivalent Virtex2 was over $500 little more than a year ago. I tried to check the nearest equivalent Virtex4 (XC4VLX40) but the distributor I checked did not have that model in their database. (and the XC3S1500 was the largest Spartan3 in stock)
With a Spartan3, their 90MHz design would probably reach 150MHz or more then again, I have not seen any mention of what FPGA they used. (I'm presuming they used something older)
Actually, Intel's research (before HT became reality) said that on average, the instruction decoder was issuing just under 2.5 instructions per tick out of a maximum of 3... so instruction decoder throughput in single-threaded mode is about 75% of maximum.
On AMD's side, the decoder has quadruple outputs and IIRC, AMD's average is 3 out of 4 so again 75% from maximum.
By adding SMT, Intel gave the P4 the potential to keep all instruction ports busy and AMD plans to do the same next year... a single-core A64 with SMT would be interesting but we will have to settle with dual-core dual-threaded A64s and P4s which should be interesting as well.
How do AMD and Intel manage to get 75% single-threaded when we know they will be stalled by RAM? Simple, out-of-order execution - most CPUs can look 32-128 instructions ahead to find something to do while stalled, this is necessary to maximize single-thread performance and would become unnecessary if apps and CPUs became massively multi-threaded, which appears to be what Sun is gunning for.
As far as concurrent SMT is concerned, I think four threads per CPU core will turn out to be the practical maximum for desktop chips. We will probably see this happen once the A64/P4/PM are upgraded to six execution ports, three or four years from now.
The only reason Sun can think of doing a SMTx8/32 chip is because their CPUs runs at ~1GHz. At higher speeds, they would not have the necessary timing margins to fit the extra logic to efficiently shuffle execution states between "reserve" and "active" threads.
Cable modems share the same RG59/RG6/??? cable so anything sent or received is also receivable by everyone in the same network segment... that's why cable modems are usually setup to use 3DES or AES encryption to preserve point-to-point (modem to head-end) confidentiality.
Any unencrypted transmission over a common carrier is open to sniffing and men-in-the-middle attacks so it would not be sane for an ISP or LAN of any kind to use any such media (air, power-line, phone-line, fiber, etc.) unencrypted - they are all interceptable.
AMD64 is not any less real 64bits than the PowerPC... the PowerPC was originally a 32bits chip and got 64bits extensions a few years before AMD/Intel's x86 chips.
How many pure 64bits CPUs are out there? All those I know are based on some older 8/16/32 bits, expanding and warping the instruction set and architecture along the way. AMD/Intel only carry more legacy cruft than many others along with the irregular and somewhat crufty x86 instruction set tradition... and all these bugs^H^H^H^H^H errata that had to become features for backwards compatibility.
It would make everyone's life easier if legacy architectures could be weeded out... Intel tried to go clean-slate with Itanium and that turned Itanic. Old habbits die hard - few people are willing to abandon an architecture everybody is familiar with and start anew.
1) Mailing the claim can cost as much as $10 with reception confirmation and other such options... make that $2 for plain enveloppe and international postal charges. 2) Cashing the refund can cost over $5 for people without USA-funds banking accounts. 3) Most rebates I have seen have a disclaimer that says they will be honoured at the manufacturer's sole discretion.
Because of this, I only buy into rebates when the base price suits me - FutureShop having a $110 CDN Audigy2 sale plus $45 mail-in is nice when the next lowest regular price around is $115 - this is how I discovered that USA rebates cost about $10 to claim... so that $45 rebate barely covered taxes, postage and cash-in costs so the card cost me about $110 net.
All in all, rebates are annoying and doubly so when they are in some other funds, not worth bothering with if under $20 - I prefer waiting a little longer until the "rebate" price becomes the regular price since rebates usually mean pending price adjustments and new models.
Since when is "America" a country? Last time I checked, there was more than one "America" and these were considered continents...... AFAIK, the USA has not conquered Canada yet - but with Bush in charge, who knows.
