I don't understand why this should be a HTTP header and muddy things up with side-band information.  Why not just put it in-band, in the JSON request itself?  The current clients and servers are all looking for specific fields, so add a field there?  Then the parsing is very well defined instead of ad-hoc.

Pfffft, ya never know! lol

here's the same data, with the running total.

http://pastebin.com/jXqmLKuc

(pffff excel...Perl for the win)

How about the obvious again...that there may have been other 'stop loss' sales in the system, and people were (although not many) committing buy and sell orders on the way down outside of the single large sale going on thereby greatly increasing the number of transactions happening as the same coins traded multiple times on the way down.

I'm sure there were losers and winners in this whole game...also, i don't even see remotely close to a million in that list, 237457.099 is the total.

Now shoo troll.

Just FYI, added a mirror on a host of mine at http://dl0.7td.org/pub/mirror/linuxcoin/linuxcoin-v0.2a.iso -- I also updated the page for linuxcoin in the wiki at https://en.bitcoin.it/wiki/LinuxCoin

Drive on!

5870's have 11 different sensors.  I dunno about others.  They're all visible ont he i2c/smbus connection, the AMD GPU Clock tool in windows shows them all, I think GPU-Z does too, I haven't ever actualyl had a GPU based card in a Linux box, I'll know more about that next week (I work with servers.  Most of the machines I use only see a keyboard long enough to get serial console or their IPKVM working)

I don't think this is news to anyone who overclocks, or if it is, they shouldn't be overclocking.  You really need to watch ALL of your available temperatures.  The problems happen when even one exceeds the limits.  The sensor is invariably some distance away from the heat source even so it's reading will be a bit cooler (usually only a couple degC)

Many stock cards have very poor VRM cooling, or poor cooling in general.  I just got a couple 6870 Sapphires (never owned a sapphire, nor a 6870) and the VRM heatsinks are way too small to be doing overclocking.  The primary HSF looks decent though.  Untested since I'm still waiting on the motherboard to show up.

No source, no reality.  Article 1 sec 8 states what the US Gov *CAN* do, and specifically it says that it CAN MINT COIN, that it can regulate the value of that coin.  It does not give the US Gov *ANY* legal standing to make ANY other form of currency, trade, barter, etc, illegal.  Period.  Read the constitution.  That which is not EXPRESSLY allowed is forbidden -> http://en.wikipedia.org/wiki/Everything_which_is_not_forbidden_is_allowed it applies specifically to the gov, if they're not EXPLICITY, DIRECTLY allowed something, then it's forbidden.  Basically, the most they can try to do is say "bitcoin is worthless here in the US in terms of the dollar"  -- they can say that until they're blue in the face, but if private entities accept/trade it, well, nothing they can actually do.  Just like now.

Before you get too excited you might want to visit the various calculators and figure out how long it would take you to (solo) mine a block at the CURRENT difficulty, and keep in mind the more people mining,t he higher the difficulty.  This is why mining pools have become common.  http://www.alloscomp.com/bitcoin/old_calculator.php gives you a pretty good idea of the average, best, and worst cases for finding a block at a given hash rate.

That said, if you still want to do it, you have to edit your bitcoin.conf file ont he wallet machine and add rpcuser and rpcpassword, as well as rpcallowip, then you can connect to it remotely exactly as you would a pool with the username/password you set using the hostname or IP of the machine running the wallet.

Again, I recommend just joining a good pool, like eligius, slush's, or (no personal experience) btcguild.  There are other advantages than just getting your payouts much more consistently, pools do what's known as "long polling" which basically gives the client a kick as soon as the block is solved so the client/miner isn't working on a block that has already been solved.

GPU mining (should) barely use any CPU at all.  Latest startup on my windows poclbm-mod miner has poclbm using about 8 CPU minutes over about 9 hours.  Compare that to task manager using 15 minutes in 3 hours, and IDK how much in chrome/etc.  It barely even uses 1% of the CPU, the bitcoin/bitcoind probably uses more CPU to handle block processing.

As for if a 1kW PSU will be enough, and how efficient it will be, depends ENTIRELY on the PSU.  Remember most of the load is going onto the +12V rails in that scenario, and so the PSU has to be appropriately sized.  I've a 700W PSU that has PCIe connectors but can not have more than about 375W across both +12V rails on the PCI-express connectors (200W on one, 175W on the other rail) so it would maybe run 2 5870s.  I've another (newer) 400W PSU that can actually do almost as well (18A on either +12V rail, but a combined total of 350W).  So really, you have to see the plate on the PSU to be able to have any hope of telling.  That said, most of you're big 1kW+ PSUs are designed pretty good and heavy on +12V these days (unlike my 700W example)

He didn't even bother saving the file either LOL

I'd be interested in the 5870 (Diamond) and the OCZ PSU...I'll send you a PM...

I'm not exactly sure where the statement that CPU miners are traffic ineffective comes from, nor where/why the community's popular view is such.  The 4way SSE algorithm puts one of my (VMWare ESX hosted so the GPU in it isn't usable) 6 core AMD Phenom's at 18khash/sec, and at a far lower $ cost than the spare cycles I run my GPU miners on with a pair of 5870's.  That single CPU (which is a 2.7ghz X6, 1045T) does 18k/sec at a similar cost to a 5870.  GPUs are definitely more efficient, but a well written algo making use of CPU SIMD instructions will come well within shooting distance of all but the latest highest end cards.  My dual quad core opteron running at 2.0Ghz does another 14k/sec, granted that older box isn't anywhere near as efficient either, but the CPU miners could be faster, the SSE algo may not be the fasted algorithm in all cases (it IS way faster than the c algo or crypto++).

Now if you mean because each of the threads from each of the miners has to request work individually instead of as a team, I can understand that, but you're still only doing around 400hits/sec (if the getworks/sec is to be any indication of the hitrate) so I definitely agree that the amt of traffic exchanged for each getwork is probably quite a bit.  One thing I noticed is the diablo GPU miner does not back off immediately when it gets an error.  It retries immediately quite a few times before finally backing off and the site does 500 errors right at the time the network adds a new block, which is definitely causing more traffic for you.  Probably doubling your actual traffic just sending out 500 errors instead of forcing clients to wait a moment.

Heck with blocking a bunch of people with auth errors you might make things worse because they'll all start hammering.

I know it's going to take quite some work to make BPM fly though, and it won't be easy.