most Gentoo users also use ReiserFS I will give you the SuSE default (though SuSE's install base isn't that high, and in the lion's share of large scale SuSE deployments I've worked with lately, they used autoyast files that explicitly specified ext3 instead of reiser), but I think Gentoo users generally end up thinking what they and their closest friends do with it is indicative of all Gentoo users. Most Gentoo users I knew went for ext3, I went for JFS.
they did finally concede that $.002 is different from.002 No they didn't. If they did, they would have adjusted his bill to 72 cents. Instead they did a full refund (what I would expect), apologizing for the situation, but never ever admitting they said something patently false, just implying that he wasn't understanding them correctly, and a full refund for a show of good customer satisfaction on an isolated incident. Clearly the callers making several calls to different reps have shown that verbally Verizon reps are saying a very incorrect thing (.002 cents per kB), but the writing may be correct (the only in writing thing I've seen officially is $.002/kB). More distressing that when faced with the task of selecting the correct rate from a multiple choice question of 'is it point zero zero two dollars or point zero zero two cents?', they always answer wrongly. Misreading it without the context of the misunderstanding is one thing, but to do so when asked directly to distinguish the two is insane.
Even before this happened, SuSE had already announced a move away from it. Reiser will remain a fringe filesystem, along with JFS and XFS. Each has their merits, but the ext[234] filesystems perform sufficiently for general usage, are the most widely understood in terms of data recovery, are the most widely tested filesystems, and for more high-performance requirements (i.e. database), allocating a raw block device to the software without a filesystem is not uncommon.
S3 is plain old suspend/sleep. hibernate/deep sleep implies suspend to disk and total power down, and is S4. And the word S3-induced makes no sense, S3 is a state entered into, not an active thing.
a) Maybe, if they weren't *so* eager to explicitly say and write.002 cents/kb, even when *asked* directly at bid time that they meant what they meant, to give the opportunity to clear it up, and even wrote it down on the quote, and recorded many times even faced with the explicit question to clarify..
b) Depending on his unlimited data plan, he may pay 60 dollars a month for about 2 mbit/second throughput. In court he can note that fact, and use his plan to download a 40 megabyte file, which would take about 10-15 minutes. If you can run through 35 megabytes in under 10 minutes and you could do that many many times in a month, all for 60 bucks, so it isn't so unreasonable to believe the salesman.
Liked the first result, but more significantly, an obvious flaw is brought out by that result. Note the harry potter box set lists high. Of course it would, if you own the book independently, you wouldn't buy the box set and vice-versa. To say Harry potter fans don't like Harry Potter books is rather stupid.
I will say that on my set some detail seemed to be obvious in the backdrop suggesting somehow it was more than a simple upsample. However, I was very disappointed that, effectively, 1080i didn't look appreciably less aliased than normal. Maybe some 2D textures are larger, but 3D models were indeed rendered in low res and upsampled?
Aliasing has always been my biggest gripe about PS2. Sure, geometric complexity is nice to improve, but aliasing has been the most jarring thing in 3D graphics to me.
That's not to say it is a significant finding yet, as others have pointed out, the material is an exported good so just because they trace it to the reactor does not yet conclusively link it anywhere.
I played Gran Turismo 4 in 1080i. It's the one and only PS2 game to support it, afaik. Technically HD, but admittedly since the geometric complexity was still PS2 levels it wasn't that much better...
That's why it isn't by and large traditional old GPS, it's aGPS. That's why they also can quickly get location lock instead of taking some time. And yes, I believe for E911 operation phones are required to implement some sort of way to give precise location, such as aGPS.
Speaking in the theoretical, let's say some company has significant content they wish to sell in a downloadable way, and want to have people able to download at 500 KB/s consistently.
If they use HTTP or some other similar strategy, they will have to pay for whatever connection/infrastructure that can support 500 KB/s * num of concurrent clients. To acheive that, the price they need to charge per their business case is 15 dollars a month.
Now with Bittorrent (or something like it), they can skimp a little on their infrastructure (though not too much, they need to be good seeds). Also, they can cheaply and easily implement load-sharing in a distributed fashion with seeds from various sites. Anyway, let's say to give that 500 KB/s rate to the same concurrent connections, they only have to invest 70% of the cost of the traditional method. In exchange for customers permitting their (for most people) unused upstream, they can offer the service at 4 bucks/month cheaper. Though not obvious, this is in a roundabout way like 4$/month compensation.
If you had a choice between a torrent-like service and HTTP service, the former costing 11$/month and the latter 15$/month, would you still choose to spend more money on the principle of not having your upstream used for their profit?
The problem with it is that companies are tempted to cut back too drastically on costs and give poor service. Download a popular linux distribution via torrent and you will see that bittorrent can be a very powerful/speedy strategy if someone invests in good number/quality of seed connections.
For example I'm level 20 going into some place with several level 38 monsters. When they die, they give about as much experience as three easier-to-kill things on par. I can either sit with a whole lot of enemies (getting more money) using just attack, or go after a high level, do the 'quickening' attacks a handful of times, then run away to recharge MP.
As far as not very good, maybe so, but going onto various forums I'm consistently lower than other people. One example was at Raithwall's tomb, my average member level was 13 or so, and people on forums were saying they were about level 18 or so. I didn't have to level then, though it was tough, but when I was level 16 going against Tiamat, that just was overwhelming and got devastated.
On a general level it just seems to move away from the aspects of traditional J-RPGs, but for the most part not to be original, but to mimic western-style RRGs, particularly MMORPGs.
The battle system is the most blatantly MMORPG-ish aspect. I admit random encounter rates could be maddening, but I would have much preferred a chrono-trigger style system, where encounters are obvious, and battle transition is natural, but action is more traditional J-RPG. I know a lot of people say "you don't want to admit all those selections can be distilled to a nearly equivalent set of rules" or "it's the same thing, just allowing automation". I find it difficult to distill rules that would act as I would all the time. A prime example is healing, the health level I wait for before I heal depends not on the proportion of current to max health, but to the approximate damage per hit the enemies are doing compared to remaining health. If I see an enemy that can do 90% worth of damage on occasion, my strategy for healing will me more aggresive, whereas if an enemy that does 11% of damage per hit, rules needed for the 90% case cause excessive unneeded healing. If I try to do it manually, there's no way for it to at least designate a default action, so you have to be mindful to hit x and select each one to start attacking by default. Also, in previous games if I was in a precarious fight, I might choose to leave a character with their 'action' gauge full, but skipping to other characters so they could heal instantly, but know action gauge starts filling specifically for the action being taken. Now the ability to turn on auto-attack might have been fine, because in random battles at least I would frequently just hammer attack, but the system just is frustrating.
The general UI is too MMORPG as well. To their credit, they filled the town with a lot more 'people' to make it more realistic than the huge cities with maybe 9 or 10 people common before, but they didn't want to bother making dialog for all the characters, so they put little 'talky' icons or 'shop' icons over people's heads to distinguish actual NPCs from mere decoration. This is jarring to me. Similarly, in the fields where enemies could be encountered, the health bar over every character's head is annoying to, having it off the screen in a status area solely was less distracting.
The graphics are mostly great (though the character design is not to my liking), but the music is merely 'not bad' to me, it doesn't stand out like some of the themes from previous games did. Same for the story and characters, fairly uninspired and flat (23 hours into the game at least). I didn't like FF5 or FF8s stories/characters for much the same reasons, but I still remember some of their music.
The level grinding is ridiculous in this game, absolutely have to run around in circles most of the time, rarely had to do that before.
The world layout is fine and more realistic than most 'world-map' style settings, but does allow the freedom that was lacking in FFX even early on. When I get what I'm sure is inevitable, an airship, I suspect the interface will be more like FFX, and that will disappoint me too. There was a nice sense to actively flying over the world map. Particularly after it being so tedious, experiencing the ease of flight is kinda cool.
"When comparing quadcore approaches, aggragate memory performance of multisocket, multi-memory controller designs is irrelevant."
Not necessarily, if the quadcore hinders the clock of the FSB compared to what it could be, it's highly relevant.