AFAIK, patent protection is a per-country thing. Patent can be registered only with a number of countries (key markets) at ~$10k a pop or registered internationally (to every country where the patent is deemed valid) for ~$100k. In either case, only one patent for any given "invention" can be valid internationally.
If someone in Taiwan built and sold items that used some technology patented only in the USA, the taiwanese company could still sell outside the USA market - at least until whoever patented the thing decided to pay $10k for protection in supplemental markets or $100k for international coverage.
One "problem" with saving Hubble is that now, image processing techniques have made it possible to merge observations from multiple ground-based telescope and achieve better-than-Hubble resolution.
With Hubble rendered nearly obsolete by ground-based computing and sensing advances, repairing Hubble is most likely not worth it unless it is also upgraded. Assembling Hubble on ground took months, upgrading would require significant (possibly delicate) disassembly and subsequent re-assembly which probably are not reasonably feasible in open space.
This depends... Enron and Nortel's management had quite a bit of trouble with the SEC and shareholder class-action suits when their suspicious accounting practices and stock manipulations came to light.
If the court concludes that SCO's IBM lawsuit was only a diversionary tactic meant to float stock prices before management dumped theirs, they will almost certainly get investigated by the SEC and get a class-action case from their shareholders seeking compensation.
So not all is lost for the loss-making shareholders yet.
My ISP claims that less than 10% of its subscribers generate 99% of its traffic and the top 1% accounts for more than 50%, with average usage being under 2GB/month after discarding the top and bottom 5%.
Since less than 10% of people account for 99% of bandwidth, I'd say Bittorrent is the distribution equivalent of Linux on the desktop: both are well-known by the geek comunity but far from being ubiquitous in the mass market.
With all the laws and other nuisances or nonsense that will be used to outlaw, impede or prevent online distribution, it seems unlikely that BitTorrent and others will be allowed to become truly mainstream.
That's where nVidia's new SLI and ATI's AMR come in handy!
For games, frame rendering usually begins by clearing the frame buffer since the whole screen usually needs redrawing but for regular apps, updating only active windows when changes occur should be far less expensive in rendering time - these would consist mainly of mapping 2D surfaces on some polygons, something even the oldest 3D cards should be able to do decently fast.
I was thinking only about the actual contents...... but I suppose it is reasonable to presume simplist people who watch brainwash-quality shows are indeed more likely to buy everything they see and generate more advertisement revenue.
Guess good (broadcast) shows beyond the occasional exceptions are going to disappear.
As a sidenote, I think all first and second seasons of TNG, DS9 and Voyager generally sucked compared to later seasons so Enterprise is not much of an exception.
I always thought "flavoured" usually implies artificially so - pork-flavoured pork is redundant since pork should implicitly taste like pork. How many "flavoured" things actually contain the real ingredient? Most meat favours are mostly salt and a dehydrated soup base while most others are based on extracts or completely simulated.
As I said previously, most common ingredients probably made at least one allergy victim at some point in time so the only complete allergy list is one that makes the ingredient listing redundant.
This is almost in the same domain as "Warning: contents may be hot." (on coffee cups)
Listing peanuts on the allergy warning of a jar marked 'Peanut Butter' in 1" tall lettering seems just as redundant as labelling "hot" steaming coffee.
Most allergy warning labels I have seen go like "May contain trace amounts of almond, sunflower seeds, sesame seeds, other seeds and nuts". Repeating official ingredients there is redundant: since there probably exist(ed) one person who is allergic to each common ingredient, the allergy warning list would not be complete until the whole ingredient list is repeated anyway, plus potential contaminents.
Internet is a really good place for advertisement desensitivation training.
Firefox with anti-stuff extentsions is great... among other things, my eyes can finally get a break from flashy flash ads thanks to Flashblock. (It also helps taming some flash-abusing sites.)
whereas with AMD Opteron SMP systems, every CPU gets its own dedicated HyperTransport link to its memory
I was not aware there was such a thing as "Hypertransport" DRAM... last time I checked, the Opterons and Athlon-FX had on-chip dual-channel DDR controllers.
BTW, single-core Opterons also share some RAM bandwidth with the other CPUs over the HT links and multi-core Opterons would also share RAM/HT bandwidth along its cores. Single-core Athlon64s have only their own DDR controller so dual-core Athlon64s will have to share RAM bandwidth somewhat like the Xeons do - without the extra package/bus/chipset/others latencies.
Since I just upgraded my PC and laptop, dual-threading dual-core CPUs should be standard fare by the time my next upgrades are due. As a moderate-heavy multitasker, I am looking forward to that and quad/octuple-channel FB-DIMM.
It always irritates me to some extent when I see people use "barrier" interchangeably with "landmark".
Breaking a barrier means working around or proving wrong something that used to be considered a fundamental limit. There is no fundamental limit to how much RAM can be put on a video card other than the number of data/address lines and available physical space.
A landmark is a point bound to be crossed sooner or later. Video cards will undoubtedly get more RAM sooner or later and 1GB low-end video cards will probably be common ten years from now.
Since there are 2GB SO-DIMMs available today, ATI and nVidia would most likely be able to release a 2GB video card as soon as this year if they really wanted to and enough people actually wanted/needed them.
So, the only reason we do not have >2GB video cards yet is because nobody needs them and by the time people would, the GPU would be completely obsolete.
Manufacturers are most likely not that stupid... most probably have more or less trivial flag-kill switches: changing some jumper (or solder bridge/resistor) configuration on the board, firmware upgrade, magic code/switch, etc. Of course, reworking PCBs is not free and is a major hassle if that PCB is buried in some other assembly boxed in some depot. So, magic code which can be put on stickers before shipping would be the most practical approach for appliances.
For PC-based cards, the flag probably is only a bit in the video stream which compliant players/recorders would use to disable inappropriate actions and a simple software upgrade would fix this. (So would using non-flag-compliant software.)
The new news season must be over so the slashdot news are showing reruns.
"Slashdot: yesterday's news, stuff that mattered" Is definitely an appropriate signature... I do not remember seeing such a steady stream of reruns in the past.
But a per-cell 10A current limit would waste about 1W with 0.01ohm MOSFET switches... and current-sensing MOSFET usually drain ~1-10% through the sense lead as well. For a 12-cells battery, this is 12W more that the battery package must dissipate on top of the cells' internal losses during normal operation.
Since cells are unlikely to short one another out unless the battery pack has been physically compromised - at which point individual cells may have also been damaged, per-cell current limit seems simply wasteful overkill - pack-wise current limit would suffice for most handling accidents.
On the other hand, I am totally for built-in charge-controllers/equalizers - they prevent cell reversals and optimize battery capacity, incidentally reducing the risk of (sometimes catastrophic) premature cell failures.
In any case, no amount of charge control or current limiting can protect against manufacturing or engineering defects or end-user stupidity.
I prefer real-world black magic like the "Magic Switch" - stuff that should (not) have any effect in the real world yet actually does.
I did read some fake-telling posts... I just felt like proving that the 7805 they hypothetically used would be long dead before any sensible CPU heat got generated.
Slashdot Mirror
And get banned on subscription tracker sites unless you also modify the client to report bogus but consistent amounts of data.
The BT tracker does not know jack about how much you uploaded to / download from everyone... ... all it knows is how much your client claimed it uploaded and downloaded.
[ Amount Uploaded]&downloaded=[Amount Downloaded]&left=[Amount Left]&key=[Private ID]
Here is all that a peer sends to a tracker while doing tracker updates:
GET [InfoHash]?peer_id=[PeerID]&port=[Port]&uploaded=
The rest is the usual generic HTTP header stuff such as application name, encoding and compression options.
Since I did write a BT client in late 2003 (and currently am in the middle of rewriting it), I probably know what I am writing about.
Normal BT trackers only know how much a peer claimed to have uploaded and downloaded. The only real way to detect leeches would be to get feedback from specially written BT clients about actual peer behavior and report to the tracker. Generic BT clients do not do any of this so anyone who knows a tiny bit of python or Java could modify a BT client to report 10X as much upload as actual and be virtually freed from upload/download ratios.
I do not think this design would fit in the smallest FPGAs... but a Spartan3-1500 (XC3S1500) does cost only $100 while the nearest equivalent Virtex2 was over $500 little more than a year ago. I tried to check the nearest equivalent Virtex4 (XC4VLX40) but the distributor I checked did not have that model in their database. (and the XC3S1500 was the largest Spartan3 in stock)
With a Spartan3, their 90MHz design would probably reach 150MHz or more then again, I have not seen any mention of what FPGA they used. (I'm presuming they used something older)
Actually, Intel's research (before HT became reality) said that on average, the instruction decoder was issuing just under 2.5 instructions per tick out of a maximum of 3... so instruction decoder throughput in single-threaded mode is about 75% of maximum.
On AMD's side, the decoder has quadruple outputs and IIRC, AMD's average is 3 out of 4 so again 75% from maximum.
By adding SMT, Intel gave the P4 the potential to keep all instruction ports busy and AMD plans to do the same next year... a single-core A64 with SMT would be interesting but we will have to settle with dual-core dual-threaded A64s and P4s which should be interesting as well.
How do AMD and Intel manage to get 75% single-threaded when we know they will be stalled by RAM? Simple, out-of-order execution - most CPUs can look 32-128 instructions ahead to find something to do while stalled, this is necessary to maximize single-thread performance and would become unnecessary if apps and CPUs became massively multi-threaded, which appears to be what Sun is gunning for.
As far as concurrent SMT is concerned, I think four threads per CPU core will turn out to be the practical maximum for desktop chips. We will probably see this happen once the A64/P4/PM are upgraded to six execution ports, three or four years from now.
The only reason Sun can think of doing a SMTx8/32 chip is because their CPUs runs at ~1GHz. At higher speeds, they would not have the necessary timing margins to fit the extra logic to efficiently shuffle execution states between "reserve" and "active" threads.
Cable modems share the same RG59/RG6/??? cable so anything sent or received is also receivable by everyone in the same network segment... that's why cable modems are usually setup to use 3DES or AES encryption to preserve point-to-point (modem to head-end) confidentiality.
Any unencrypted transmission over a common carrier is open to sniffing and men-in-the-middle attacks so it would not be sane for an ISP or LAN of any kind to use any such media (air, power-line, phone-line, fiber, etc.) unencrypted - they are all interceptable.
AMD64 is not any less real 64bits than the PowerPC... the PowerPC was originally a 32bits chip and got 64bits extensions a few years before AMD/Intel's x86 chips.
How many pure 64bits CPUs are out there? All those I know are based on some older 8/16/32 bits, expanding and warping the instruction set and architecture along the way. AMD/Intel only carry more legacy cruft than many others along with the irregular and somewhat crufty x86 instruction set tradition... and all these bugs^H^H^H^H^H errata that had to become features for backwards compatibility.
It would make everyone's life easier if legacy architectures could be weeded out... Intel tried to go clean-slate with Itanium and that turned Itanic. Old habbits die hard - few people are willing to abandon an architecture everybody is familiar with and start anew.
1) Mailing the claim can cost as much as $10 with reception confirmation and other such options... make that $2 for plain enveloppe and international postal charges.
2) Cashing the refund can cost over $5 for people without USA-funds banking accounts.
3) Most rebates I have seen have a disclaimer that says they will be honoured at the manufacturer's sole discretion.
Because of this, I only buy into rebates when the base price suits me - FutureShop having a $110 CDN Audigy2 sale plus $45 mail-in is nice when the next lowest regular price around is $115 - this is how I discovered that USA rebates cost about $10 to claim... so that $45 rebate barely covered taxes, postage and cash-in costs so the card cost me about $110 net.
