The "problem" you are noticing is that most software is not programmed to take advantage of multiple execution cores.
The problem in a nutshell is that writing parallel execution routines in software is not trivial.
What you point out is exactly the problem that many have been "freaking out" about for a while. That multi-core is all fine and dandy for workloads that can leverage parallelism. But for a lot of applications this is very difficult to accomplish.
In the case of this "computer" at this university, it's likely a number crunching "computer" or supercomputer. Very likely to be just a gang of machines networked together to process ridiculously parallel problems.
Not something you'll ever boot Vista on and expect to run Half Life any faster on...
TPC-C and especialy TPC-H for DW benchmarking. Buy the benchmark kit and run it...done. If you're really serious have it independantly audited and submit the results to the TPC.org. You could probably wrangle up some sponsors to help foot the bill.
If you've ever run any large site implementations of MS Exchange or MS-SQL, your limiting factor on x86 is almost always memory sometimes I/O...rarely CPU. I/O has scaled quite nicely with the ability to buy cheap PCI-X HBAs and lots of them and fast CPUs to handle interrupts, so if you've figured all that out, most shops end up being constrained by memory. This move will put serious pressure on the commercial *nix world, especially Oracle, DB2 implementations and to a lesser extent Notes on *nix.
Apparently Microsoft didn't see longterm potential for Alpha and Power CPU architectures for their commerical OSes. They do offer "high-end" version of Windows 2000/2003 server for the Itanium but you can probably count the number of customers on one hand. This annoucement of x86-64 only for their forthcoming server software stops short of spelling out the death of the Itanium support for the same reasons as the end of support for Power and Alpha architectures.
It's interesting to note that their custom OS for the Xbox 360 which I understand to be an windows XP derivative is running on the custom 3 core POWER processor. Not sure, but I doubt for something demanding high-performance that you'd want to emulate x86 instead of just re-compiling.
As for Linux and multiple architectures, it's hard to find commercial support from a big name for architectures other than x86, Itanium and Power...for end-user OS anyways. The reality is that the x86 architecture has volume out the wazoo compared to all other archs. and from a price performance perspective pretty much crushes everything out there...mostly economics that are leading the industry down this path.
Based on the reporting over at the register the densest you can go with these CPUs is 1 CPU per RU. So at most 42 * 70 watts or about 3KW of "straight" CPU power per rack...I haven't heard of any blade config that Sun has in the works. Unfortunately efficiency per rack/sq. ft. is dictated by the server form factor that Sun chooses...no different than other vendors except that there are many more server form factor choices for other CPU architectures than SPARC.
Again it all boils down to performance/per watt/$. Let's assume the capital cost to purchase is the same for the dual socket/quad core opteron X4100 as a 1U T1000...how many frames/web pages/compiles/queries/files served can the opteron do versus the SPARC T1? This is the precious info that Sun will probably never release and for good reason. Instead they'll say that their platform is 2.5 times more power efficient! So what, if I have to buy 5 times more servers to do the same work than I'm much further behind.
My understanding is that each physical core has 4 virtual cores, 3 of the 4 are in effect stopped until the one currently running stalls. So my understanding is at most you'll have 8 threads con-currently executing and 24 waiting. I imagine the OS scheduler in Solaris will be heavily tuned to make it all very smooth...likely preference will be to use the paramaters in the fair share scheduler to ensure no thread sits for too long.
Agreed that power consumption is important...I personally think liquid cooling will come back as the solution of choice (think hermetically sealed CPU in fluorinet)...but I digress.
I would love to see future SPEC benchmarks include a power consumption measurement for CPU bound tests.
Unfortunately claiming that your CPU consumes very little power means nothing unless you can quantify that it was doing more useful work per cycle per watt than the competition. Using the power consumption argument alone is very misleading, but typical Sun marketecture.
They choose to hype it's tree saving power versus it's actual performance.
Before you beat me up I agree that efficiency is an important thing to measure, and most systems benchmarks do not take efficiency into account...that being said I can bet that single threaded performance is lousy (I know they never claimed it was going to be great) but more importantly my bet is that aggregate throughput won't be that great either...my only rationale is that Sun is notorious for making crazy claims and the best they can come up with is a tree saving comparison. Puhlease.
