| Home | Forums | Reviews | Guides | Newsgroups | Register | Search |
 |
| Thread Tools |
|
Guest
Posts: n/a
|
|
|
|
||
|
|
| |
|
Guest
Posts: n/a
|
|
|
|
|
||
|
|
| |
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
I too think the octacores planned for Nehalem would be more than enough. By
the time software catches up Intel will be down the road two more microarchtectures anyway. Image running Data Center on a 64 octacore box. "My Photoshop is too slow. What am I doing wrong?"  "Charlie Russel - MVP" <> wrote in message news:1DD88A46-2460-440C-A09A-... > Could be. But I've seen enough studies to know that it doesn't matter what > you do with a "normal" OS and applications, at some point the addition of > processors is counter-productive. (this assumes that both the OS and the > applications you're running are not specifically written for HPC, which is > a different kettle of fish.) Now in a Server environment, that's a bit > different, but as a workstation? Even someone who multi-tasks as much as I > do is pretty much maxed out at 4 cores. And honestly, I don't even push > that fourth one if I don't do virtualization. Now, virtualizing? Oh, yeah, > now we want some cores. And lots and lots and lots of memory. > > -- > Charlie. > http://msmvps.com/blogs/xperts64 > http://mvp.support.microsoft.com/profile/charlie.russel > > "Colin Barnhorst" <> wrote in message > news:5317700C-B8DF-4AEE-90C9-... >> Maybe Intel has in mind leveraging the QuickConnect technology (instead >> of the FSB) to achieve some gain there. >> >> "Charlie Russel - MVP" <> wrote in message >> news:FBBD4AA0-DA0A-4983-86E0-... >>> Oh, I'm sure it will. But given the overhead of handling extra CPUs, the >>> actual _gain_ may not be all that much. Since HyperThreading isn't >>> nearly as efficient as a full core. >>> >>> -- >>> Charlie. >>> http://msmvps.com/blogs/xperts64 >>> http://mvp.support.microsoft.com/profile/charlie.russel >>> >>> "Colin Barnhorst" <> wrote in message >>> news:3FF15687-E98A-43ED-8655-... >>>> Charlie, I believe Intel's next new microarchitecture, Nehalem, will >>>> support hyperthreading. It is about a year away. See the link: >>>> http://en.wikipedia.org/wiki/Nehalem...rchitecture%29 >>>> >>>> "Charlie Russel - MVP" <> wrote in >>>> message news:832A59A6-74B1-45A7-9DFC-... >>>>>I wouldn't assume that, personally. (And that is NOT based on any >>>>>privileged knowledge of what's coming, just my SWAG based on the public >>>>>evidence.) I would be very surprised if the next version of a client >>>>>virtualization solution didn't support more than a single core. And I >>>>>think 64-bit guests are a must. >>>>> >>>>> I might add that I think the days of hyperthreading are behind us. >>>>> Multiple cores has really replaced the need for it, and the gains just >>>>> aren't there, IMO. >>>>> >>>>> -- >>>>> Charlie. >>>>> http://msmvps.com/blogs/xperts64 >>>>> http://mvp.support.microsoft.com/profile/charlie.russel >>>>> >>>>> "Bender" <> wrote in message >>>>> news:EBA53A9B-93DC-4523-97B3-... >>>>>> My main concern is in being able to use more than one processor that >>>>>> virtual pc currently is limited to. When the 8 core, with Hyper >>>>>> Threading for 16 simultaneous threads comes out, only being able to >>>>>> use one core would seem very inadequate. I don't want to run a >>>>>> separate computer just to run server so I can use Hyper-V. >>>>>> I am assuming that Hyper-V will be the only option that allows >>>>>> multiple cores and a 64 bit guest. That virtual PC won't be rewritten >>>>>> when Hyper-V can do the job. >>>>>> >>>>>> >>>>>> "Bender" <> wrote in message >>>>>> news:0994A907-8449-43B5-AA2B-... >>>>>>> When does the Vista client version come out? January 2010 ? >>>>>>> >>>>>>> "Charlie Russel - MVP" <> wrote in >>>>>>> message news:FE27DFE1-E550-4710-B144-... >>>>>>>> http://msmvps.com/blogs/xperts64/default.aspx >>>>>>>> >>>>>>>> Download today, or wait till early July for WU/MU of your existing >>>>>>>> environments >>>>>>>> >>>>>>>> -- >>>>>>>> Charlie. >>>>>>>> http://msmvps.com/blogs/xperts64 >>>>>>>> http://mvp.support.microsoft.com/profile/charlie.russel >>>>>>>> >>>>>>>> >>>>> >>>> >>> >> > |
|
|
|
|
|||
|
Guest
Posts: n/a
|
Yes. And until we get RAM that uses less energy (and produces less heat), we
have a problem anyway. I'm now favouring Opterons again, just because I don't have to use FBDIMMs. -- Charlie. http://msmvps.com/blogs/xperts64 http://mvp.support.microsoft.com/profile/charlie.russel "Colin Barnhorst" <> wrote in message news:0060F7B8-852F-4D19-8D5D-... >I too think the octacores planned for Nehalem would be more than enough. >By the time software catches up Intel will be down the road two more >microarchtectures anyway. Image running Data Center on a 64 octacore box. >"My Photoshop is too slow. What am I doing wrong?" > > "Charlie Russel - MVP" <> wrote in message > news:1DD88A46-2460-440C-A09A-... >> Could be. But I've seen enough studies to know that it doesn't matter >> what you do with a "normal" OS and applications, at some point the >> addition of processors is counter-productive. (this assumes that both the >> OS and the applications you're running are not specifically written for >> HPC, which is a different kettle of fish.) Now in a Server environment, >> that's a bit different, but as a workstation? Even someone who >> multi-tasks as much as I do is pretty much maxed out at 4 cores. And >> honestly, I don't even push that fourth one if I don't do virtualization. >> Now, virtualizing? Oh, yeah, now we want some cores. And lots and lots >> and lots of memory. >> >> -- >> Charlie. >> http://msmvps.com/blogs/xperts64 >> http://mvp.support.microsoft.com/profile/charlie.russel >> >> "Colin Barnhorst" <> wrote in message >> news:5317700C-B8DF-4AEE-90C9-... >>> Maybe Intel has in mind leveraging the QuickConnect technology (instead >>> of the FSB) to achieve some gain there. >>> >>> "Charlie Russel - MVP" <> wrote in >>> message news:FBBD4AA0-DA0A-4983-86E0-... >>>> Oh, I'm sure it will. But given the overhead of handling extra CPUs, >>>> the actual _gain_ may not be all that much. Since HyperThreading isn't >>>> nearly as efficient as a full core. >>>> >>>> -- >>>> Charlie. >>>> http://msmvps.com/blogs/xperts64 >>>> http://mvp.support.microsoft.com/profile/charlie.russel >>>> >>>> "Colin Barnhorst" <> wrote in message >>>> news:3FF15687-E98A-43ED-8655-... >>>>> Charlie, I believe Intel's next new microarchitecture, Nehalem, will >>>>> support hyperthreading. It is about a year away. See the link: >>>>> http://en.wikipedia.org/wiki/Nehalem...rchitecture%29 >>>>> >>>>> "Charlie Russel - MVP" <> wrote in >>>>> message news:832A59A6-74B1-45A7-9DFC-... >>>>>>I wouldn't assume that, personally. (And that is NOT based on any >>>>>>privileged knowledge of what's coming, just my SWAG based on the >>>>>>public evidence.) I would be very surprised if the next version of a >>>>>>client virtualization solution didn't support more than a single core. >>>>>>And I think 64-bit guests are a must. >>>>>> >>>>>> I might add that I think the days of hyperthreading are behind us. >>>>>> Multiple cores has really replaced the need for it, and the gains >>>>>> just aren't there, IMO. >>>>>> >>>>>> -- >>>>>> Charlie. >>>>>> http://msmvps.com/blogs/xperts64 >>>>>> http://mvp.support.microsoft.com/profile/charlie.russel >>>>>> >>>>>> "Bender" <> wrote in message >>>>>> news:EBA53A9B-93DC-4523-97B3-... >>>>>>> My main concern is in being able to use more than one processor that >>>>>>> virtual pc currently is limited to. When the 8 core, with Hyper >>>>>>> Threading for 16 simultaneous threads comes out, only being able to >>>>>>> use one core would seem very inadequate. I don't want to run a >>>>>>> separate computer just to run server so I can use Hyper-V. >>>>>>> I am assuming that Hyper-V will be the only option that allows >>>>>>> multiple cores and a 64 bit guest. That virtual PC won't be >>>>>>> rewritten when Hyper-V can do the job. >>>>>>> >>>>>>> >>>>>>> "Bender" <> wrote in message >>>>>>> news:0994A907-8449-43B5-AA2B-... >>>>>>>> When does the Vista client version come out? January 2010 ? >>>>>>>> >>>>>>>> "Charlie Russel - MVP" <> wrote in >>>>>>>> message news:FE27DFE1-E550-4710-B144-... >>>>>>>>> http://msmvps.com/blogs/xperts64/default.aspx >>>>>>>>> >>>>>>>>> Download today, or wait till early July for WU/MU of your existing >>>>>>>>> environments >>>>>>>>> >>>>>>>>> -- >>>>>>>>> Charlie. >>>>>>>>> http://msmvps.com/blogs/xperts64 >>>>>>>>> http://mvp.support.microsoft.com/profile/charlie.russel >>>>>>>>> >>>>>>>>> >>>>>> >>>>> >>>> >>> >> > |
|
|
|
|
|||
|
|
| |
|
| Thread Tools | |
|
|
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Probing Change from B2 to RTM?? | =?Utf-8?B?Sm9lbCBEIEtyYWZ0?= | ASP .Net | 2 | 11-14-2005 03:04 PM |
| ASP.NET 2.0 RTM breaks FormsAuthentication.SetAuthCookie cookie | Bill Henning | ASP .Net | 8 | 11-09-2005 04:48 PM |
| Calendar control Beta 2 to RTM - Selected Dates not showing | King Adrock | ASP .Net | 0 | 11-08-2005 02:59 PM |
| Server 2003 sp1 RTM | meanoldman | MCSE | 1 | 03-31-2005 08:22 PM |
| windows xp rtm | =?Utf-8?B?Q2xheXRvbiBSaXBsZXk=?= | Microsoft Certification | 2 | 04-10-2004 11:52 PM |
