AMD has been a fab generation behind for years, and still spanked Intel prior to the Core 2 release. And there's far more factors than just clock speed, number of cores and number of dies. Intel's replacement for their aging frontside bus is years behind; the first chip they announced would use it (Whitefield) was canceled back in 2005, and now they're promising it will debut in the later half of 2008. An integrated memory controller, which AMD has had since 2001, will debut at the same time. Using a dedicated L2 cache and adding an L3 cache has the potential to reduce cache contention. More advanced hardware virtualization (including I/O virtualization) could improve performance in consolidation (one of the areas that heavily parallel designs make the most sense).
Of course it's all speculation one way or another until real hardware leaks out. And of course subsequent releases from Intel may well erase some of AMDs advantages. But the converse is also true.
Sun put eight in-order single-issue integer only cores on a single die. AMD is putting eight full superscalar cores with branch prediction, virtualization extensions, vector units, blah, blah, etc, etc. Very different design philosophies producing chips with very different aims.
Sun's foray into more traditional processor designs - the Rock - isn't expected to ship until 2008 and will feature only four cores.
The only designs actually on the market with eight traditional cores would be the IBM POWER4 and POWER5 lines, but those are dual core dies in a multi-chip module, reminiscent of the Intel scheme.
Super Mario Bros.: The Lost Levels (orig SMB2 in Japan) (NES)
Super Mario Bros. 2 (NES)
Super Mario Bros. 3 (NES)
Dr. Mario (NES)
Super Mario Kart (SNES)
Yoshi's Safari (SNES)
Super Mario World 2: Yoshi's Island (SNES)
Super Mario RPG: Legend of the Seven Stars (SNES)
Mario Party (N64)
Mario Party 2 (N64)
Mario Golf (N64)
Mario Tennis (N64)
Paper Mario (N64)
Mario Party 3 (N64)
As well a number of Zelda titles:
Zelda II: The Adventure of Link (NES)
The Legend of Zelda: Majora's Mask (N64)
The Legend of Zelda: Oracle of Seasons (GBC)
The Legend of Zelda: Four Swords (GBA)
Then there are other franchises, like Metroid:
Metroid (NES)
Super Metroid (SNES)
Metroid Fusion (GBA)
Metroid: Zero Mission (GBA)
Or Kirby:
Kirby's Avalanche (SNES)
Kirby Super Star (SNES)
Kirby's Dream Land 3 (SNES)
Kirby 64: The Crystal Shards (N64)
Kirby & the Amazing Mirror (GBA)
There's plenty of other games out there - the original Super Smash Bros, the entire Megaman series, the Wario games, the remaining Castlevania games, the first Star Fox, etc.
And of course there's plenty of games left from other platforms. Plenty of titles continue to roll in from the Turbographix-16 and Sega Genesis. There are MSX games in Japan that could potentially be introduced to the US market. Neo Geo games are in the pipeline.
Given their current strategy of trickling the top tier games one a week (or less, some weeks) they've got enough material for a good while, even if they don't add any more systems to their repetoire.
Considering they have 9 of the top 10 best selling games (of all time, on any platform) and 14 of the top 20, they'd have to actively sabotage themselves to not dominate the top five list.
Anecdotes are nice. Hard numbers are nicer. North American sales of the PSP dropped from 183k to 181k the month after they dropped the price, while Nintendo DS sales remained steady at 471k. In Japan the situation is even worse for Sony, with the DS consistently outselling the PSP more than five to one.
If they're in stock for your area, it's highly unusual by most accounts. With the exception of bundles no online vendor has had them in stock for any length of time since launch and plenty of cities have no local stock. Friends both local and remote reported taking weeks to find them, even polling local vendors daily. Often when stock did come in it was already sold by the time they got to the store.
Depends on what you mean by "generic" components. There are plenty of generic servers cases out there designed for storage applications. A 2U rack enclosure with 12 drive bays can be had for around a grand. Add a twelve port raid controller for six or seven hundred and you've got capacity to expand to 6TB raw even if you stick to 500GB drives (currently best price per GB).
This doesn't strike me as having much to do with ZFS at all. You've been able to do a home grown NAS / SAN box for years on the cheap using commodity equipment. Take ZFS out of the picture and you just need to use a hardware raid controller or a block level RAID (like dmraid on Linux or geom on FreeBSD). There are even canned solutions for this, like OpenFiler.
That being said, this sort of solution may or may not be appropriate, depending on site needs. Sometimes support is worth it.
You're also grossly overestimating the cost of an entry-level iSCSI SAN solution. Even going with EMC, hardly the cheapest of vendors, you can pick up a 6TB solution for about $15k, not $50k. Go with a second tier vendor and you can cut that number in half.
Not that I think that home grown storage is necessarily a good fit for To be fair, Seagate does list MTBF if you look at the data sheet for SATA drives actually sold for enterprise applications.
I'll agree, however, that home grown solutions are only approrpriate for a limited number of applications where outlay costs are more important that reliability and support.
Upgrading a single package from upstream is far less disruptive than doing a full platform upgrade.
If you want to compile all by yourself, you should be using Gentoo, LFS or some other distro like that.
No thanks, I stick to using a stable platform and packaging any additional or upgraded packages I need.
Got the picture here ? It is not about "version churn game". It is about using the correct tool for the correct job. Or fulfilling the requirements. Or whatever you want to call it.
I'm not about to upgrade a couple hundred servers because I need to run newer versions of a handful of packages. It's really not a big deal to spin a few RPMs and set up a repository.
That reminds me of my Kubuntu experience. Every time I want to upgrade, there is always something that doesn't work. This is true for flash and mplayer. I'm honestly surprised that both aren't installed automatically. How else do they expect us to use YouTube?
Unforutnately there are sticky legal issues revolving redistribution. A useful build of mplayer (i.e. including mp3 and dvd support) is probably illegal, at least technically, due to patent issues.
I find it so surprising that I can use an application in 1 version of a distribution, but a new version of another distribution seems to have no knowledge of its existance.
Heh, the track record of consistency in the desktop is pretty piss poor. Both major camps have had far to much flux in terms of both binary interface and methodology. It's gotten better, and hopefully the success of some of the freedesktop.org will keep things on track.
It's not like I'm a newbie. I've been using Linux since RH 6.0. At this stage in the game, it should be totally automatic for a basic workstation, while allowing a user to partition a disk, if he chooses. Everything else should be installed automatically, and then configured after installation.
Thankfully seems to be a trend in that direction. Red Hat and Fedora, at least, have moved a fair amound of configuration into the Setup Agent that gets run on first boot. And I was just reading a discussion on the OpenSUSE forums earlier today talking about simplifying the install process.
We've definitely come a long way since Red Hat 6 (started with Slackware 95 myself; been doing Red Hat since 4.2) but there's plenty of room for improvement.
Source builds can be automated with shell scripts, though it doesn't address all the issues - the user still needs to know to open a terminal and navigate to the proper mount poiunt, dependency issues don't go away, super-user privileges are still necessary for system-wide installation. And what does it really buy you? I mean aside from excesively long install times and added complexity.
I'm generally of the opinion that source-level build and install scripts are best optimized for packagers and distributers. A typical end user really has no need or no desire to touch source code.
Otherwise stock RHEL? Would be curious to see the exact errors (and maybe the output of "rpm -qa --qf '%{name}-%{version}.%{arch}\n'").
I version-revved Ruby as well (to fix an issue with gems), but I took the somewhat simpler route of backporting from RHEL5. After installing the SRPM the only modification needed to the spec file is changing libX11-devel to xorg-x11-devel in the BuildRequires line. After that, all it takes is an "rpmbuild -ba ruby.spec" and you have shiny new packages.
(I'm a big fan of RPM even if you're building from source. Keeps everything nicely managed by the packaging system and makes repeatable builds a lot easier. Even manually version revving without a backport is usually just a matter of dropping a new tarball in/usr/src/redhat/SOURCES and updating Version, Revision and %changelog in the spec file.
Not everyone plays the version churn game, especially enterprise users. There's a reason RHEL4 has support until 2012 (hell, 2.1 doesn't fall out of support until 2009).
"Dear user: Insert the CD. Type make all; make install. Press return and go for coffee."
And come back to:
make: *** No rule to make target `all'. Stop.
And that's assuming the user knew enough to open a terminal and navigated to the appropriate directory, which you left out. Let's say after they got that error they look at INSTALL and discover your instructions forgot to include 'configure', and now they get:
configure: error: Enchant library not found or too old. Use --disable-spell to build without spell plugin.
Let's say they go ahead and disable spell-checking, even though it would be a useful feature. They type 'make install' again, like the instructions said, and get:
Maybe they're not quite frustrated enough to give up just yet, and they do a google search and discover that you forgot to tell them to run as root. Hurrah! It installed! They type the application name and get
cannot open display 0:0
Whoops, still root. Maybe they realize this (smart user!), exit, and try running again as themselves. Oh, damn, the application installed in/usr/local and there's an old copy in/usr, so they end up launching the old copy instead. Even if that wasn't the case, most users are going to want menu entries and icons without having to set them up manually.
We've all got our switches, lights, and knobs to deal with, Elgan. I mean, down here there are literally hundreds and thousands of blinking, beeping, and flashing lights, blinking and beeping and flashing - they're FLASHING and they're BEEPING. I can't stand it anymore! They're BLINKING and BEEPING and FLASHING! Why doesn't somebody pull the plug!
20/29 to 28/35mpg may not be as drastic as the Prius or Civic, but I wouldn't call it slight either. Likewise I wouldn't call a bump from 244 to 253hp "much higher". Were you maybe comparing the 4 cylinder models to the 6 cylinder hybrid?
Plenty of other hybrids don't focus on power. The Ford Escape takes a signifigant hit in power on the Hybrid (153hp and 152lb-ft of torque drops to 133 and 124) and bumps fuel economy from 23/26 to 34/30.
The hybrid model of the Saturn VUE does increase horsepower over the other 4 cyl model (144hp and 152lb-ft to 170 and 160) but still manages to push economy from 22/27 to 27/32.
Or how about the Nissan Altima? Power drops from 170hp and 175lb-ft to 158 and 162. Fuel economy jumps from 26/35 to 41/36.
The Chevy Silverado hybrid is a bit of a weird beast. Only 2mpg increase with no power increase (it's literally the same v8 as it's more conventional brothers), but it's using a less complicated partial hybrid system so the cost increase is minimal. Real selling point here, in my opinion, is that the generator can be used to drive a 20A circuit with standard L5-15 sockets. Allowing people to run tools off their truck on job sites is a fabulous idea.
In other words, you need to look a bit closer at what's out there. Most hybrids on the market are decidedly not tuned to increased performance. In the case of Honda it makes perfect sense to position the Accord as the high performance hybrid since they already have an economy minded model in the Civic.
may only get an extra 2mpg, but they also don't get any additional power. In that case, honestly, the real boon is that the generator can be used to drive a 20A circuit powering standard L5-15 sockets.
You'd think with the EIA open already, you might take the time to look up fuel sources for electricity production. Petroleum accounts for a whopping 3%. The leading sources are coal (49.7%), nuclear (19.3%), natural gas (18.3%) and hydroelectric (6.5%). Lollipops and unicorn kisses would be lumped under "other renewables" (2.3%).
Generally you don't optimize your design for corner cases.
No-one would ever need any more than 640KB...
Computers are general purpose machines. A media for storing video has a fairly well defined scope. Sure, there will be exceptions (television series, multi-movie sets, dozens of hours of special features few people will bother watching, etc). There's a proven workaround though - multi-disc packages. So long as you don't need to flip or swap discs to watch a movie, Joe Consumer isn't likely to care.
A better analogy might be shopping for a car. Say you don't have kids and don't have any vocational need for hauling capacity. Do you buy a GMC Yukon because someday you might go somewhere with eight friends or find yourself moving to a new apartment? Or do you buy a sensible car and say "we can always take two cars or make a few extra trips"?
True, I forgot the option for uncompressed audio. A 7.1 TrueHD track can burn through a maximum of 8.1GB/hr, even though in reality average bitrate should be far lower. Seems kind of silly to me since humans won't be able to perceive a difference, but the specs sound impressive at least.:) [Disclaimer: I actually store audio tracks lossless, but that's to support transcoding to a variety of formats without worrying about artifacts. Not something I'd consider a primary concern on movie discs though.]
Slashdot Mirror
AMD has been a fab generation behind for years, and still spanked Intel prior to the Core 2 release. And there's far more factors than just clock speed, number of cores and number of dies. Intel's replacement for their aging frontside bus is years behind; the first chip they announced would use it (Whitefield) was canceled back in 2005, and now they're promising it will debut in the later half of 2008. An integrated memory controller, which AMD has had since 2001, will debut at the same time. Using a dedicated L2 cache and adding an L3 cache has the potential to reduce cache contention. More advanced hardware virtualization (including I/O virtualization) could improve performance in consolidation (one of the areas that heavily parallel designs make the most sense).
Of course it's all speculation one way or another until real hardware leaks out. And of course subsequent releases from Intel may well erase some of AMDs advantages. But the converse is also true.
Sun put eight in-order single-issue integer only cores on a single die. AMD is putting eight full superscalar cores with branch prediction, virtualization extensions, vector units, blah, blah, etc, etc. Very different design philosophies producing chips with very different aims.
Sun's foray into more traditional processor designs - the Rock - isn't expected to ship until 2008 and will feature only four cores.
The only designs actually on the market with eight traditional cores would be the IBM POWER4 and POWER5 lines, but those are dual core dies in a multi-chip module, reminiscent of the Intel scheme.
Erm, there's a fair number of Mario titles left:
Super Mario Bros.: The Lost Levels (orig SMB2 in Japan) (NES)
Super Mario Bros. 2 (NES)
Super Mario Bros. 3 (NES)
Dr. Mario (NES)
Super Mario Kart (SNES)
Yoshi's Safari (SNES)
Super Mario World 2: Yoshi's Island (SNES)
Super Mario RPG: Legend of the Seven Stars (SNES)
Mario Party (N64)
Mario Party 2 (N64)
Mario Golf (N64)
Mario Tennis (N64)
Paper Mario (N64)
Mario Party 3 (N64)
As well a number of Zelda titles:
Zelda II: The Adventure of Link (NES)
The Legend of Zelda: Majora's Mask (N64)
The Legend of Zelda: Oracle of Seasons (GBC)
The Legend of Zelda: Four Swords (GBA)
Then there are other franchises, like Metroid:
Metroid (NES)
Super Metroid (SNES)
Metroid Fusion (GBA)
Metroid: Zero Mission (GBA)
Or Kirby:
Kirby's Avalanche (SNES)
Kirby Super Star (SNES)
Kirby's Dream Land 3 (SNES)
Kirby 64: The Crystal Shards (N64)
Kirby & the Amazing Mirror (GBA)
There's plenty of other games out there - the original Super Smash Bros, the entire Megaman series, the Wario games, the remaining Castlevania games, the first Star Fox, etc.
And of course there's plenty of games left from other platforms. Plenty of titles continue to roll in from the Turbographix-16 and Sega Genesis. There are MSX games in Japan that could potentially be introduced to the US market. Neo Geo games are in the pipeline.
Given their current strategy of trickling the top tier games one a week (or less, some weeks) they've got enough material for a good while, even if they don't add any more systems to their repetoire.
You do realize people can choose not to buy games they think are too expensive, right?
Considering they have 9 of the top 10 best selling games (of all time, on any platform) and 14 of the top 20, they'd have to actively sabotage themselves to not dominate the top five list.
It would suck if it didn't, like, come with the nunchuck.
Anecdotes are nice. Hard numbers are nicer. North American sales of the PSP dropped from 183k to 181k the month after they dropped the price, while Nintendo DS sales remained steady at 471k. In Japan the situation is even worse for Sony, with the DS consistently outselling the PSP more than five to one.
If they're in stock for your area, it's highly unusual by most accounts. With the exception of bundles no online vendor has had them in stock for any length of time since launch and plenty of cities have no local stock. Friends both local and remote reported taking weeks to find them, even polling local vendors daily. Often when stock did come in it was already sold by the time they got to the store.
Depends on what you mean by "generic" components. There are plenty of generic servers cases out there designed for storage applications. A 2U rack enclosure with 12 drive bays can be had for around a grand. Add a twelve port raid controller for six or seven hundred and you've got capacity to expand to 6TB raw even if you stick to 500GB drives (currently best price per GB).