"When AMD announces a single die processor with multiple integrated memory controllers then it matters."
Barcelona is that.
"Offsetting AMD's memory throughput advantages are Intel's much larger caches."
True, it brings more problems into the space of operating entirely in cache, but it doesn't entirely offset it. Again, if you can't operate over a given small amount of time within either cache, AMD may do better. If you exceed the AMD cache size but fit in Intel's, Intel will absolutely do better. It's copmlicated, but not that bad.
"Meanwhile, there is no evidence that Intel MCM approach is substancially inferior performance-wise." The exception being that in the short term we see Intel clocking down FSB for quad-core. I believe Intel will have 1.3 GHz FSB again shortly, but at that point the Dual core FSB might have been higher than that. This meshes with a lot of predictions that multi-die design would be a greater load on the bus.
"I would much prefer a single processor, single memory controller system with 90% of the performance of the AMD dual proc beast. Of course, I pulled that number out of my ass..."
That is very true on the home desktop market, when I gave the qualification of 'high end' (i.e. clusters), whichever platform you choose each node is frequently multi-socket.
I doubt intercore communication in the Intel design is any better than the Barcelona (it's impossible to say this far out in any event). What is known (I have worked with some of the quad core Intel stuff) is that the two dies produce a higher load on the FSB and require the FSB be clocked down from the equivalent dual-core model. This means that AMDs remaining advantage over Intel's offerings is made more drastic (aggregate memory performance, particularly in multi-socket configurations). I.e. an Intel that thoroughly spanks a high-end two socket AMD offering linpack wise (4 flops/clock), will offer as low as 33% of the stream performance as the AMD offering. So, particularly at the high end, there remains no clear answer about which solution to pick, as Intel currently far and away has the best performance once the data has reached the cache, but if the data set being operated on within small periods of time exceeds the cache, AMD can still win. This is one of the reasons hpcc has merit for measuring multiple aspects of a cluster (i.e. aggregate memory performance, node interconnect, as well as traditional linpack tests), it's not so simple to say what is the best unless an architecture allows for superior performance across the board.
I'm a fan of JRPGs in general, and the FF series has generally been solid, though not always great. (Ignoring FFXI because the concept never appealed to me so I can't judge).
I've been playing FFXII a while and am underwhelmed, a few points -Graphically, beautiful, no complaints really. Well, except no progressive scan support. -The cities, admittedly, feel like realistic cities, well, except for the MMORPG-style icons popping up everywhere, which are actually kinda annoying. I am never bothered by the notoriously sparse cities in JRPGs, but I do have to concede they could be judged weird (a huge metropolis with maybe 13-17 people you can see is not atypical in JRPGs). -The battle system is really uninspired. Doing it manually is just not made remotely feasible in the game, and distilling your strategy into fairly hard and fast rules makes for very boring fights. Now admittedly, high rates of random encounters in previous games were monotonous as hell, but the difficult fights were more interesting, and you had the opportunity to contemplate at most any given context the risk of damage and whether heal or attack is good. In XII, the decisions are either done in advance (i.e. if health 50%, heal will happen, no thinking), or else not given a good opportunity to recognize your situation if trying the manual approach. I really liked Chrono Trigger, no random encounters, no separate battle screen, yet a traditional JRPG battle system -The story. I'm fairly far in, and it remains one of the most boring FF stories ever. Basically there is little depth and it's a fairly cookie cutter high-fantasy story with some Squareisms tossed in (Moogles, Chocobos). It's not the first fairly boring story (FFV was very much along the same lines, and FFVIII was boring to me, but maybe a little more interesting, VI, VII, and X are fairly strong IMO). -The characters. The characters are all very very flat and lacking depth, similar to the story -The music. The music is also very ho-hum. FF has a tradition of memorable music, even among other games in the series I wasn't crazy about. FFXII consists of fairly generic background music with some hints of remixes of some of their staple music from the past. I have some FF tracks play on occasion in my car, but nothing from XII appeals.
All in all it feels like they wanted to make a generic western RPG with little story and heavy inspiration from the popular MMORPGs, but done very well graphically.
This is the first single-player FF series title in a long time I think I'll pass on buying.
The thing is you have to hit the x button, and navigate every time anyone completes any action to 'try' the old way. So it's significantly a bigger PITA to do it manually.
I'll wager the majority of the population goes to vote with nothing more than the TV ads and aggressive mailings/telemarketing calls to guide them, i.e. underinformed to misinformed.
This means: -The candidate that will push policies that align well with the views of the majority of the constituency may lose to the slime ball who sinks the lowest and sways the most 'sheep' to vote for them. A strong factor is stressing things that have nothing to do with policy, or when they do focus on the opponent's voting history/plans they do so in a headline sensationalist way (i.e. 'He voted against the PATRIOT act, would you want an unpatriotic representative?'). -Your pool of likely candidates to win consist solely of those who already powerful and wealthy, who can either afford to fund such shallow campaigns or align themselves with those who do, at which point you can be assured they are likely to have no grasp on the 'common' person's experience, and/or have agreed to completely be the bitch of the rich and powerful. If a very level-headed but middle-class person of no particular extraordinary means who is independent would be the best candidate and widely popular, he will still not get into the media outlets. This used to be an unfortunate fact that people could not easily overcome, but now with the internet people have a venue to campaign, but *only* if the voters would proactively research the candidates. That's the up and down of the internet, anyone can publish, but only those that want to read will.
A lot of people argue that the uninformed balance out between both parties, and that's a bad statement to make. First the implication is that no independent or third party candidates should ever matter. Secondly, truthfully speaking, there are almost *no* uninformed voters, just those with useless information or bad information they are brainwashed into thinking is important. If they were truly uninformed, they wouldn't even recognize the words 'democrat' or 'republican' and being significant and different from 'independent', 'libertarian', 'green party' or what have you.
The result is that we have a system consisting of two parties that have ceased to mean much at all, with independent and third party candidates consisting largely of overly extreme viewpoints. If someone were, say, a republican truer to the original creed of reduced government, but not overboard, in a sane system that candidate could be libertarian, but if they are moderate they will jump into the republican camp even though the party doesn't represent that anymore, because it's the closest 'realistic' fit. Republicans *vaguely* are about reinforcing the current social structure and trusting business to manage the welfare of the economy, but not much more can really be said, some would cut back on government size and spending, some would increase both. Democrats are vaguely about increasing social programs, but again it's not a hard rule and not much else can be discerned. For example, based on the widespread principles, the occurrence of democrats rightfully criticizing a lack of fiscal responsibility in the republican party should be exactly backwards, but it's happening.
About all I can say is right on, "go vote" is empty without research. Go vote if you care enough to actively research what you vote on. If you rely on commercials or otherwise are passive on info gathering, *DON'T VOTE*. If you care, you'll research for yourself, if you don't care, don't vote.
They used to at least work at dominating new markets, and used to succeeding, they've stopped trying and just expect markets to roll over for them now...
Here they obviously have decided that the Zune will be the killer player and they don't need their former partners because "they're microsoft". They screw over existing customers and partners, to bring forth a product that may or may not be adequate (the iPod is sure as hell adequate by the market behavior), all the while not even bothering to undercut their competition in price. Considering that the iPod has, to use their terminology an 'ecosystem' of OEM support (perhaps most notable umpteen different car stereos that can take an iPod and interface intelligently with it). Consumer wise their strategy seems to be fairly boneheaded and assumes success.
Another good example of late is their HPC move 'Compute Cluster Edition'. Here they are trying to enter a market absolutely dominated by linux. Linux proves invaluable to HPC configurations because of the sheer flexibility and power of the system, with either home-grown support with absolute zero licensing fees, or working with vendors that get their licensing fees, but don't enforce things in a draconian fashion. Now an associate of mine has their 'solution' demoed and was expecting that it would work okay, but probably miss some of the point, but it worked like crap on what it did do on top of *completely* missing the point, with the MS person just shrugged and said "it's our initial release, we will improve." I fully expected they would have done better, and also expected that they would recognize the competitive landscape and aggressively price. I didn't expect them to give it away, but their per-node licensing exceeds even the 'entry server' licenses of RH and Novell, and experience teaches that commonly clusters go no higher than the 'workstation' pricing for nodes.