It will be interesting to see what real world benchmark they can concoct that will show one of their single socket (8 core) servers (the T1000/T2000) using this chip outperforming a cheap dual socket (quad core) opteron server...maybe the the energy TCO will sway it. I think most commercial customers care about response time more than throughput...unless you're fully batch oriented. We'll see.
Seriously, increase your alertness by carefully consuming stimulants. Energy drinks like Redbull or a can of Rockstar...will really pick you up and as long as you're counting your calories and alloting for them you'll still be ok.
Or you could go for calorie free caffeinated water from www.buzzwater.com which is the best solution...caffeine actually speeds up your matabolism so you'll feel pumped from the extreme caffeine dosage and burn more calories.
Microsoft will likely never beat Google at it's game unless Microsoft spun off search and un-emcumbered it allowing it to do the right thing.
One of Google's key success factors has been their open source approach to delivering and developing their product offerings. The very foundation of Google is Open source backed which is the antithesis of Microsoft.
Even if MS engineers came up with a whiz bang search technology, they would force their search division to write it in.NET, host it on Windows, integrate it with IIS, leverage MS SQL, utilize WinFS...etc basically slowing them down and making them un-competitive. In the meantime Google engineers could take the same ideas and implement them much quicker with less restraint because they wouldn't get a black eye if suddenly they wanted to leverage Solaris, or Zeus, or python...or you get the idea.
I do say though that it "feels" like we are finally living in some interesting times again in IT where there are some serious players competing in the industry...
Just a guess, perhaps some Windows Group Policy admin can jump in and correct this. But doesn't Office/Windows allow you to set the default save as file type to whatever you want? Any organization that wanted to ensure interoperability with 3rd parties should be able to easily change the default.
It's certainly an option that can be set individually under Options : Save. Any decent group policy admin could roll that out to all users at their next login...
The original reclamation of Tibet was brutal. But what the Chinese did to the Tibetans is no more brutal than what the "Americans", "Canadians", "Mexicans", "Peruvians", "Bolivians",...etc did to the natives in the Americas. If anything the Tibeting history is more complex and less brutal.
Humans seem to me to be territorial and prone to violence. I'm not really condoning it, but why else would you explain the sordid history of humanity killing each other over the same piece of dirt over and over again.
I not sure I would say that I'm as hard on the Chinese communists as you are. I think the Chinese people in power woke up about 20 years ago and realized that communism and a state run economy were not going to work. They are quickly opening up their economy and they are gradually making the political transformation into something less brutal than what it was 40 years ago...will they fully eliminate communism...time will tell. Maybe they'll succeed with a new hybrid style of government where others have failed?
Strategically they will publicly deny this up until the point they are close to having a product ready for Linux. My guess is some preliminary analysis into the cost of the port has been done and the economics of supporting a Office on Linux, nothing more.
If there is a cross over point where the MS stategists feel they have more to gain by porting Office to Linux they will...that probably isn't for some years yet though, if ever. As protective as MS is, it doesn't blindly ignore trends or profittable business opportunities. Witness their 180 degree change of course when the Internet emerged as the unifying network of choice in the 90s.
If I was an MS shareholder I would be very dissapointed if MS suddenly started to play fair and eroded the value of my investment by supporting anything that increases competition.
Is this the same Lloyd Braun?
on
YahooTV
·
· Score: 3, Funny
George's Nemesis on Seinfeld? Serenity now...insanity at Yahoo later!
Re:Mac's Milk - Bloody Zit Froster
on
The Slurpee at 40
·
· Score: 1
To quote the famous Napolean Dynamite:
"GROSS!"
Re:Exactly what *is* the Dell aversion to AMD?
on
Dell Dumping Itanium
·
· Score: 1
While I'm a fan of the Opteron and AMD in general, Dell's decision makes sense to me.
Not knowing the explicit details of the Intel / Dell relationship there must be financial incentives for Dell to remain loyal to their partner Intel. And they don't necessarily have to be illegal or anti-competitive. Fact of the matter is that Dell hands off a lot of R&D to partners like Intel. It's doubtful that they would get similar levels of R&D support from AMD. Secondly there is a cost to doing business with more than one supplier, vendor and contract negotiations are not trivial matters, especially when dealing in the volume that Dell has.
I suspect one of the main deterrants for Dell has been availability of supply from AMD...AMD to this day struggles to supply it's customers with chips...something that Dell is not used to as numero uno with Intel.
It's funny Sun claimed 15x performance increase with Niagara about 16 months ago, but they never bothered to put that claim into any context. 15x the then 900 MHz SPARC III, I doubt it seriously. I doubt even 15x their low-end SPARC IIe in their now discontinued blades.
It appears that Sun engineers have hit a MHz wall sooner than the likes of Intel/AMD/IBM and are going extreme parallelism.
Based on what I've read the Niagara CPU will only be deployed in a single slot server...the only thing it might be useful for is front-end web servers and light-duty app servers. It doesn't sound like FP performance will be too exciting so I doubt it will find it's way into renderfarms.
I would like to see a showdown between the IBM/Toshiba Cell and Niagara.
It's my opinion that the Sun engineering team are in serious trouble.
"Don't count Sun out yet...it employs many smart people."
But is led by a couple of nutjobs (McNealy and Schwartz)...witness the hell that Carly did to HP, McNealy and Schwartz may not be as bad as Carly, but they certainly aren't doing that great with all the talent and technology assets they have at their disposal. Yes they have smart people doing cool stuff, but their ability to execute sucks ass, the SPARC has been a dog for at least 5 years now and they still don't fully admit it for obvious reasons and their delays at releasing viable product is saddening. Most CEOs would be bounced for a failure that was the SPARC 5...imagine how much money was lost right there alone! Not just a complete CPU scrapped but an entire line of systems designed for it...ouch.
Sun's track record with mergers, biggest failure in my mind is Cobalt guys and Netscape software...they have a good chance of destroying what little STK has.
STK is an OK storage company, I would argue they only really excel in TAPE and mainframe based storage solutions. Do they have any real software assets?
And while Sun has partnered with STK for Tape solutions for many years now and was a reseller how does owning them make their technology portfolio better? I also wonder what this deal will do to their relationships with other storage vendors, notable Hitachi for the high end Lightning series storage.
This doesn't change my opinion about Sun's future, I still see them eroding into irrelevance unless they change leadership.
The only things keeping Sun alive is the smart people who do good work to release solid products at Sun and the die hard customers still willing to pay a premium for technology that can be had elsewhere better, faster, cheaper.
I was wondering if it supports console redirect to the serial port or even better still serial over IP to provide a headless/keybordless managenent using the extra NIC?
That would make for a slick *nix based system in the closet...
I wasn't clear, what I meant to say was that RedHat needs to incorporate a solution natively that would be similar to VMWare...I'm aware of VMWare's long history of supporting and working with Linux, this is great but costs a lot of money. When compared to Solaris which is now basically free and does have virtualization built in it's a tough call for some.
I'm no Solaris or Windows fan per se, but RHEL is still missing a few things:
-Xen or virtualization solution like VMWare, Virtual Server, Solaris Zones -Fair Share Scheduler like in Solaris -Better management tools with better documentation, particularly GUI tools to displace Windows installs
RedHat needs to integrate/clone/whatever the following solutions:
-A fully supported Samba + LDAP solution like IDEALX, to eliminate the need for MS ActiveDisease -A mail/groupware solution with a client (suggest latching onto Mozilla project) to displace the killer MS Exchange+Oultook combo and Lotues Notes.
Don't get me wrong RHEL is a very good product given its age, but it could be better that's all.
If they can do the above in the next 12-18 months RedHat will be a serious contender...
15K RPM drives have been around for a long time, does anyone know if this is the fastest they can spin them from a physics perspective, do the bits start to fly off? Seriously what is preventing them from ramping up the speed further?
I would think in the server world where fast throughput could be used that they would at least be pushing 20K+ RPM drives by now...
Slashdot Mirror
The "problem" you are noticing is that most software is not programmed to take advantage of multiple execution cores.
The problem in a nutshell is that writing parallel execution routines in software is not trivial.
What you point out is exactly the problem that many have been "freaking out" about for a while. That multi-core is all fine and dandy for workloads that can leverage parallelism. But for a lot of applications this is very difficult to accomplish.
In the case of this "computer" at this university, it's likely a number crunching "computer" or supercomputer. Very likely to be just a gang of machines networked together to process ridiculously parallel problems.
Not something you'll ever boot Vista on and expect to run Half Life any faster on...
TPC-C and especialy TPC-H for DW benchmarking. Buy the benchmark kit and run it...done. If you're really serious have it independantly audited and submit the results to the TPC.org. You could probably wrangle up some sponsors to help foot the bill.
Good luck.
If you've ever run any large site implementations of MS Exchange or MS-SQL, your limiting factor on x86 is almost always memory sometimes I/O...rarely CPU. I/O has scaled quite nicely with the ability to buy cheap PCI-X HBAs and lots of them and fast CPUs to handle interrupts, so if you've figured all that out, most shops end up being constrained by memory. This move will put serious pressure on the commercial *nix world, especially Oracle, DB2 implementations and to a lesser extent Notes on *nix.
Apparently Microsoft didn't see longterm potential for Alpha and Power CPU architectures for their commerical OSes. They do offer "high-end" version of Windows 2000/2003 server for the Itanium but you can probably count the number of customers on one hand. This annoucement of x86-64 only for their forthcoming server software stops short of spelling out the death of the Itanium support for the same reasons as the end of support for Power and Alpha architectures.
It's interesting to note that their custom OS for the Xbox 360 which I understand to be an windows XP derivative is running on the custom 3 core POWER processor. Not sure, but I doubt for something demanding high-performance that you'd want to emulate x86 instead of just re-compiling.
As for Linux and multiple architectures, it's hard to find commercial support from a big name for architectures other than x86, Itanium and Power...for end-user OS anyways. The reality is that the x86 architecture has volume out the wazoo compared to all other archs. and from a price performance perspective pretty much crushes everything out there...mostly economics that are leading the industry down this path.
My only bitch about this chip, is Sun hasn't released performance info.
Based on the reporting over at the register the densest you can go with these CPUs is 1 CPU per RU. So at most 42 * 70 watts or about 3KW of "straight" CPU power per rack...I haven't heard of any blade config that Sun has in the works. Unfortunately efficiency per rack/sq. ft. is dictated by the server form factor that Sun chooses...no different than other vendors except that there are many more server form factor choices for other CPU architectures than SPARC.
Again it all boils down to performance/per watt/$. Let's assume the capital cost to purchase is the same for the dual socket/quad core opteron X4100 as a 1U T1000...how many frames/web pages/compiles/queries/files served can the opteron do versus the SPARC T1? This is the precious info that Sun will probably never release and for good reason. Instead they'll say that their platform is 2.5 times more power efficient! So what, if I have to buy 5 times more servers to do the same work than I'm much further behind.
My understanding is that each physical core has 4 virtual cores, 3 of the 4 are in effect stopped until the one currently running stalls. So my understanding is at most you'll have 8 threads con-currently executing and 24 waiting. I imagine the OS scheduler in Solaris will be heavily tuned to make it all very smooth...likely preference will be to use the paramaters in the fair share scheduler to ensure no thread sits for too long.
Agreed that power consumption is important...I personally think liquid cooling will come back as the solution of choice (think hermetically sealed CPU in fluorinet)...but I digress.
I would love to see future SPEC benchmarks include a power consumption measurement for CPU bound tests.
Unfortunately claiming that your CPU consumes very little power means nothing unless you can quantify that it was doing more useful work per cycle per watt than the competition. Using the power consumption argument alone is very misleading, but typical Sun marketecture.
They choose to hype it's tree saving power versus it's actual performance.
Before you beat me up I agree that efficiency is an important thing to measure, and most systems benchmarks do not take efficiency into account...that being said I can bet that single threaded performance is lousy (I know they never claimed it was going to be great) but more importantly my bet is that aggregate throughput won't be that great either...my only rationale is that Sun is notorious for making crazy claims and the best they can come up with is a tree saving comparison. Puhlease.
