Anything about the release date and the prise? I so hope, that it' would be early september ...
Printable View
Anything about the release date and the prise? I so hope, that it' would be early september ...
I'm curious about the DIMM slots. Shouldn't they have 3 colors now - one for each of the 3 channels?
Andypro1,
There's no standard for slot coloring so not really. Some makers choose color per channel, some assign same color for those slots that you use for DC.
Why? You'd want to populate all three channels anyway so they're just differentiated with two colors to avoid populating two channels with 2+1 DIMM.Quote:
Originally Posted by Andypro1
No? You want to fill 3 at a time. Like all the black.Quote:
Originally Posted by Andypro1
I think it's a great addition. I'm planning to get 2 15K Savvio's (one for OS, one for apps), a :banana::banana::banana::banana:load of RAM, and disable the page file. Gets rid of the regret of not going all out with RAID and buying a good hardware controller (too much $$$ and after careful thought, too little reward for my needs).Quote:
Originally Posted by lowfat
PS. the SAS controller isn't a RAID controller. That would cost a few hundred extra bucks if it was even possible (and I don't think they have software RAID for SCSI/SAS). They just added it so people can use SAS drives, which again I think is an awesome addition.
SAS support is nice - but then again SAS HDDs are now almost useless for desktops simply due to the existence of SSDs.
SAS drives are very fast, and their write speeds are faster than a lot of SSD drives. I think people now focus on synthetic benchmarks way too much, and they don't realize that SSD's aren't as great as they seem (in their current state) when it comes to real applications.Quote:
Originally Posted by largon
For an OS drive, for me at least, access speeds ('snappy feel') is more important than throughput, and the only thing that beats flash for low latency is a RAM-drive, which isn't even close in terms of ease of use, capacity, price, and so on.Quote:
Originally Posted by Baron_Davis
Baron_Davis,
As far as I understand, write speed is the least important factor for a storage solution's performance in consumer applications.
the die is actually smaller then 2 Merom stuck together to form a MCM quad core Kentsfield. The socket and package are larger due to all the extra pins needed to run 2 QPI lanes and 3 memory channels.Quote:
Originally Posted by largon
EDIT: oh you ment Tylersburg. Well it needs to carry IO for 2 QPI lanes, the logic for PCI control and all the logic required to control the power and clocks for all the seperate parts of 2 Nehalem die.
I think he meant the chipset, compared to its function. Given that the memory controller is gone one might not expect it to be big.Quote:
Originally Posted by Blauhung
Well, whenever you install a program or update your drivers, your write speed is a big factor. I'm no performance expert, but I'd like for someone to show me some real life benchmarks where SSD's are destroying or even beating 15K RPM SAS drives (or even Velociraptors). You can get a 15K Savvio 36GB drive for around 300 just for reference.Quote:
Originally Posted by largon
Say hello to my next mobo! :eek:
I (like most here) can't wait to see what Nehalem will bring to the table, it's great to see some pics of Tylersburg mobo's surfacing. :)
Am I correct that X58 is made on 65nm?
How about X38?
Baron_Davis,
Indeed, writes are "only" needed in select instances in normal consumer use. Huge majority of storage accesses are reads... Not even sequential, but random reads. Then ofcourse, if one does things heavily involving the storage - like video editing 'n' such - it's a different matter.
X38 and X48 (basically same chip) are made on 90 nm, as is P35. P45 is made on 65 nm. and so will X58.
Must be some serious logic in there...
X58's approx. transistor count?
the board looks a bit bigger hope it will be able to fit on current cases
its not bigger ;)Quote:
Originally Posted by ultimate11
http://img176.imageshack.us/img176/3783/1324vd0.jpg
I'm also worried about the size and TDP of the X58. Looks too big for what it's supposed to be.
I still think its dumb they no longer have the IHS, thats was a Good addition.
while I don't think I should give out actual die sizes, I can probably say that Tylersburg is smaller then both Bearlake (G35/P35) and Seaburg (X5400) are larger then Tylersburg, with Eaglelake (G45/P45) being smaller then Tylersburg by about the same ammount that Bearlake and Seaburg are larger. While I don't know much about what is taking up all the space that should have been freed up by no longer having integrated graphics on die like the P35/P45 have, I do know that since Tylersburg is a server chipset, there is a lot of stuff devoted to handling all the cross traffic that will be going through the QPI links. Just keep in mind that the X58 platform is a server chip that was fit into the enthusiast space because it has the PCIe lanes and overclocking functionality we want.
when 775 x58?
up to 24GB ram support?
ASUS already uses 16 phase VRM on many of their higher end boards (including the P45 based P5Q Deluxe and above).Quote:
Originally Posted by largon
Will there be P55? Haven't really stayed on top of motherboard releases.
I wonder what cooling they will put on the "northbridge." It seems like anything of significant size would block the first PCI-E slot, otherwise the layout looks good.