PowerShell gives you full access to.Net and Server 2012 allows every feature of the server to be configurable via powershell. There is nothing you can do with the UI that can't be done with the cli... finally....
People still use the start button? I haven't had to use it since Win7. This is actually a large reason why MS got rid of it. Their metrics showed the average person almost never used it.
I would love to see someone put a 120ton telescope into space. The clarity of a telescope is mostly a function of the amount of light it can collect, so one would still need a very large space based telescope to compete with the ground based one.
I love Chan9 and MS Research and I think a lot of what MS makes is "cool", but we are all human and mistakes WILL be made. Linux has a great track record.
This is also why BTRFS will take a while to get traction in the Enterprise. EXT4 and ZFS are still getting bug fixes.
There should be a law stating that any EULA/ToS contract longer than 8 lines of Font 10 on standard paper, requires the end user consulted by a professional lawyer and written signature of both the end user and the lawyer to be made valid. So if you have a web service with a long EULA/ToS and the customer clicks "ok", without the signatures on file, the end-user is not bound to the contract as there is no way in hell that they could have read and understood a 10 page ToS.
That should fix long unreadable contracts. Just limited to short unreadable contracts, but it's much harder to sneak something in.
Optimizing for newer 3D hardware and mutil-core CPUs typically means making non-3D hardware and single-core computer's slower. The general rule for ANYTHING in life is keep up or get left behind. Enjoy your old computers all you want, but if you don't like the way opensource is moving, fork the project and do it yourself.
At least you have the option for leaner distros. That means there are enough like minded people to at least maintain code for you.
As for me, I want my $300 GPU and 12 thread CPU to be used. Not inefficiently just for the sake of using, but my OS/Software should be capable of making use in the case it is needed.
Maybe this push will get the OpenGL people to finally get some sort of support for multi-threading. Last I checked, it only had internal multi-threading support, but not exposed through the API, and the main devels said they had no intent of adding it within their current foreseeable future.
No huge hurry, since only DX11 current supports multi-threading and AMD drivers still don't expose this and most games still don't make use of it. I just see Linux as a platform where people are more likely to attempt multi-threaded rendering, instead of releasing yet another DX9 game for quick money.
All houses? I remember asking my dad why some of the appliances had 4 prongs. Although, our breaker boxes tended to have something like 30+ 25amp breakers and every outlet had GFCI. Lighting, computer room, and entertainment room were on one phase and the garage, kitchen, laundry room, etc were on the other phase.
You could start a power-saw without the lights dimming.
Mind you, we were relatively poor. $60k/year for a 5 person house and my dad is a HUGE stickler for prices, so no commercial appliances.
You do not mod someone down merely because their logic stinks. If the person was trolling, that would be something different. GP did not seem to be, however.
blarkon clearly has done no research, does not understand the issue, or is trolling. I don't think infinitelink should have mentioned that he wanted to mod the other guy down, but infinitelink is correct. blarkon starts off his arguement jumping strait into "Google is stealing content", which is obviously false, unless one uses a really perverted meaning of content.
Personally, I think blarkon is trolling and has done a rather "good" job of it.
Nah, don't rewrite, just amend in the ability to allow citizens to issue something like a "class action suit" against a politician, then allow the a jury to assign whatever punishment they deem fit.
People voted you in, they didn't like when you did, they're sending you to life in prison. Better be careful what you do.
"depending on the manufacturing process you get between 5 and 20% unused space where 'bad' blocks come to live"
It's not just for "bad blocks" but for general wear leveling. The SSD doesn't just set aside 20% and not ever use it except corner cases, it uses it all the time, but it doesn't report that space when showing the drive size.
An extreme example of this is assume your SSD is 100% full and you want to change a block of data. Because there is no more free space, there is no more blocks to swap out with for wear leveling. If you have X% reserved, then you always have some blocks to swap out with.
Also, your theoretical best amount of data that can bet written to a SDD is a product of the amount of total storage and the amount of writes that can be handled. The practical amount of data that can be written is more a function of the interaction of how full the drive is and how much data is being written.
It's a balancing act of the trade off between usable storage and how fast the drive wears out.
ZFS supports rate limiting the cache drives, which is indicated by an interval time in units of ms and a transfer amounts indicated in an amount of bytes.
8,388,608 every 100ms would net you 80MB/s max. If you have more than one vdev for the cache, the total is split evenly among the vdevs.
