ZOMG, that is wicked power. And setop boxes and routers???? Imagine your DVR being uber kick ass!!

64bit & 8 Cores??? I dont get it. why didnt they just codename it "Viagra" instead???

Too bad this isn't windows compatible.

kwchang007Too bad this isn't windows compatible.

It's vice versa, Windows hasn't run on anything but x86 and IA64 since NT4. Microsoft could port it however. On the other hand, these CPU's aren't meant for desktop work, no desktop applications really benefit from 64 threads. Same goes for the T1, it had very specific uses.

DanTheBanjomanIt's vice versa, Windows hasn't run on anything but x86 and IA64 since NT4. Microsoft could port it however. On the other hand, these CPU's aren't meant for desktop work, no desktop applications really benefit from 64 threads. Same goes for the T1, it had very specific uses.

yup. if windows wants to their server software to run with this chip then they will have to program for it. but i doubt anything windows could ever program could handle that much power. freebsd ftw!

Easy Rhinoyup. if windows wants to their server software to run with this chip then they will have to program for it. but i doubt anything windows could ever program could handle that much power. freebsd ftw!

Earlier versions of NT could work on Alpha, PowerPC, MIPS and Sparc if not more. (Xbox360 uses a modified Win2K kernel still, so PowerPC is still possible) The T2 is still a Sparc. And Windows can easily handle that many threads and a lot of RAM, so it could be done if there was a bigger market.

DanTheBanjomanEarlier versions of NT could work on Alpha, PowerPC, MIPS and Sparc if not more. (Xbox360 uses a modified Win2K kernel still, so PowerPC is still possible) The T2 is still a Sparc. And Windows can easily handle that many threads and a lot of RAM, so it could be done if there was a bigger market.

yea they 'could work' but they never worked to the processors full potential.

Easy Rhinoyea they 'could work' but they never worked to the processors full potential.

Actually they did, Alpha, MIPS and PowerPC versions were fully functional and available to customers up to NT4. RC's of Win2K for Alpha exist as well. What makes you think they worked worse than the x86 editions?

DanTheBanjomanActually they did, Alpha, MIPS and PowerPC versions were fully functional and available to customers up to NT4. RC's of Win2K for Alpha exist as well. What makes you think they worked worse than the x86 editions?

no windows server OS has ever been able to handle the full potential of server processors like the ultrasparc. those processors were and are made specifically for unix/linux operating systems.

Easy Rhinono windows server OS has ever been able to handle the full potential of server processors like the ultrasparc. those processors were and are made specifically for unix/linux operating systems.

Ultrasparc is not specifically a server processor. There are plenty of Sparcstations out there. Though why can't Windows utilize their full potential? You told me twice now without any argument.

DanTheBanjomanUltrasparc is not specifically a server processor. There are plenty of Sparcstations out there. Though why can't Windows utilize their full potential? You told me twice now without any argument.

i didnt say it was specifically a server chip, i said windows has never been able to handle it as a server chip. windows cant utilize the full potential of an ultrasparc chip such as this because it doesnt handle memory as efficiently and it has lousy API programming. but you dont have to take my word for it, just ask any IT guy that worked with unix and windows in the mid 90s and they will tell you the same. and im not trying to turn this into a windows sucks post, cause i like windows. just not as a server because it cant match unix or linux on my simple dual core cpu, it certainly cant handle it on an 8 core 64thread beast designed for the highest demands and specifically geared to linux.

Easy Rhinoi didnt say it was specifically a server chip, i said windows has never been able to handle it as a server chip. windows cant utilize the full potential of an ultrasparc chip such as this because it doesnt handle memory as efficiently and it has lousy API programming. but you dont have to take my word for it, just ask any IT guy that worked with unix and windows in the mid 90s and they will tell you the same. and im not trying to turn this into a windows sucks post, cause i like windows. just not as a server because it cant match unix or linux on my simple dual core cpu, it certainly cant handle it on an 8 core 64thread beast designed for the highest demands and specifically geared to linux.