It will be interesting to see what real world benchmark they can concoct that will show one of their single socket (8 core) servers (the T1000/T2000) using this chip outperforming a cheap dual socket (quad core) opteron server...maybe the the energy TCO will sway it. I think most commercial customers care about response time more than throughput...unless you're fully batch oriented. We'll see.
Flame away sun biggots.
Seriously, increase your alertness by carefully consuming stimulants. Energy drinks like Redbull or a can of Rockstar...will really pick you up and as long as you're counting your calories and alloting for them you'll still be ok.
Or you could go for calorie free caffeinated water from www.buzzwater.com which is the best solution...caffeine actually speeds up your matabolism so you'll feel pumped from the extreme caffeine dosage and burn more calories.
Just don't get too addicted!
Good luck.
Microsoft will likely never beat Google at it's game unless Microsoft spun off search and un-emcumbered it allowing it to do the right thing.
.NET, host it on Windows, integrate it with IIS, leverage MS SQL, utilize WinFS...etc basically slowing them down and making them un-competitive. In the meantime Google engineers could take the same ideas and implement them much quicker with less restraint because they wouldn't get a black eye if suddenly they wanted to leverage Solaris, or Zeus, or python...or you get the idea.
One of Google's key success factors has been their open source approach to delivering and developing their product offerings. The very foundation of Google is Open source backed which is the antithesis of Microsoft.
Even if MS engineers came up with a whiz bang search technology, they would force their search division to write it in
I do say though that it "feels" like we are finally living in some interesting times again in IT where there are some serious players competing in the industry...
Eventually technology will replace refs. Maybe once it's 100% infallable, maybe before then.
Soon enough technology will replace everything...including you and whatever it is you probably do!
Patience my child! Mwuuuhahahahahaha!
Just a guess, perhaps some Windows Group Policy admin can jump in and correct this. But doesn't Office/Windows allow you to set the default save as file type to whatever you want? Any organization that wanted to ensure interoperability with 3rd parties should be able to easily change the default.
It's certainly an option that can be set individually under Options : Save. Any decent group policy admin could roll that out to all users at their next login...
I don't feel like I'm defending the Chinese anymore than I would defend the Lion for eating a gazelle. Life sucks and can be incredibly unfair...
The original reclamation of Tibet was brutal. But what the Chinese did to the Tibetans is no more brutal than what the "Americans", "Canadians", "Mexicans", "Peruvians", "Bolivians",...etc did to the natives in the Americas. If anything the Tibeting history is more complex and less brutal.
Humans seem to me to be territorial and prone to violence. I'm not really condoning it, but why else would you explain the sordid history of humanity killing each other over the same piece of dirt over and over again.
I not sure I would say that I'm as hard on the Chinese communists as you are. I think the Chinese people in power woke up about 20 years ago and realized that communism and a state run economy were not going to work. They are quickly opening up their economy and they are gradually making the political transformation into something less brutal than what it was 40 years ago...will they fully eliminate communism...time will tell. Maybe they'll succeed with a new hybrid style of government where others have failed?
Strategically they will publicly deny this up until the point they are close to having a product ready for Linux. My guess is some preliminary analysis into the cost of the port has been done and the economics of supporting a Office on Linux, nothing more.
If there is a cross over point where the MS stategists feel they have more to gain by porting Office to Linux they will...that probably isn't for some years yet though, if ever. As protective as MS is, it doesn't blindly ignore trends or profittable business opportunities. Witness their 180 degree change of course when the Internet emerged as the unifying network of choice in the 90s.
If I was an MS shareholder I would be very dissapointed if MS suddenly started to play fair and eroded the value of my investment by supporting anything that increases competition.
George's Nemesis on Seinfeld? Serenity now...insanity at Yahoo later!
Lloyd Braun
To quote the famous Napolean Dynamite:
"GROSS!"
While I'm a fan of the Opteron and AMD in general, Dell's decision makes sense to me.
