http://www.hexus.net/content/item.php?item=25269

I'm sure there are SPCR readers that are looking for "high performance" silent workstations, like myself. My Q6600 just isn't cutting it any more, i7s are too expensive, and previously AMD's 6-core champ had a rather higher 125W TDP. So now they're releasing this 45nm edition, maybe this will be much easier to cool, and lighter on the electricity bill...

That is good news that they have the X6 at 95W, now where is my $100 X4 960T?

TDP 95W really power is between 95 and 125W depending voltage

It could be 125W at 1.35V, 120W at 1.325V, 1.30V and 115W and still be 95W TDP

Idle power is still the same.

I agree the idle power will be the same... but not with your first comments. TDP is about the maximum power at stock speeds/settings with the highest load a CPU can see. AMD specs ( http://products.amd.com/en-us/DesktopCP ... 641&id=652 ) clearly indicates that the new chip runs 1.075-1.375 vs 1.125-1.40V for the old -- ie, it requires less voltage to run. It also has a higher max safe temp... which is a little curious.

This could mean that 125TDP version and 95 running at the same voltage 1.375 are 125 and 95 real wattages, that would be great, or maybe 95TDP uses the same crystal just undervolted to 1.225V, becasause there is just no other way doing this without making a radical improvement, but i am no expert. If is the first case, it would be good to know:)

Max temp increase could mean that leakage has been minimised by further straining the strainded silicon, but likely they just came up with that number, or is it something else?

Radical improvements are common in the first year of a new process or design. This isn't even that radical. A 65W X6 would be radical.

IMO the reduced TDP could be a combination of a new stepping and reduced auto-overclocking.
Still, seems good value for money if one can use all six cores. And with a good heatsink it should run cool with fans speeds under 800 rpm.

Same stepping E0, just lower VID that is 1.10 or 1.15 to achieve lower watts. it overtakes 4.0Ghz low voltage too. no wonder they charge more for it.

This was an expected design but finally! Now I can jump on the X6 bandwagon!

Someone at xtreme systems got their hands on a 95watt 1055t [url]http://www.xtremesystems.org/forums/showthread.php?t=254765]Link[/url]

Highlights:
-135watt system power usage in Prime95, 76watt idle (CnQ) with IGP.
-1.15v stock voltage in windows (without CnQ)
-3.7ghz @ 1.236v
-4ghz @ 1.284
-4.2ghz @ 1.44v


The clock at 1.44v is a little disappointing considering the 4ghz volt. Still a pretty amazing chip. Hexcore 4ghz chip in a mini-itx anyone?

45nm was 2 years ago. Not suitable for silent PCs.

Yeah, I noticed that when this topic was first posted. All AMD CPUs since the AII/PII are 45nm. The 125W -> 95W version is not from a change in manufacturing process but just a tweak to the existing process... Maybe even just a reduction in voltage. AMD seems to use unnecessarily high voltages on their CPUs as most can be undervolted by about 0.2V comfortably if not more.

Manufacturing process (90nm, 65nm, 45nm) is not directly related to having a silent PC. Intel's first 65nm CPUs were hotter than their 90nm CPUs at the time due to leakage and stuff. Usually a smaller transistor size means it needs less power and can run cooler. But as long as your cooling is good enough to handle the heat, you can still have a silent PC with a hot CPU. It's just a little more difficult.

I totally agree. These chips are not suitable for 'silent' PCs, but they can certainly be used in quiet ones.

My current build has a 125w 1055t overclocked to 3.5GHz on stock voltage. The CPU is undervolted to 0.75v 800MHz at idle. CPU and case fans are about 500rpm at idle and 1200/1000 rpm at load running LinX. It's certainly audible at full load, but not loud by my standard. I also tested undervolting at stock speeds, and I could go down to about 1.1v.

I wouldn't be surprised if these 95watt 1055t can be undervolted to below 1v or even 0.95v. Then they would be in the order of 80 watt at full load. That's certainly manageable in a quiet build.

Nice link and post. Thanks

Any word on these in the states yet?

With which mobo/software you do that (undervolting the CPU only at idle P-state)?

What can I say, you've got one hell of a chip. Anybody else managed to undervolt a Thuban this low at idle ???