I'm not taking it as a Windows sucks thread, I'm seriously wondering why Windows wouldn't handle Sparcs as well as Linux though. I have problems believing memory management would be an issue. In my experience Windows servers aren't any worse than Linux servers on x86 hardware, I see no reason that they would be on other platforms. In fact, both Linux and Windows are limited to 64 CPU's, so that's not an argument.

According to this article Unix usually wins benchmarks, however apparently it has nothing to do with scaling to more CPU's. ie Windows wouldn't crap out on a Sparc because it can handle 64 threads. The difference would be the same as if it had a single CPU capable of a single thread. So how would Windows be worse on a Sparc than it is on x86? Unless of course you will be stating that Windows isn't fully utilizing x86 CPU's either which makes this a Windows vs Unix issue and not why Windows wouldn't run on Sparc.

DanTheBanjomanSo how would Windows be worse on a Sparc than it is on x86? Unless of course you will be stating that Windows isn't fully utilizing x86 CPU's either which makes this a Windows vs Unix issue and not why Windows wouldn't run on Sparc.

that is what i am getting at

Easy Rhinothat is what i am getting at

So basically you're saying Windows is less efficient than Unix on either platform? (apart from the other half of the tests that are in favor of Windows) How is that relevant to the fact that Windows could run on a Sparc just as well as on x86?

DanTheBanjomanSo basically you're saying Windows is less efficient than Unix on either platform? (apart from the other half of the tests that are in favor of Windows) How is that relevant to the fact that Windows could run on a Sparc just as well as on x86?

i said that 'i doubt windows could ever program anything that could handle that much power.' where have i contradicted anything you have said so far?

Easy Rhinoi said that 'i doubt windows could ever program anything that could handle that much power.' where have i contradicted anything you have said so far?

Well, Windows scales the same as Unix. Say Windows performs at 80% of Unix on a single core, it performs at 80% on 8 cores/64 threads compared to Unix. So they can handle that much power.

DanTheBanjomanWell, Windows scales the same as Unix. Say Windows performs at 80% of Unix on a single core, it performs at 80% on 8 cores/64 threads compared to Unix. So they can handle that much power.

i wouldnt say %80 is handling that much power. i would say it is crippling it.

Easy Rhinoi wouldnt say %80 is handling that much power. i would say it is crippling it.

You're missing the point. 80% is just a random number. The point is that Windows scales just as well. According to you there is not a single reason to use Windows because it cripples everything. I'm saying the reasons to do so would apply to high end servers as well.
And like I mentioned, there are also situations in which Windows wins. Besides that think of big MSSQL or Exchange servers. Claiming Windows is a bad performer per definition is really just fanboyism.

DanTheBanjomanYou're missing the point. 80% is just a random number. The point is that Windows scales just as well. According to you there is not a single reason to use Windows because it cripples everything. I'm saying the reasons to do so would apply to high end servers as well.
And like I mentioned, there are also situations in which Windows wins. Besides that think of big MSSQL or Exchange servers. Claiming Windows is a bad performer per definition is really just fanboyism.

windows doesnt scale as well and ive never seen a benchmark that shows windows performs better than unix as a server in any catagory. i dont see any reason to use windows as a server unless you are too lazy to understand unix and its shear power. and im not being a fanboy, im just talking from experience.

Easy Rhinowindows doesnt scale as well and ive never seen a benchmark that shows windows performs better than unix as a server in any catagory. i dont see any reason to use windows as a server unless you are too lazy to understand unix and its shear power. and im not being a fanboy, im just talking from experience.

If you actually believe there is no reason to use Windows on a server you are very narrow minded. Or the experience you speak of is very limited of course.

Generally, large servers use linux, because windows either isn't designed to handle the type of processor, or windows can not handle that many cpus. 64 threads over 8 cores would look to the os as 64 cores (I think...). I believe that Windows is stuck at 64 cores max (at least the consumer versions) so if you add another cpu in on the board then the system can't handle it. It's not how well each operating system scales to the amount of threads that are able to be handled, but more of how much resources the os consumes and how well the programs are programmed for that particular os. If windows could handle enough threads, and did support that cpu it would still be down to individual programmers to make use of that, or have a server that does a bunch of stuff at once.