Initial TBred2 Results . .

Add Your Comments

I’ve gone to a number of forums and looked to see how pioneers were doing with 2400+.

Though it’s still early, a few observations can be made:

Forget An Easy 2.4GHz Presuming high-end air and not too crazy voltage (up to 1.95V) and high-end air, what can realistically be expected from the CPUs out there is somewhere between 2.2-2.3GHz. If you’re not comfortable cranking up voltage past 1.85V, expect a result closer to 2.2 than 2.3.

To realistically get 2.4GHz, get a Peltier or more.

Also remember that these chips are .13 micron, and the default voltage is 1.65V, not 1.75V. 2V+ will probably cause more burnouts than earlier generations of AMD CPUs.

These results also help to explain the absence of 2600+s. If the 2400+s could usually do 2.4GHz without too much sweat, there would be no reason for 2600+ not to be around, since there would be sufficient headroom to guarantee 2.13. Given the current results, there doesn’t seem to be that kind of headroom.

Just Getting The Act Together Virtually all 2400+ out there were made week 39 or later; there are already week 41 chips out there. This is a strong indicator AMD has only been able to make these in somewhat reasonable bulk for about the last month. It also indicates that these are getting shipped out as quickly as they are being made, with little inventory floating around the channels.

Now there is bulk and there is bulk. While the number of places that offer 2400+s have grown the last couple weeks, it still remains to be seen whether or not whether these places can keep them in stock. Given all the other evidence, though, you’d probably have to give the benefit of the doubt to AMD for 2400+s at this point.

Where Will Pricing Go? Pricing still remains pretty high, and a good test of AMD’s ability to get product out will be the price.

If prices drop significantly (i.e. more than 10%) over the next two weeks in the US, that’s probably a pretty good sign AMD can reliably supply these things. If not, scarcity will keep prices relatively high.

The appearance of TBredBs in lower speeds (i.e. 2200+ or 2000+) would be another good indicator, but no sign of that yet.

Variations There’s already been a number of variations on the initial AIUAB 2400+. AIUCB is a pretty common code, and I’ve seen codes starting with “AJ” and “AIV.”

While AMD is known to put out a lot of variants on a model of CPU, four variations within a few weeks is quite a bit even for them, so they are obviously scrambling to improve the product.

Conclusions? 2.2-2.3 GHz is OK, and will keep AMDers in the game against PIV C1/DCDDR, but 2.4 would have better. At 2.2-2.3GHz TBredBs will edge the PIV C1s running at a bit over 3GHz, but the PIVDCDDR will do considerably better in memory bandwidth.

Ed

Leave a Reply

Your email address will not be published. Required fields are marked *