MS has turned into a company expecting success to fall into their laps, and I hope this complacency will bite them in the ass.
Slashdot Mirror
I mean, they're both the same if you look at them on paper-wise
Even before this happened, SuSE had already announced a move away from it. Reiser will remain a fringe filesystem, along with JFS and XFS. Each has their merits, but the ext[234] filesystems perform sufficiently for general usage, are the most widely understood in terms of data recovery, are the most widely tested filesystems, and for more high-performance requirements (i.e. database), allocating a raw block device to the software without a filesystem is not uncommon.
S3 is plain old suspend/sleep. hibernate/deep sleep implies suspend to disk and total power down, and is S4. And the word S3-induced makes no sense, S3 is a state entered into, not an active thing.
a) Maybe, if they weren't *so* eager to explicitly say and write .002 cents/kb, even when *asked* directly at bid time that they meant what they meant, to give the opportunity to clear it up, and even wrote it down on the quote, and recorded many times even faced with the explicit question to clarify..
b) Depending on his unlimited data plan, he may pay 60 dollars a month for about 2 mbit/second throughput. In court he can note that fact, and use his plan to download a 40 megabyte file, which would take about 10-15 minutes. If you can run through 35 megabytes in under 10 minutes and you could do that many many times in a month, all for 60 bucks, so it isn't so unreasonable to believe the salesman.
Liked the first result, but more significantly, an obvious flaw is brought out by that result. Note the harry potter box set lists high. Of course it would, if you own the book independently, you wouldn't buy the box set and vice-versa. To say Harry potter fans don't like Harry Potter books is rather stupid.
I will say that on my set some detail seemed to be obvious in the backdrop suggesting somehow it was more than a simple upsample. However, I was very disappointed that, effectively, 1080i didn't look appreciably less aliased than normal. Maybe some 2D textures are larger, but 3D models were indeed rendered in low res and upsampled?
Aliasing has always been my biggest gripe about PS2. Sure, geometric complexity is nice to improve, but aliasing has been the most jarring thing in 3D graphics to me.
In the last paragraph, RTFA.
That's not to say it is a significant finding yet, as others have pointed out, the material is an exported good so just because they trace it to the reactor does not yet conclusively link it anywhere.
I played Gran Turismo 4 in 1080i. It's the one and only PS2 game to support it, afaik. Technically HD, but admittedly since the geometric complexity was still PS2 levels it wasn't that much better...
That's why it isn't by and large traditional old GPS, it's aGPS. That's why they also can quickly get location lock instead of taking some time. And yes, I believe for E911 operation phones are required to implement some sort of way to give precise location, such as aGPS.
Speaking in the theoretical, let's say some company has significant content they wish to sell in a downloadable way, and want to have people able to download at 500 KB/s consistently.
If they use HTTP or some other similar strategy, they will have to pay for whatever connection/infrastructure that can support 500 KB/s * num of concurrent clients. To acheive that, the price they need to charge per their business case is 15 dollars a month.
Now with Bittorrent (or something like it), they can skimp a little on their infrastructure (though not too much, they need to be good seeds). Also, they can cheaply and easily implement load-sharing in a distributed fashion with seeds from various sites. Anyway, let's say to give that 500 KB/s rate to the same concurrent connections, they only have to invest 70% of the cost of the traditional method. In exchange for customers permitting their (for most people) unused upstream, they can offer the service at 4 bucks/month cheaper. Though not obvious, this is in a roundabout way like 4$/month compensation.
If you had a choice between a torrent-like service and HTTP service, the former costing 11$/month and the latter 15$/month, would you still choose to spend more money on the principle of not having your upstream used for their profit?
The problem with it is that companies are tempted to cut back too drastically on costs and give poor service. Download a popular linux distribution via torrent and you will see that bittorrent can be a very powerful/speedy strategy if someone invests in good number/quality of seed connections.
So if 90 degree is worst, 90 degree is medium, 135 degree is better, 180 degree should be best.
I think my work place should let me work lying down completely, for my health....
Yeah, you gamers are such dweebs, can't believe you'd say such stupid stuff.
What the hell is this number over my head counting down that suddenly appeared?
For example I'm level 20 going into some place with several level 38 monsters. When they die, they give about as much experience as three easier-to-kill things on par. I can either sit with a whole lot of enemies (getting more money) using just attack, or go after a high level, do the 'quickening' attacks a handful of times, then run away to recharge MP.
As far as not very good, maybe so, but going onto various forums I'm consistently lower than other people. One example was at Raithwall's tomb, my average member level was 13 or so, and people on forums were saying they were about level 18 or so. I didn't have to level then, though it was tough, but when I was level 16 going against Tiamat, that just was overwhelming and got devastated.
am frustrated with the game.
On a general level it just seems to move away from the aspects of traditional J-RPGs, but for the most part not to be original, but to mimic western-style RRGs, particularly MMORPGs.
The battle system is the most blatantly MMORPG-ish aspect. I admit random encounter rates could be maddening, but I would have much preferred a chrono-trigger style system, where encounters are obvious, and battle transition is natural, but action is more traditional J-RPG. I know a lot of people say "you don't want to admit all those selections can be distilled to a nearly equivalent set of rules" or "it's the same thing, just allowing automation". I find it difficult to distill rules that would act as I would all the time. A prime example is healing, the health level I wait for before I heal depends not on the proportion of current to max health, but to the approximate damage per hit the enemies are doing compared to remaining health. If I see an enemy that can do 90% worth of damage on occasion, my strategy for healing will me more aggresive, whereas if an enemy that does 11% of damage per hit, rules needed for the 90% case cause excessive unneeded healing. If I try to do it manually, there's no way for it to at least designate a default action, so you have to be mindful to hit x and select each one to start attacking by default. Also, in previous games if I was in a precarious fight, I might choose to leave a character with their 'action' gauge full, but skipping to other characters so they could heal instantly, but know action gauge starts filling specifically for the action being taken. Now the ability to turn on auto-attack might have been fine, because in random battles at least I would frequently just hammer attack, but the system just is frustrating.
The general UI is too MMORPG as well. To their credit, they filled the town with a lot more 'people' to make it more realistic than the huge cities with maybe 9 or 10 people common before, but they didn't want to bother making dialog for all the characters, so they put little 'talky' icons or 'shop' icons over people's heads to distinguish actual NPCs from mere decoration. This is jarring to me. Similarly, in the fields where enemies could be encountered, the health bar over every character's head is annoying to, having it off the screen in a status area solely was less distracting.
The graphics are mostly great (though the character design is not to my liking), but the music is merely 'not bad' to me, it doesn't stand out like some of the themes from previous games did. Same for the story and characters, fairly uninspired and flat (23 hours into the game at least). I didn't like FF5 or FF8s stories/characters for much the same reasons, but I still remember some of their music.
The level grinding is ridiculous in this game, absolutely have to run around in circles most of the time, rarely had to do that before.
The world layout is fine and more realistic than most 'world-map' style settings, but does allow the freedom that was lacking in FFX even early on. When I get what I'm sure is inevitable, an airship, I suspect the interface will be more like FFX, and that will disappoint me too. There was a nice sense to actively flying over the world map. Particularly after it being so tedious, experiencing the ease of flight is kinda cool.
The time may come when the firefox logo is visible. Of course, then we must all panic as the giant space fox has come to hump the planet...
"When comparing quadcore approaches, aggragate memory performance of multisocket, multi-memory controller designs is irrelevant."
Not necessarily, if the quadcore hinders the clock of the FSB compared to what it could be, it's highly relevant.
"When AMD announces a single die processor with multiple integrated memory controllers then it matters."
Barcelona is that.
"Offsetting AMD's memory throughput advantages are Intel's much larger caches."
True, it brings more problems into the space of operating entirely in cache, but it doesn't entirely offset it. Again, if you can't operate over a given small amount of time within either cache, AMD may do better. If you exceed the AMD cache size but fit in Intel's, Intel will absolutely do better. It's copmlicated, but not that bad.
