Computex 2005 Day 2 - ULi's new Athlon 64 Chipset, G70, CrossFire & more
by Anand Lal Shimpi on June 1, 2005 1:57 AM EST- Posted in
- Trade Shows
Shuttle Introduces 945G and Pentium M SFFs
Despite the declining interest in SFFs from other manufacturers, Shuttle had a strong showing of SFF systems based on all of the latest chipsets.
First the obligatory 945G based SFF:
Of course the 945G based SFF does support Intel's Pentium D processors:
Quite possibly the more interesting SFF was the new Pentium M based SFF from Shuttle:
And once more we got to see Shuttle's HTPC SFF also built around the Pentium M processor:
Both Pentium M SFFs use the Intel 915GM chipset.
We mentioned in our first coverage of Shuttle's Pentium M solution at CES that the solution was going to launch with support for a CableCard reader. The CableCard reference design is complete and is ready to go into the next version of Media Center Edition, however it seems that concerns over DRM are preventing it from coming to market. Despite the product being done, it doesn't look like this XPC will be able to launch with CableCard support by the end of this year.
Facebook
Twitter
RSS
53 Comments
View All Comments
Icehawk - Wednesday, June 1, 2005 - link
Anyone have links for more on the i-Ram? ETA on the device?I have some older DDR sticks just sitting around that are essentially worthless, seems like the i-RAM would make me a nice swap disk. Too bad it doesn't offer a port on the back for an optional wall wart though.
Also, on the AGP & PCIe boards - can you run both at the same time? Not SLI, I mean multi-monitor stuff.
EODetroit - Wednesday, June 1, 2005 - link
#38/#41... I can't wait for those iRAM things to come out.My plan: Stripe or JBOD two of those suckers with 4GB each in them together, then load my World of Warcraft files on it, then instantaneously load new zones and massive PvP battles while everyone else is lagging out!
Woohoo!
mlittl3 - Wednesday, June 1, 2005 - link
Oops, a little mistake in my post, #38. Actually a big mistake.Please ignore all the memory bandwidth numbers. I'm getting my bits mixed up with my bytes.
DDR200 -> 1600 Mbytes/s not 1600 Mbits/s
DDR333 -> 2700 Mbytes/s not 2700 Mbits/s
The SATA numbers and everything else is correct, however. I think.
Now I'm frustrated with myself. :)
gibhunter - Wednesday, June 1, 2005 - link
You people need to get hold of yourselves and calm down about ULI. They used to make the worst pos chipsets as ALI and that's why they went under. Now rebranded as ULI you think they will be any better? Let me save you the trouble. Their chipsets are still piece of shit. I have a socket 939 motherboard that's just sitting there cause it wouldn't run at high performance settings. In fact, it corrupted my winXP install halfway through intslling drivers on it. Fresh install of windows gave me all kinds of problems thanks to a driver cd that installed winME USB drivers. So much for trying to save some time. Trust me, they went under for a reason. Stay away and stay happy. I now have two other socket 939 systems on VIA and Nvidia chipsets and sailing has been nothing but smooth. Can't say that for POS ULI which even if it runs stable can't match the other two in performance.erinlegault - Wednesday, June 1, 2005 - link
I want one of those nForce 4 Pro dual socket Tyan motherboard's with dual core Opteron 2xx's.That should last a while through our current multitasking paradigm.
mlittl3 - Wednesday, June 1, 2005 - link
#13 and anyone else, please, please read this.The Gigabyte iRAM is ONLY powered by the PCI slot. No, I repeat, NO data is moved through the PCI slot. When the PC is off, the RAM is powered by the rechargable battery.
The reason for only supporting DDR200 is because data transfers through the SATA port (the iRAM is seen as a SATA hard drive by the PC) on the PCB. So the bandwidth of SATA is 150 Mbytes/s (1200 Mbits/s) and the bandwidth of DDR200 is 200 Mbytes/s (1600 Mbits/s). Any faster memory will work (DDR333, DDR400, etc.) but will be clocked at DDR200 on the iRAM because the bottleneck is the SATA transfer bandwidth.
If they move the card to SATA II, 300 Mbytes/s (2400 Mbits/s) then they will support DDR333 which has a bandwidth of 337 Mbytes/s (2700 Mbits/s).
The iRAM will have latencies of tens of nanoseconds instead of 10 milliseconds. This is three orders of magnitude (1000 times) faster.
I know no one is going to read this and I'm going to get more frustrated.
monsoon - Wednesday, June 1, 2005 - link
looking forward to real world tests of the GIGABYTE i-RAM and the ZALMAN "hamster" cooler !;)Aquila76 - Wednesday, June 1, 2005 - link
LOL. I can't help thinking the Zalman CPU cooler should have a hamster running in it!ViperV990 - Wednesday, June 1, 2005 - link
First I was wondering why there was a VGA-out on the A8N-SLI Premium, then I realized the rear I/O panel shown is not the one from the A8N. In fact, it doesn't match anything on the page.ryanv12 - Wednesday, June 1, 2005 - link
My girlfriend is going abroad in China this summer. I think I should have her pick me up one of these ULI boards with AGP and PCI-E if they aren't going to be releasing them in the states ;)