Not knowing the explicit details of the Intel / Dell relationship there must be financial incentives for Dell to remain loyal to their partner Intel. And they don't necessarily have to be illegal or anti-competitive. Fact of the matter is that Dell hands off a lot of R&D to partners like Intel. It's doubtful that they would get similar levels of R&D support from AMD. Secondly there is a cost to doing business with more than one supplier, vendor and contract negotiations are not trivial matters, especially when dealing in the volume that Dell has.
I suspect one of the main deterrants for Dell has been availability of supply from AMD...AMD to this day struggles to supply it's customers with chips...something that Dell is not used to as numero uno with Intel.
Is this released yet? If not when is the release date for North America?
It's funny Sun claimed 15x performance increase with Niagara about 16 months ago, but they never bothered to put that claim into any context. 15x the then 900 MHz SPARC III, I doubt it seriously. I doubt even 15x their low-end SPARC IIe in their now discontinued blades.
It appears that Sun engineers have hit a MHz wall sooner than the likes of Intel/AMD/IBM and are going extreme parallelism.
Based on what I've read the Niagara CPU will only be deployed in a single slot server...the only thing it might be useful for is front-end web servers and light-duty app servers. It doesn't sound like FP performance will be too exciting so I doubt it will find it's way into renderfarms.
I would like to see a showdown between the IBM/Toshiba Cell and Niagara.
It's my opinion that the Sun engineering team are in serious trouble.
"Don't count Sun out yet...it employs many smart people."
But is led by a couple of nutjobs (McNealy and Schwartz)...witness the hell that Carly did to HP, McNealy and Schwartz may not be as bad as Carly, but they certainly aren't doing that great with all the talent and technology assets they have at their disposal. Yes they have smart people doing cool stuff, but their ability to execute sucks ass, the SPARC has been a dog for at least 5 years now and they still don't fully admit it for obvious reasons and their delays at releasing viable product is saddening. Most CEOs would be bounced for a failure that was the SPARC 5...imagine how much money was lost right there alone! Not just a complete CPU scrapped but an entire line of systems designed for it...ouch.
Sun's track record with mergers, biggest failure in my mind is Cobalt guys and Netscape software...they have a good chance of destroying what little STK has.
STK is an OK storage company, I would argue they only really excel in TAPE and mainframe based storage solutions. Do they have any real software assets?
And while Sun has partnered with STK for Tape solutions for many years now and was a reseller how does owning them make their technology portfolio better? I also wonder what this deal will do to their relationships with other storage vendors, notable Hitachi for the high end Lightning series storage.
This doesn't change my opinion about Sun's future, I still see them eroding into irrelevance unless they change leadership.
The only things keeping Sun alive is the smart people who do good work to release solid products at Sun and the die hard customers still willing to pay a premium for technology that can be had elsewhere better, faster, cheaper.
I was wondering if it supports console redirect to the serial port or even better still serial over IP to provide a headless/keybordless managenent using the extra NIC?
That would make for a slick *nix based system in the closet...
I wasn't clear, what I meant to say was that RedHat needs to incorporate a solution natively that would be similar to VMWare...I'm aware of VMWare's long history of supporting and working with Linux, this is great but costs a lot of money. When compared to Solaris which is now basically free and does have virtualization built in it's a tough call for some.
I'm no Solaris or Windows fan per se, but RHEL is still missing a few things:
-Xen or virtualization solution like VMWare, Virtual Server, Solaris Zones
-Fair Share Scheduler like in Solaris
-Better management tools with better documentation, particularly GUI tools to displace Windows installs
RedHat needs to integrate/clone/whatever the following solutions:
-A fully supported Samba + LDAP solution like IDEALX, to eliminate the need for MS ActiveDisease
-A mail/groupware solution with a client (suggest latching onto Mozilla project) to displace the killer MS Exchange+Oultook combo and Lotues Notes.
Don't get me wrong RHEL is a very good product given its age, but it could be better that's all.
If they can do the above in the next 12-18 months RedHat will be a serious contender...
15K RPM drives have been around for a long time, does anyone know if this is the fastest they can spin them from a physics perspective, do the bits start to fly off? Seriously what is preventing them from ramping up the speed further? I would think in the server world where fast throughput could be used that they would at least be pushing 20K+ RPM drives by now...