All in all, rebates are annoying and doubly so when they are in some other funds, not worth bothering with if under $20 - I prefer waiting a little longer until the "rebate" price becomes the regular price since rebates usually mean pending price adjustments and new models.
Since when is "America" a country? Last time I checked, there was more than one "America" and these were considered continents... ... AFAIK, the USA has not conquered Canada yet - but with Bush in charge, who knows.
AFAIK, patent protection is a per-country thing. Patent can be registered only with a number of countries (key markets) at ~$10k a pop or registered internationally (to every country where the patent is deemed valid) for ~$100k. In either case, only one patent for any given "invention" can be valid internationally.
If someone in Taiwan built and sold items that used some technology patented only in the USA, the taiwanese company could still sell outside the USA market - at least until whoever patented the thing decided to pay $10k for protection in supplemental markets or $100k for international coverage.
One "problem" with saving Hubble is that now, image processing techniques have made it possible to merge observations from multiple ground-based telescope and achieve better-than-Hubble resolution.
With Hubble rendered nearly obsolete by ground-based computing and sensing advances, repairing Hubble is most likely not worth it unless it is also upgraded. Assembling Hubble on ground took months, upgrading would require significant (possibly delicate) disassembly and subsequent re-assembly which probably are not reasonably feasible in open space.
This depends... Enron and Nortel's management had quite a bit of trouble with the SEC and shareholder class-action suits when their suspicious accounting practices and stock manipulations came to light.
If the court concludes that SCO's IBM lawsuit was only a diversionary tactic meant to float stock prices before management dumped theirs, they will almost certainly get investigated by the SEC and get a class-action case from their shareholders seeking compensation.
So not all is lost for the loss-making shareholders yet.
My ISP claims that less than 10% of its subscribers generate 99% of its traffic and the top 1% accounts for more than 50%, with average usage being under 2GB/month after discarding the top and bottom 5%.
Since less than 10% of people account for 99% of bandwidth, I'd say Bittorrent is the distribution equivalent of Linux on the desktop: both are well-known by the geek comunity but far from being ubiquitous in the mass market.
With all the laws and other nuisances or nonsense that will be used to outlaw, impede or prevent online distribution, it seems unlikely that BitTorrent and others will be allowed to become truly mainstream.
That's where nVidia's new SLI and ATI's AMR come in handy!
For games, frame rendering usually begins by clearing the frame buffer since the whole screen usually needs redrawing but for regular apps, updating only active windows when changes occur should be far less expensive in rendering time - these would consist mainly of mapping 2D surfaces on some polygons, something even the oldest 3D cards should be able to do decently fast.
I was thinking only about the actual contents... ... but I suppose it is reasonable to presume simplist people who watch brainwash-quality shows are indeed more likely to buy everything they see and generate more advertisement revenue.
Guess good (broadcast) shows beyond the occasional exceptions are going to disappear.
As a sidenote, I think all first and second seasons of TNG, DS9 and Voyager generally sucked compared to later seasons so Enterprise is not much of an exception.
... or else, we'll replace the show with yet another reality TV show.
Every "smarter-than-average" TV show usually getting cancelled and replaced by "reality" shows or worse is a frightening trend.
To me, it feels like the medias are sending a very strong "Thinking's BAD" message. (So I presume the USPTO's staff must be watching reality TV 24/7.)
Where I live, what people commonly call "tin foil" is actually aluminum and would be completely useless for blocking magnetic stripes.
I always thought "flavoured" usually implies artificially so - pork-flavoured pork is redundant since pork should implicitly taste like pork. How many "flavoured" things actually contain the real ingredient? Most meat favours are mostly salt and a dehydrated soup base while most others are based on extracts or completely simulated.
As I said previously, most common ingredients probably made at least one allergy victim at some point in time so the only complete allergy list is one that makes the ingredient listing redundant.
This is almost in the same domain as "Warning: contents may be hot." (on coffee cups)
Listing peanuts on the allergy warning of a jar marked 'Peanut Butter' in 1" tall lettering seems just as redundant as labelling "hot" steaming coffee.