Samsung 830s do very well as they fill up. Triple core CPU in the controller helps with quick garbage collection and remapping. The issue is mostly moot. Something like a 20%-30% drop for your more average SSDs. Still leaps and bounds better than rust buckets, and it only really affects writes.
Yay, race conditions(sudden power loss is an interrupt). I guess SSDs need to remap blocks atomically or update state in a way that it cane pick up where it left off.
The firmware is supposed to detect when the reserved wear leveling space has about run out and the cells are about at their limit. At this point the entire drive should suddenly enter a "real only" state with only the most recently written data subject to corruption.
This is my understanding, but I guess it's up to the firmware to handle this and I'm not sure is this is industry standard.
GPU gained more than 20x work per power used in the past 5 years. Should a 5 year old 50watt GPU be pushed over a 200watt modern GPU? Also, most modern GPUs are idle most of the time. Idle power draw is the overwhelming average load. Even when loading the GPU in games, most games can't push GPUs. Most of my games leave my GPU over 50% idle and I have an "old" ATI6950. I've been playing WoW with "Ultra" settings and that's only putting me about 8% GPU load.
I also have Civ5, BC2, and a few other games that can actually load the GPU, and even then mine is in the 60%-80% range.
My guess is the biggest benefit to lower peak TDP GPUs is not needing as much cooling to handle peak load. If GPUs really are idle most of the time, then it's mostly wasted potential. Game reviews will show little difference between high and low end and people will gravitate towards the lower end to save money. No point in forcing regulation when the market should fix itself.
Anyway, "efficiency" has been a huge priority for the past many years. Datacenters are wanting power efficient number crunchers for a while now.
Slashdot Mirror
PowerShell gives you full access to .Net and Server 2012 allows every feature of the server to be configurable via powershell. There is nothing you can do with the UI that can't be done with the cli... finally....
People still use the start button? I haven't had to use it since Win7. This is actually a large reason why MS got rid of it. Their metrics showed the average person almost never used it.
I'm watching Steam and Intel graphics drivers on Linux. Until I can play my games, my MSDN sub will let me play with whatever OS I want.
I would love to see someone put a 120ton telescope into space. The clarity of a telescope is mostly a function of the amount of light it can collect, so one would still need a very large space based telescope to compete with the ground based one.
I thought posting false details is not only against the ToS, but is against the law. We're supposed to break the law to protect ourselves?
I don't see PC-BSD in there. Mostly just augmented FreeBSD to be more desktop friendly.
I love Chan9 and MS Research and I think a lot of what MS makes is "cool", but we are all human and mistakes WILL be made. Linux has a great track record. This is also why BTRFS will take a while to get traction in the Enterprise. EXT4 and ZFS are still getting bug fixes.
Reading != comprehension
And comprehending a 16 page list of assorted exceptions doesn't help.
Speaking about reading comprehension, it seems you didn't notice this being implied, so I see you a perfect example to show my point.
There should be a law stating that any EULA/ToS contract longer than 8 lines of Font 10 on standard paper, requires the end user consulted by a professional lawyer and written signature of both the end user and the lawyer to be made valid. So if you have a web service with a long EULA/ToS and the customer clicks "ok", without the signatures on file, the end-user is not bound to the contract as there is no way in hell that they could have read and understood a 10 page ToS.
That should fix long unreadable contracts. Just limited to short unreadable contracts, but it's much harder to sneak something in.
Teach lawyers to be concise in their wording.
Optimizing for newer 3D hardware and mutil-core CPUs typically means making non-3D hardware and single-core computer's slower. The general rule for ANYTHING in life is keep up or get left behind. Enjoy your old computers all you want, but if you don't like the way opensource is moving, fork the project and do it yourself.
At least you have the option for leaner distros. That means there are enough like minded people to at least maintain code for you.
As for me, I want my $300 GPU and 12 thread CPU to be used. Not inefficiently just for the sake of using, but my OS/Software should be capable of making use in the case it is needed.
Maybe this push will get the OpenGL people to finally get some sort of support for multi-threading. Last I checked, it only had internal multi-threading support, but not exposed through the API, and the main devels said they had no intent of adding it within their current foreseeable future.
No huge hurry, since only DX11 current supports multi-threading and AMD drivers still don't expose this and most games still don't make use of it. I just see Linux as a platform where people are more likely to attempt multi-threaded rendering, instead of releasing yet another DX9 game for quick money.