This doesn't strike me as having much to do with ZFS at all. You've been able to do a home grown NAS / SAN box for years on the cheap using commodity equipment. Take ZFS out of the picture and you just need to use a hardware raid controller or a block level RAID (like dmraid on Linux or geom on FreeBSD). There are even canned solutions for this, like OpenFiler.
That being said, this sort of solution may or may not be appropriate, depending on site needs. Sometimes support is worth it.
You're also grossly overestimating the cost of an entry-level iSCSI SAN solution. Even going with EMC, hardly the cheapest of vendors, you can pick up a 6TB solution for about $15k, not $50k. Go with a second tier vendor and you can cut that number in half.
Not that I think that home grown storage is necessarily a good fit for
To be fair, Seagate does list MTBF if you look at the data sheet for SATA drives actually sold for enterprise applications.
I'll agree, however, that home grown solutions are only approrpriate for a limited number of applications where outlay costs are more important that reliability and support.
Upgrading a single package from upstream is far less disruptive than doing a full platform upgrade.
If you want to compile all by yourself, you should be using Gentoo, LFS or some other distro like that.
No thanks, I stick to using a stable platform and packaging any additional or upgraded packages I need.
Got the picture here ? It is not about "version churn game". It is about using the correct tool for the correct job. Or fulfilling the requirements. Or whatever you want to call it.
I'm not about to upgrade a couple hundred servers because I need to run newer versions of a handful of packages. It's really not a big deal to spin a few RPMs and set up a repository.
That reminds me of my Kubuntu experience. Every time I want to upgrade, there is always something that doesn't work. This is true for flash and mplayer. I'm honestly surprised that both aren't installed automatically. How else do they expect us to use YouTube?
Unforutnately there are sticky legal issues revolving redistribution. A useful build of mplayer (i.e. including mp3 and dvd support) is probably illegal, at least technically, due to patent issues.
I find it so surprising that I can use an application in 1 version of a distribution, but a new version of another distribution seems to have no knowledge of its existance.
Heh, the track record of consistency in the desktop is pretty piss poor. Both major camps have had far to much flux in terms of both binary interface and methodology. It's gotten better, and hopefully the success of some of the freedesktop.org will keep things on track.
It's not like I'm a newbie. I've been using Linux since RH 6.0. At this stage in the game, it should be totally automatic for a basic workstation, while allowing a user to partition a disk, if he chooses. Everything else should be installed automatically, and then configured after installation.
Thankfully seems to be a trend in that direction. Red Hat and Fedora, at least, have moved a fair amound of configuration into the Setup Agent that gets run on first boot. And I was just reading a discussion on the OpenSUSE forums earlier today talking about simplifying the install process.
We've definitely come a long way since Red Hat 6 (started with Slackware 95 myself; been doing Red Hat since 4.2) but there's plenty of room for improvement.
Source builds can be automated with shell scripts, though it doesn't address all the issues - the user still needs to know to open a terminal and navigate to the proper mount poiunt, dependency issues don't go away, super-user privileges are still necessary for system-wide installation. And what does it really buy you? I mean aside from excesively long install times and added complexity.
I'm generally of the opinion that source-level build and install scripts are best optimized for packagers and distributers. A typical end user really has no need or no desire to touch source code.
Otherwise stock RHEL? Would be curious to see the exact errors (and maybe the output of "rpm -qa --qf '%{name}-%{version}.%{arch}\n'").
/usr/src/redhat/SOURCES and updating Version, Revision and %changelog in the spec file.
I version-revved Ruby as well (to fix an issue with gems), but I took the somewhat simpler route of backporting from RHEL5. After installing the SRPM the only modification needed to the spec file is changing libX11-devel to xorg-x11-devel in the BuildRequires line. After that, all it takes is an "rpmbuild -ba ruby.spec" and you have shiny new packages.
(I'm a big fan of RPM even if you're building from source. Keeps everything nicely managed by the packaging system and makes repeatable builds a lot easier. Even manually version revving without a backport is usually just a matter of dropping a new tarball in
Not everyone plays the version churn game, especially enterprise users. There's a reason RHEL4 has support until 2012 (hell, 2.1 doesn't fall out of support until 2009).