Most allergy warning labels I have seen go like "May contain trace amounts of almond, sunflower seeds, sesame seeds, other seeds and nuts". Repeating official ingredients there is redundant: since there probably exist(ed) one person who is allergic to each common ingredient, the allergy warning list would not be complete until the whole ingredient list is repeated anyway, plus potential contaminents.
Internet is a really good place for advertisement desensitivation training.
Firefox with anti-stuff extentsions is great... among other things, my eyes can finally get a break from flashy flash ads thanks to Flashblock. (It also helps taming some flash-abusing sites.)
whereas with AMD Opteron SMP systems, every CPU gets its own dedicated HyperTransport link to its memory
I was not aware there was such a thing as "Hypertransport" DRAM... last time I checked, the Opterons and Athlon-FX had on-chip dual-channel DDR controllers.
BTW, single-core Opterons also share some RAM bandwidth with the other CPUs over the HT links and multi-core Opterons would also share RAM/HT bandwidth along its cores. Single-core Athlon64s have only their own DDR controller so dual-core Athlon64s will have to share RAM bandwidth somewhat like the Xeons do - without the extra package/bus/chipset/others latencies.
Since I just upgraded my PC and laptop, dual-threading dual-core CPUs should be standard fare by the time my next upgrades are due. As a moderate-heavy multitasker, I am looking forward to that and quad/octuple-channel FB-DIMM.
It always irritates me to some extent when I see people use "barrier" interchangeably with "landmark".
Breaking a barrier means working around or proving wrong something that used to be considered a fundamental limit. There is no fundamental limit to how much RAM can be put on a video card other than the number of data/address lines and available physical space.
A landmark is a point bound to be crossed sooner or later. Video cards will undoubtedly get more RAM sooner or later and 1GB low-end video cards will probably be common ten years from now.
Since there are 2GB SO-DIMMs available today, ATI and nVidia would most likely be able to release a 2GB video card as soon as this year if they really wanted to and enough people actually wanted/needed them.
So, the only reason we do not have >2GB video cards yet is because nobody needs them and by the time people would, the GPU would be completely obsolete.
Manufacturers are most likely not that stupid... most probably have more or less trivial flag-kill switches: changing some jumper (or solder bridge/resistor) configuration on the board, firmware upgrade, magic code/switch, etc. Of course, reworking PCBs is not free and is a major hassle if that PCB is buried in some other assembly boxed in some depot. So, magic code which can be put on stickers before shipping would be the most practical approach for appliances.
For PC-based cards, the flag probably is only a bit in the video stream which compliant players/recorders would use to disable inappropriate actions and a simple software upgrade would fix this. (So would using non-flag-compliant software.)
The new news season must be over so the slashdot news are showing reruns.
"Slashdot: yesterday's news, stuff that mattered"
Is definitely an appropriate signature... I do not remember seeing such a steady stream of reruns in the past.
But a per-cell 10A current limit would waste about 1W with 0.01ohm MOSFET switches... and current-sensing MOSFET usually drain ~1-10% through the sense lead as well. For a 12-cells battery, this is 12W more that the battery package must dissipate on top of the cells' internal losses during normal operation.
Since cells are unlikely to short one another out unless the battery pack has been physically compromised - at which point individual cells may have also been damaged, per-cell current limit seems simply wasteful overkill - pack-wise current limit would suffice for most handling accidents.
On the other hand, I am totally for built-in charge-controllers/equalizers - they prevent cell reversals and optimize battery capacity, incidentally reducing the risk of (sometimes catastrophic) premature cell failures.
In any case, no amount of charge control or current limiting can protect against manufacturing or engineering defects or end-user stupidity.
I prefer real-world black magic like the "Magic Switch" - stuff that should (not) have any effect in the real world yet actually does.
I did read some fake-telling posts... I just felt like proving that the 7805 they hypothetically used would be long dead before any sensible CPU heat got generated.