AMD's 890GX Chipset - Same Graphics, Better South Bridge
by Anand Lal Shimpi on March 2, 2010 4:36 AM EST- Posted in
- CPUs
Say Goodbye to ACC, Say Hello to ASUS
One thing AMD neglected to mention in its press presentation about the 8-series chipset was the fact that it removed support for Advanced Clock Calibration (ACC) from AMD’s BIOS. You may remember that ACC was the nifty tool that first let us overclock Phenom processors higher, but later allowed us to unlock disabled cores on Phenom II X2, X3 and Athlon II X3 processors. You heard it right, apparently ACC is gone.
Well, not exactly. Apparently ASUS has figured out a way to unlock cores despite not having easy access to whatever it is ACC did. And you have three ways to enable it:
1) Turn this switch on:
Technically it's the switch to the right of this one, but you get the point
2) Hold the “4” key during POST
3) Or enable it in the BIOS
Nifty. As far as I can tell, ASUS and ASRock are the only companies implementing ACC at this point. I suspect the rest will follow once they figure out how to do it.
Facebook
Twitter
RSS
65 Comments
View All Comments
semo - Tuesday, March 2, 2010 - link
Hi Anand,I noticed something strange with my Sharkoon ( http://www.sharkoon.com/html/produkte/docking_stat...">http://www.sharkoon.com/html/produkte/d...tations/... )device recently and I thought you might find it interesting. I had a 3.5" SATA drive connected to it and I just switched off my PC. When it turned off, the case fans kept spinning (CPU fan spinning up/down constantly) and the card reader/temp sensor turned red and started beeping. It kept doing after I took off the power cord. After 10 mins of looking around and head scratching I remembered that the Sharkoon has power going to it. Unplugged that and the PC shut down. I don’t think I’ve seen this anywhere else but the Sharkoon’s USB power is actually bidirectional (the DriveLink at least). That doesn’t happen usually I think and maybe different motherboards won’t like this.
That’s unfortunate that your C300 died. I wonder why if a non essential device like drive fails, the system doesn’t POST. There shouldn’t be such a condition ever (I actually have one SATA drive that does that actually). Something as simple as removing the DVD drive belt can cause the system to POST or at least take much longer to do so. Why, the thing is not essential?
Also do you have an explanation why the Vertex LE has such good write performance compared to read. I’ve assumed that you expect the opposite from NAND flash.
Looking to full review of the 890FX, hopefully it will be more polished!
SunLord - Tuesday, March 2, 2010 - link
that has to be a defect... There is no way anyone would design something to send power into a computer via a usb port it would cause all kinda of bad voodoo for the systemSunLord - Tuesday, March 2, 2010 - link
Why does the index indicates that the 890GX is DirectX 10.1 and has UVD2 while the 790GX is DX10 and UVD1 if they are exactly the same? Is the index wrong or do these changes require no hardware tweaks?Anand Lal Shimpi - Tuesday, March 2, 2010 - link
I've updated the article a bit. The move from DX10 to 10.1 in AMD's case didn't require much of a change. Technically the 890GX is more like a 785G/790GX hybrid. Either way, performance is identical between all of the cores clocked at 700MHz.Take care,
Anand
SunLord - Wednesday, March 3, 2010 - link
Put some active cooling on it and overclock it!psychobriggsy - Tuesday, March 2, 2010 - link
At least the southbridge is better featured, with SATA3 and GigE, even if the former wasn't really tested in this review, and the latter wasn't utilised.Shame that AMD didn't bump the shader count to (e.g.,) 60, it would have made a massive difference since Intel actually put some effort in on their recent attempts. Then again, Cedar 5450 should have had 160 shaders in my opinion to make it a reasonable low-end purchase.
Clearly it's a tide-over chipset until Llano changes everything.
nice123 - Tuesday, March 2, 2010 - link
They can't boost it to 60 sadly because they are arranged in blocks of 40 - the next step up is 80, which is of course Radeon 5450 territory since they decided not to add any more shaders to that and kept it exactly the same performance as the 4550.fiki959 - Tuesday, March 2, 2010 - link
I am little disappointed with the new chipset. But there is a reason that AMD didn't improve IGP performance because doing so will probably hurt radeon 5450 sales. An improvement of 30-50% will bring the IGP very close to low end dedicated cards so maybe that is the main reason or staying with 55nm process have something to with the decision I don't know..By the way I see some Athlon 2 laptops in my country, some review for these CPUs please.
shrihara - Tuesday, March 2, 2010 - link
If it has USB 3.0 which is backward compatible, then what is the need of having USB 2.0 along with that? I was hoping that AMD 890 will come up with only USB 3.0 on board like SATA 6GBps.strikeback03 - Tuesday, March 2, 2010 - link
because for whatever reason AMD didn't include USB3 and they didn't want to spend the money/PCIe lanes on a bunch of external USB3 controllers?