backtop


Print 27 comment(s) - last by kenyee.. on Feb 6 at 10:39 PM


AMD "Trevally" block diagram courtesy of Chile Hardware
AMD gets ready to take on Intel's "Santa Rosa" with its next-generation "Trevally" mobile platform

AMD is currently providing partners with its next-generation Trevally mobile reference design for Socket S1 Turion processors. Trevally features AMD’s upcoming RS690T north bridge and new SB700 south-bridge.

In its mobile configuration, the RS690T supports side port and MXM technologies for video upgrades. Notebook manufacturers are able to implement an external frame buffer for the integrated Radeon X700-derived graphics core using the available side port. Manufacturers are free to use 512-mbit DDR2 memory for the external frame buffer.

While the RS690T has an integrated Radeon X700-derived graphics core, it features eight PCIe lanes for an external graphics or greater video output capabilities. The eight PCIe lanes are routable to an MXM-slot for graphics card upgrades. Additionally, the eight PCIe lanes are SDVO compatible in case manufacturers want to implement TV, VGA and DVI outputs. Two PCIe x1 links are available for PCIe Ethernet and ExpressCard on RS690T too.

Connecting to the RS690T IGP-north-bridge via a PCIe x4 interconnect is the new SB700 south-bridge. New features of the SB700 include more USB and SATA 3.0Gbps ports. SB700 increases the amount of supported USB 2.0 ports to 12 from the previous 10. Additionally, two USB 1.1 ports are supported for 14-supported USB ports.  

AMD increased the total amount of SATA 3.0Gbps ports to six with the SB700; the SB600 south-bridge only supports four SATA 3.0Gbps ports. While chipset manufacturers occasionally remove parallel ATA support completely with more SATA ports, AMD has opted to retain parallel ATA support. The Trevally reference platform uses the ATA 66/100/133 interface for a flash memory module, similar to Intel’s Robson technology.

It is unknown when AMD’s Trevally platform will make its appearance, especially since RS690 is still unavailable.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Sideport Memory
By IntelUser2000 on 2/5/2007 10:14:32 PM , Rating: 1
Latency was never a big problem as bandwidth for graphics. You can see from the "Xpress 200 for A64" review by AT that the dedicated sideport memory+UMA was only few % faster than the UMA-only version. The dedicated sideport one was even slower.

By adding dedicated memory, it contradicts the first point of IGP, lower cost.




RE: Sideport Memory
By ADDAvenger on 2/5/2007 10:54:53 PM , Rating: 3
quote:
By adding dedicated memory, it contradicts the first point of IGP, lower cost.


Not if this stuff falls inbetween a standard IGP and discrete graphics in price and performance. Not to mention this would likely be more power efficient than a discrete card, which could (depending on how well it works) be a huge plus to mobile professionals.


RE: Sideport Memory
By IntelUser2000 on 2/5/2007 11:30:52 PM , Rating: 2
quote:
Not if this stuff falls inbetween a standard IGP and discrete graphics in price and performance. Not to mention this would likely be more power efficient than a discrete card, which could (depending on how well it works) be a huge plus to mobile professionals.


Initial benchmarks for the elder version of the chipset, the RS600 for desktop shows that it performs like an X300. A standard value discrete video card nowadays is really X1300/7300. That's about to be replaced with G80/R600 based ones very soon. Integrated cards are in their own class.


RE: Sideport Memory
By ADDAvenger on 2/5/2007 11:47:23 PM , Rating: 2
Yes, this is no replacement if you want discrete graphics, but it could work well if you just want a beefier IGP.

Were those R600 tests done with system memory or this sideslot thing?


RE: Sideport Memory
By ADDAvenger on 2/5/2007 11:48:05 PM , Rating: 2
Shoot, I'll just ask for the link, that'll be quicker than a bunch of questions


RE: Sideport Memory
By IntelUser2000 on 2/5/2007 11:58:49 PM , Rating: 2
Go find out at HKEPC. Sure, it doesn't have the sideport thing, but it won't matter a lot anyway. The first sideport memory offered 1.4GB/s for the faster one, which is less than what single channel DDR200 can offer.

"As we mentioned before, the SidePort memory interface is a single 32-bit channel, which at 350MHz provides 1.4GB/s of bandwidth to the integrated graphics core. At 200MHz SidePort can only provide 800MB/s of bandwidth,"

Seriously, they won't put 2GHz 32-bit memory or something like that for sideport. This is mainly for power consumption.


RE: Sideport Memory
By heffeque on 2/6/2007 11:26:50 AM , Rating: 2
Anything is better than those crappy Intel integrated graphics. It's good to see that integrated graphics doesn't mean crappy anymore, it just means 'slow'.


RE: Sideport Memory
By Goty on 2/5/2007 10:54:57 PM , Rating: 2
Well, IGPs are also supposed to be lower power solutions, which I would think would be the main aim for a mobile chipset over cost.


RE: Sideport Memory
By JimFear on 2/6/2007 7:06:07 AM , Rating: 3
I believe the Frame Buffer is put in place to enable the laptop to go into an extreme low power state. As the memory controller is built in to the CPU it can never be turned off because the graphics chipset would lose access to memory and throw a hissy fit, but having its own frame buffer the CPU can go into a low power/idle state without problems.

Someone please correct me if i'm wrong but thats what I understood of it.


"Mac OS X is like living in a farmhouse in the country with no locks, and Windows is living in a house with bars on the windows in the bad part of town." -- Charlie Miller

Related Articles













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki