65-watt AM2 X2's
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
65-watt AM2 X2's
I noticed that there are new 65w versions of the X2, as opposed to the 89w models. I've seen a few entries on some stores, but there isn't much info. I'm VERY interested in getting a hold of one as I'm in the process of building a system now.
Well, I for one am not convinced its been axed.Yeti wrote:Ah, I see. I'll just go with a normal 3800 now, and will upgrade to something faster and cooler down the road. The 3600 isn't anywhere to be found except in Asia, as it was axed for some reason.
-
- Posts: 2198
- Joined: Thu Feb 10, 2005 11:20 am
- Location: TN, USA
http://forums.silentpcreview.com/viewtopic.php?t=30626
Stripping the non 3800+ chips from the list you get:
X2 3800+, BH9-E4, 59°C, 1.35V, 53.1W
X2 3800+, BH9-E4, 59°C, 1.35V, 53.1W
X2-3800+, JH9-E6, 61°C, 1.35V, 59.4W
X2-3800+, JH9-E6, 61°C, 1.35V, 59.4W
X2 3800+, BH9-E4, 61°C, 1.35V, 59.4W
X2 3800+, BH9-E4, 61°C, 1.35V, 59.4W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2-3800+, JH9-E6, 63°C, 1.35V, 65.6W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2 3800+, JH9-E6, 63°C, 1.35V, 65.6W
X2-3800+, JH9-E6, 65°C, 1.35V, 71.9W
X2 3800+, BH9-E4, 67°C, 1.35V, 78.1W
X2-3800+, JH9-E6, 67°C, 1.35V, 78.1W
X2-3800+, BH9-E4, 71°C, 1.35V, 89.0W
Not a lot of parts above the 65W TDP there.
Now as to the undervolt claim, Lets compare for a second. The AM2 version is the Windsor core. The voltage for Windsor VCore is: 1.30 V - 1.35 V. and the Stepping is F. Note that the list above is chock full of E4 and E6 steppings (socket 939). So those are almost certainly Manchester cores (Manchester can be either E4 or E6 but E6 can also be Toledo).
Both Toledo and Manchester are rated at 1.35 to 1.40 which is higher than the F stepping. Yet Windsor uses more watts at the same clock speed than Toledo and Manchester. Why that is is beyond the scope of this post. It is enough to know that the stock voltage of a chip does not indicate overall power usage and you can't compare Manchesters to Windsors soley by voltage.
A Windsor Athlon 64 X2 3800+ 65W is listed as 1.20v or 1.25v.
A Windsor Athlon 64 X2 3800+ 35W is listed as 1.025v or 1.075v.
These are clearly undervolted. But are they anywhere near clearly named?
The list at the top of this post is of TDPs found in a socket 939 part that is named as a 89W part. If you already have a socket 939 board you could just buy a new socket 939 part and be reasonably assured that you would get a 65W or less TDP. Though to be sure you might have to be willing to pay a restocking fee somewhere and buy a 2nd one elsewhere as you might hit a 70 to 89W TDP Manchester occasionally.
Who knows what the actual TDPs of the so called "65W" and "35W" Windsor part you could buy are?
Stripping the non 3800+ chips from the list you get:
X2 3800+, BH9-E4, 59°C, 1.35V, 53.1W
X2 3800+, BH9-E4, 59°C, 1.35V, 53.1W
X2-3800+, JH9-E6, 61°C, 1.35V, 59.4W
X2-3800+, JH9-E6, 61°C, 1.35V, 59.4W
X2 3800+, BH9-E4, 61°C, 1.35V, 59.4W
X2 3800+, BH9-E4, 61°C, 1.35V, 59.4W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2-3800+, JH9-E6, 63°C, 1.35V, 65.6W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2 3800+, BH9-E4, 63°C, 1.35V, 65.6W
X2 3800+, JH9-E6, 63°C, 1.35V, 65.6W
X2-3800+, JH9-E6, 65°C, 1.35V, 71.9W
X2 3800+, BH9-E4, 67°C, 1.35V, 78.1W
X2-3800+, JH9-E6, 67°C, 1.35V, 78.1W
X2-3800+, BH9-E4, 71°C, 1.35V, 89.0W
Not a lot of parts above the 65W TDP there.
Now as to the undervolt claim, Lets compare for a second. The AM2 version is the Windsor core. The voltage for Windsor VCore is: 1.30 V - 1.35 V. and the Stepping is F. Note that the list above is chock full of E4 and E6 steppings (socket 939). So those are almost certainly Manchester cores (Manchester can be either E4 or E6 but E6 can also be Toledo).
Both Toledo and Manchester are rated at 1.35 to 1.40 which is higher than the F stepping. Yet Windsor uses more watts at the same clock speed than Toledo and Manchester. Why that is is beyond the scope of this post. It is enough to know that the stock voltage of a chip does not indicate overall power usage and you can't compare Manchesters to Windsors soley by voltage.
A Windsor Athlon 64 X2 3800+ 65W is listed as 1.20v or 1.25v.
A Windsor Athlon 64 X2 3800+ 35W is listed as 1.025v or 1.075v.
These are clearly undervolted. But are they anywhere near clearly named?
The list at the top of this post is of TDPs found in a socket 939 part that is named as a 89W part. If you already have a socket 939 board you could just buy a new socket 939 part and be reasonably assured that you would get a 65W or less TDP. Though to be sure you might have to be willing to pay a restocking fee somewhere and buy a 2nd one elsewhere as you might hit a 70 to 89W TDP Manchester occasionally.
Who knows what the actual TDPs of the so called "65W" and "35W" Windsor part you could buy are?
IMO the main reason for the AM2 X2s having a higher TDP is the DDR2 memory controller.
It has to deal with twice the bandwidth and to better hide latency than the one for 939.
There's also the issue of TDP binning. The 89W desktop parts are the ones that didn't qualify for 65W.
Today those 65W AM2 X2s showed up for the first time in my country. Too bad I already got my eye on a 2.4 GHz Conroe.
It has to deal with twice the bandwidth and to better hide latency than the one for 939.
There's also the issue of TDP binning. The 89W desktop parts are the ones that didn't qualify for 65W.
Today those 65W AM2 X2s showed up for the first time in my country. Too bad I already got my eye on a 2.4 GHz Conroe.