"Meanwhile, there is no evidence that Intel MCM approach is substancially inferior performance-wise."
The exception being that in the short term we see Intel clocking down FSB for quad-core. I believe Intel will have 1.3 GHz FSB again shortly, but at that point the Dual core FSB might have been higher than that. This meshes with a lot of predictions that multi-die design would be a greater load on the bus.
"I would much prefer a single processor, single memory controller system with 90% of the performance of the AMD dual proc beast. Of course, I pulled that number out of my ass..."
That is very true on the home desktop market, when I gave the qualification of 'high end' (i.e. clusters), whichever platform you choose each node is frequently multi-socket.
I doubt intercore communication in the Intel design is any better than the Barcelona (it's impossible to say this far out in any event). What is known (I have worked with some of the quad core Intel stuff) is that the two dies produce a higher load on the FSB and require the FSB be clocked down from the equivalent dual-core model. This means that AMDs remaining advantage over Intel's offerings is made more drastic (aggregate memory performance, particularly in multi-socket configurations). I.e. an Intel that thoroughly spanks a high-end two socket AMD offering linpack wise (4 flops/clock), will offer as low as 33% of the stream performance as the AMD offering. So, particularly at the high end, there remains no clear answer about which solution to pick, as Intel currently far and away has the best performance once the data has reached the cache, but if the data set being operated on within small periods of time exceeds the cache, AMD can still win. This is one of the reasons hpcc has merit for measuring multiple aspects of a cluster (i.e. aggregate memory performance, node interconnect, as well as traditional linpack tests), it's not so simple to say what is the best unless an architecture allows for superior performance across the board.
Someone in my company had business cards with the title 'Cluster Ninja'.
Offtopic: Why am I not seeing any replies to comments? Woo hoo, no one can disagree with me!
I'm a fan of JRPGs in general, and the FF series has generally been solid, though not always great. (Ignoring FFXI because the concept never appealed to me so I can't judge).
I've been playing FFXII a while and am underwhelmed, a few points
-Graphically, beautiful, no complaints really. Well, except no progressive scan support.
-The cities, admittedly, feel like realistic cities, well, except for the MMORPG-style icons popping up everywhere, which are actually kinda annoying. I am never bothered by the notoriously sparse cities in JRPGs, but I do have to concede they could be judged weird (a huge metropolis with maybe 13-17 people you can see is not atypical in JRPGs).
-The battle system is really uninspired. Doing it manually is just not made remotely feasible in the game, and distilling your strategy into fairly hard and fast rules makes for very boring fights. Now admittedly, high rates of random encounters in previous games were monotonous as hell, but the difficult fights were more interesting, and you had the opportunity to contemplate at most any given context the risk of damage and whether heal or attack is good. In XII, the decisions are either done in advance (i.e. if health 50%, heal will happen, no thinking), or else not given a good opportunity to recognize your situation if trying the manual approach. I really liked Chrono Trigger, no random encounters, no separate battle screen, yet a traditional JRPG battle system
-The story. I'm fairly far in, and it remains one of the most boring FF stories ever. Basically there is little depth and it's a fairly cookie cutter high-fantasy story with some Squareisms tossed in (Moogles, Chocobos). It's not the first fairly boring story (FFV was very much along the same lines, and FFVIII was boring to me, but maybe a little more interesting, VI, VII, and X are fairly strong IMO).
-The characters. The characters are all very very flat and lacking depth, similar to the story
-The music. The music is also very ho-hum. FF has a tradition of memorable music, even among other games in the series I wasn't crazy about. FFXII consists of fairly generic background music with some hints of remixes of some of their staple music from the past. I have some FF tracks play on occasion in my car, but nothing from XII appeals.
All in all it feels like they wanted to make a generic western RPG with little story and heavy inspiration from the popular MMORPGs, but done very well graphically.
This is the first single-player FF series title in a long time I think I'll pass on buying.
The thing is you have to hit the x button, and navigate every time anyone completes any action to 'try' the old way. So it's significantly a bigger PITA to do it manually.
