   |
|||
|
|||
|
:
News
: Archives
|
|
AMD Athlon X2 6000+ Review |
|
|
No FX for You! |
|||
|
|
|||
|
by Josh Walrath |
|||
|
A Question of Power In the Pentium 4 days AMD was sitting pretty. Their Athlon 64 parts based on their 90 nm process were the most power efficient processors on the planet, and they ran cool too. Times have obviously changed. Intel released the Core 2 Duo on their 65 nm process, and suddenly the power to performance ratio radically switched sides. Now AMD is looking more like Intel during the Prescott days. The 6000+ certainly gives the impression that it is a firebreathing beast. There is more to this story though. The 6000+ is the highest wattage processor that AMD has released for the AM2 socket. But the numbers are somewhat more interesting than what one first expects. The 6000+ has the same power rating as the old FX-62, even though it is a full 200 MHz faster. 125 watts TDP is the max amount of power, in the absolute worst case scenario that this chip will draw. What is even more interesting is that the X2 5600, which is a direct replacement for the FX-62, now has an 89 watt TDP. When we dig a little farther down the line we see that AMD has 65 watt TDP EE products starting at 2.5 GHz and going down to 1.9 GHz. These EE products are a mix of 90 nm Windsor and 65 nm Brisbane cores. At the very bottom end with the Brisbane based 1.9 GHz X2 3600, it seems the 65 watt TDP is vastly overstated! Considering that 3rd party tests have shown that the X2 4800 Brisbane is pulling at max around 40 watts, the 1.9 GHz product could easily be ¾ that number. All 65 nm parts are considered Energy Efficient. A question that comes to mind when looking at this situation is why AMD has not put the screws to the 65 nm process and released a 3.0 GHz product based on that? The answer to that is not exactly easy, and it mainly deals with the economics of node migration. From all indications AMD has a healthy 65 nm process, but it is not yet optimized for high speeds. AMD also wants to pay off their investment in their 65 nm process by selling the greatest amount of chips possible based on that node. To do that they need to set conservative speed bins. AMD has focused their 65 nm process on providing parts for the “meat and potatoes” levels of the market. This means dual core processors that cost between $100 and $250. Another factor is that AMD wants as many good dies to come off the line as possible. If they were merely focused on the highest transistor performance products, then yields and speed bins will be poor. AMD could do this by adjusting their process properties, but in a brand new process the potential for disastrous yields is huge. It makes little business sense for AMD to try this, as sales of 10,000 X2 6000+ at $450 are easily overshadowed by 100,000 lower clocked processors between $100 and $200. It simply is not an effective usage of resources, especially resources that cost as much as the 65 nm line took to develop. AMD simply does not have the resources that Intel does, and considering they have one Fab that has a working 65 nm line, it makes much more sense to produce the greatest amount of marketable products than to chase after the illusory performance crown. These are the reasons we see AMD using the older 90 nm process for these last-gasp high end parts. The constant improvement at 90 nm has allowed good yields at that speed bin, and it still allows AMD to more adequately tune their 65 nm line without having to worry about top end transistor performance. Today also marks the day when AMD is releasing their first single core 65 nm Athlon 64s. The 3500+ and 3800+ are clocked at 2.2 GHz and 2.4 GHz respectively. They will be geared towards the lower end of the spectrum in terms of price, coming in at $91 and $101 officially. These single core products are meant to be good performing parts at budget prices. They also are rated at a maximum of 45 watts TDP. While these might not be for everyone, a 2.2 and 2.4 GHz single core Athlon 64 are still nicely performing parts. These could easily find themselves in budget systems as well as HTPCs.
Chip Specs The X2 6000+ is based on the Rev. F core that was developed for use in the AM2 socket with DDR-2. 3.0 GHz Dual Core 64 KB data + 64 KB instruction L1 cache per core 1 MB L2 cache per core 125 Watt TDP Max 940 pins (AM2) DDR-2 800 max memory speed 32 bit/64 bit support 1 GHz HyperTransport Bus
If you have found this article interesting or a great help, please donate to this site.
Copyright 1999-2007 PenStar Systems, LLC. |
|||