And come back to:And that's assuming the user knew enough to open a terminal and navigated to the appropriate directory, which you left out. Let's say after they got that error they look at INSTALL and discover your instructions forgot to include 'configure', and now they get:Let's say they go ahead and disable spell-checking, even though it would be a useful feature. They type 'make install' again, like the instructions said, and get:Maybe they're not quite frustrated enough to give up just yet, and they do a google search and discover that you forgot to tell them to run as root. Hurrah! It installed! They type the application name and getWhoops, still root. Maybe they realize this (smart user!), exit, and try running again as themselves. Oh, damn, the application installed in
We've all got our switches, lights, and knobs to deal with, Elgan. I mean, down here there are literally hundreds and thousands of blinking, beeping, and flashing lights, blinking and beeping and flashing - they're FLASHING and they're BEEPING. I can't stand it anymore! They're BLINKING and BEEPING and FLASHING! Why doesn't somebody pull the plug!
20/29 to 28/35mpg may not be as drastic as the Prius or Civic, but I wouldn't call it slight either. Likewise I wouldn't call a bump from 244 to 253hp "much higher". Were you maybe comparing the 4 cylinder models to the 6 cylinder hybrid?
Plenty of other hybrids don't focus on power. The Ford Escape takes a signifigant hit in power on the Hybrid (153hp and 152lb-ft of torque drops to 133 and 124) and bumps fuel economy from 23/26 to 34/30.
The hybrid model of the Saturn VUE does increase horsepower over the other 4 cyl model (144hp and 152lb-ft to 170 and 160) but still manages to push economy from 22/27 to 27/32.
Or how about the Nissan Altima? Power drops from 170hp and 175lb-ft to 158 and 162. Fuel economy jumps from 26/35 to 41/36.
The Chevy Silverado hybrid is a bit of a weird beast. Only 2mpg increase with no power increase (it's literally the same v8 as it's more conventional brothers), but it's using a less complicated partial hybrid system so the cost increase is minimal. Real selling point here, in my opinion, is that the generator can be used to drive a 20A circuit with standard L5-15 sockets. Allowing people to run tools off their truck on job sites is a fabulous idea.
In other words, you need to look a bit closer at what's out there. Most hybrids on the market are decidedly not tuned to increased performance. In the case of Honda it makes perfect sense to position the Accord as the high performance hybrid since they already have an economy minded model in the Civic.
may only get an extra 2mpg, but they also don't get any additional power. In that case, honestly, the real boon is that the generator can be used to drive a 20A circuit powering standard L5-15 sockets.
You'd think with the EIA open already, you might take the time to look up fuel sources for electricity production. Petroleum accounts for a whopping 3%. The leading sources are coal (49.7%), nuclear (19.3%), natural gas (18.3%) and hydroelectric (6.5%). Lollipops and unicorn kisses would be lumped under "other renewables" (2.3%).
And you'll pay for insurance and repairs on them for everyone, right?
You mean like IVT and AMD-V and IVT/
Somehow more choice is supposed to be bad for "us"?
Generally you don't optimize your design for corner cases.
No-one would ever need any more than 640KB...
Computers are general purpose machines. A media for storing video has a fairly well defined scope. Sure, there will be exceptions (television series, multi-movie sets, dozens of hours of special features few people will bother watching, etc). There's a proven workaround though - multi-disc packages. So long as you don't need to flip or swap discs to watch a movie, Joe Consumer isn't likely to care.
A better analogy might be shopping for a car. Say you don't have kids and don't have any vocational need for hauling capacity. Do you buy a GMC Yukon because someday you might go somewhere with eight friends or find yourself moving to a new apartment? Or do you buy a sensible car and say "we can always take two cars or make a few extra trips"?
True, I forgot the option for uncompressed audio. A 7.1 TrueHD track can burn through a maximum of 8.1GB/hr, even though in reality average bitrate should be far lower. Seems kind of silly to me since humans won't be able to perceive a difference, but the specs sound impressive at least. :) [Disclaimer: I actually store audio tracks lossless, but that's to support transcoding to a variety of formats without worrying about artifacts. Not something I'd consider a primary concern on movie discs though.]