I'll wager the majority of the population goes to vote with nothing more than the TV ads and aggressive mailings/telemarketing calls to guide them, i.e. underinformed to misinformed.
This means:
-The candidate that will push policies that align well with the views of the majority of the constituency may lose to the slime ball who sinks the lowest and sways the most 'sheep' to vote for them. A strong factor is stressing things that have nothing to do with policy, or when they do focus on the opponent's voting history/plans they do so in a headline sensationalist way (i.e. 'He voted against the PATRIOT act, would you want an unpatriotic representative?').
-Your pool of likely candidates to win consist solely of those who already powerful and wealthy, who can either afford to fund such shallow campaigns or align themselves with those who do, at which point you can be assured they are likely to have no grasp on the 'common' person's experience, and/or have agreed to completely be the bitch of the rich and powerful. If a very level-headed but middle-class person of no particular extraordinary means who is independent would be the best candidate and widely popular, he will still not get into the media outlets. This used to be an unfortunate fact that people could not easily overcome, but now with the internet people have a venue to campaign, but *only* if the voters would proactively research the candidates. That's the up and down of the internet, anyone can publish, but only those that want to read will.
A lot of people argue that the uninformed balance out between both parties, and that's a bad statement to make. First the implication is that no independent or third party candidates should ever matter. Secondly, truthfully speaking, there are almost *no* uninformed voters, just those with useless information or bad information they are brainwashed into thinking is important. If they were truly uninformed, they wouldn't even recognize the words 'democrat' or 'republican' and being significant and different from 'independent', 'libertarian', 'green party' or what have you.
The result is that we have a system consisting of two parties that have ceased to mean much at all, with independent and third party candidates consisting largely of overly extreme viewpoints. If someone were, say, a republican truer to the original creed of reduced government, but not overboard, in a sane system that candidate could be libertarian, but if they are moderate they will jump into the republican camp even though the party doesn't represent that anymore, because it's the closest 'realistic' fit. Republicans *vaguely* are about reinforcing the current social structure and trusting business to manage the welfare of the economy, but not much more can really be said, some would cut back on government size and spending, some would increase both. Democrats are vaguely about increasing social programs, but again it's not a hard rule and not much else can be discerned. For example, based on the widespread principles, the occurrence of democrats rightfully criticizing a lack of fiscal responsibility in the republican party should be exactly backwards, but it's happening.
About all I can say is right on, "go vote" is empty without research. Go vote if you care enough to actively research what you vote on. If you rely on commercials or otherwise are passive on info gathering, *DON'T VOTE*. If you care, you'll research for yourself, if you don't care, don't vote.
They used to at least work at dominating new markets, and used to succeeding, they've stopped trying and just expect markets to roll over for them now...
Here they obviously have decided that the Zune will be the killer player and they don't need their former partners because "they're microsoft". They screw over existing customers and partners, to bring forth a product that may or may not be adequate (the iPod is sure as hell adequate by the market behavior), all the while not even bothering to undercut their competition in price. Considering that the iPod has, to use their terminology an 'ecosystem' of OEM support (perhaps most notable umpteen different car stereos that can take an iPod and interface intelligently with it). Consumer wise their strategy seems to be fairly boneheaded and assumes success.
Another good example of late is their HPC move 'Compute Cluster Edition'. Here they are trying to enter a market absolutely dominated by linux. Linux proves invaluable to HPC configurations because of the sheer flexibility and power of the system, with either home-grown support with absolute zero licensing fees, or working with vendors that get their licensing fees, but don't enforce things in a draconian fashion. Now an associate of mine has their 'solution' demoed and was expecting that it would work okay, but probably miss some of the point, but it worked like crap on what it did do on top of *completely* missing the point, with the MS person just shrugged and said "it's our initial release, we will improve." I fully expected they would have done better, and also expected that they would recognize the competitive landscape and aggressively price. I didn't expect them to give it away, but their per-node licensing exceeds even the 'entry server' licenses of RH and Novell, and experience teaches that commonly clusters go no higher than the 'workstation' pricing for nodes.
MS has turned into a company expecting success to fall into their laps, and I hope this complacency will bite them in the ass.