Perhaps no direct money made, but added value to the platform as a whole. Not to mention data-mining user patterns.
Google mentioned that they forwarded about 4bil users/month to the French sites. Even a small percentage of that volume will be a notable difference.
All houses? I remember asking my dad why some of the appliances had 4 prongs. Although, our breaker boxes tended to have something like 30+ 25amp breakers and every outlet had GFCI. Lighting, computer room, and entertainment room were on one phase and the garage, kitchen, laundry room, etc were on the other phase.
You could start a power-saw without the lights dimming.
Mind you, we were relatively poor. $60k/year for a 5 person house and my dad is a HUGE stickler for prices, so no commercial appliances.
You do not mod someone down merely because their logic stinks. If the person was trolling, that would be something different. GP did not seem to be, however.
blarkon clearly has done no research, does not understand the issue, or is trolling. I don't think infinitelink should have mentioned that he wanted to mod the other guy down, but infinitelink is correct. blarkon starts off his arguement jumping strait into "Google is stealing content", which is obviously false, unless one uses a really perverted meaning of content.
Personally, I think blarkon is trolling and has done a rather "good" job of it.
Google isn't "symbiotic", it's "parasitic"
Just like the bacteria in your gut. Go a head, take a ton of antibiotics and see how well that goes.
When determining if someone is a witch, I always carry a duck and a scale with me.
Nah, don't rewrite, just amend in the ability to allow citizens to issue something like a "class action suit" against a politician, then allow the a jury to assign whatever punishment they deem fit.
People voted you in, they didn't like when you did, they're sending you to life in prison. Better be careful what you do.
Sounds like your driver went all "Mr Bean". "I'm in a race" - Rat Race
"depending on the manufacturing process you get between 5 and 20% unused space where 'bad' blocks come to live"
It's not just for "bad blocks" but for general wear leveling. The SSD doesn't just set aside 20% and not ever use it except corner cases, it uses it all the time, but it doesn't report that space when showing the drive size.
An extreme example of this is assume your SSD is 100% full and you want to change a block of data. Because there is no more free space, there is no more blocks to swap out with for wear leveling. If you have X% reserved, then you always have some blocks to swap out with.
Also, your theoretical best amount of data that can bet written to a SDD is a product of the amount of total storage and the amount of writes that can be handled. The practical amount of data that can be written is more a function of the interaction of how full the drive is and how much data is being written.
It's a balancing act of the trade off between usable storage and how fast the drive wears out.
ZFS supports rate limiting the cache drives, which is indicated by an interval time in units of ms and a transfer amounts indicated in an amount of bytes.
8,388,608 every 100ms would net you 80MB/s max. If you have more than one vdev for the cache, the total is split evenly among the vdevs.
based on readings, not practice.
Samsung 830s do very well as they fill up. Triple core CPU in the controller helps with quick garbage collection and remapping. The issue is mostly moot. Something like a 20%-30% drop for your more average SSDs. Still leaps and bounds better than rust buckets, and it only really affects writes.
Yay, race conditions(sudden power loss is an interrupt). I guess SSDs need to remap blocks atomically or update state in a way that it cane pick up where it left off.
The firmware is supposed to detect when the reserved wear leveling space has about run out and the cells are about at their limit. At this point the entire drive should suddenly enter a "real only" state with only the most recently written data subject to corruption.
This is my understanding, but I guess it's up to the firmware to handle this and I'm not sure is this is industry standard.
efficient != low power consumption
GPU gained more than 20x work per power used in the past 5 years. Should a 5 year old 50watt GPU be pushed over a 200watt modern GPU? Also, most modern GPUs are idle most of the time. Idle power draw is the overwhelming average load. Even when loading the GPU in games, most games can't push GPUs. Most of my games leave my GPU over 50% idle and I have an "old" ATI6950. I've been playing WoW with "Ultra" settings and that's only putting me about 8% GPU load.
I also have Civ5, BC2, and a few other games that can actually load the GPU, and even then mine is in the 60%-80% range.
My guess is the biggest benefit to lower peak TDP GPUs is not needing as much cooling to handle peak load. If GPUs really are idle most of the time, then it's mostly wasted potential. Game reviews will show little difference between high and low end and people will gravitate towards the lower end to save money. No point in forcing regulation when the market should fix itself.
Anyway, "efficiency" has been a huge priority for the past many years. Datacenters are wanting power efficient number crunchers for a while now.