65nm X2 3600+
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
-
- Posts: 189
- Joined: Sat Jun 05, 2004 12:56 pm
- Location: Houston, Texas
- Contact:
65nm X2 3600+
I caught wind of this new, unlisted (OEM-only) AMD processor on The Tech Report's "deal of the week" feature. Available separately or in a discounted bundle with a Biostar TForce mobo, it is a 65nm, 65W EE Brisbane clocked at 1.9GHz with 512KB of L2 cache per core (unlike the original 90nm 3600+ at 2.0GHz and with 256x2 KB cache). The new fabrication process should lower power consumption even further compared to the 90nm 3800+ and 3600+, and it should be the dual-core desktop CPU to beat with respect to heat output. Not a bad deal in terms of price/performance, either, at least until the E4300 drops down to similar levels.
Also available at ChiefValue.com.
So who wants to give this proc a try? And report back?
Also available at ChiefValue.com.
So who wants to give this proc a try? And report back?
-
- Posts: 189
- Joined: Sat Jun 05, 2004 12:56 pm
- Location: Houston, Texas
- Contact:
If it can run at 1.0V on default frequency, it would probably be in the low 20W range. Pretty good.
For reference, my X2 3800+ at 2.2GHz and 1.2V seems to consume 40-45W and my E6600 at 2.4GHz and 1.0V seems to consume 25-30W.
A C2D E4300/E6300 is probably still a better alternative in every category except price.
For reference, my X2 3800+ at 2.2GHz and 1.2V seems to consume 40-45W and my E6600 at 2.4GHz and 1.0V seems to consume 25-30W.
A C2D E4300/E6300 is probably still a better alternative in every category except price.
Nope, both figures are under load. I'm estimating it based on my own power consumption measurements at the wall socket and figures measured by various review sites.jaganath wrote:Out of interest how are you calculating this? Is this at idle or load? If idle the X2 3800 number seems rather high.
The differences at the wall socket during idle and full load (using Orthos) are:
X2 3800+: 144-108 = 36W
E6600: 130-101 = 29W
Due to power supply efficiency (77-80% for the NeoHE 430W) and CPU power circuit efficiency (hardly higher than 85%), it's unlikely that more than 0.8*0.85 => 68% of that reaches the CPU. Probably not even that much.
Those figures, along with a bunch of idle numbers from around the web, is what I'm using to calculate my estimates. They're hardly exact, but should provide a pretty good estimate.
I'm very impressed by the E6600 at 1.0V. It uses about the same amount of power as the 3800+ EE SFF and is approximately 50% faster.
-
- *Lifetime Patron*
- Posts: 4284
- Joined: Fri Apr 04, 2003 6:21 pm
- Location: Undisclosed but sober in US
Re: 65nm X2 3600+
Crap, first the 3600+ has an odd divider, second the newegg package includes the 550 mobo as opposed to the T6100 which would also have video.tempeteduson wrote:I caught wind of this new, unlisted (OEM-only) AMD processor on The Tech Report's "deal of the week" feature. Available separately or in a discounted bundle with a Biostar TForce mobo, it is a 65nm, 65W EE Brisbane clocked at 1.9GHz with 512KB of L2 cache per core (unlike the original 90nm 3600+ at 2.0GHz and with 256x2 KB cache). The new fabrication process should lower power consumption even further compared to the 90nm 3800+ and 3600+, and it should be the dual-core desktop CPU to beat with respect to heat output. Not a bad deal in terms of price/performance, either, at least until the E4300 drops down to similar levels.
Also available at ChiefValue.com.
So who wants to give this proc a try? And report back?
-
- Posts: 1608
- Joined: Tue Jan 04, 2005 4:02 pm
- Location: United States
-
- Posts: 1608
- Joined: Tue Jan 04, 2005 4:02 pm
- Location: United States
-
- Posts: 189
- Joined: Sat Jun 05, 2004 12:56 pm
- Location: Houston, Texas
- Contact:
-
- Posts: 89
- Joined: Tue Sep 07, 2004 8:24 am
- Location: MA, USA
No.1, those are not exact measurements, No.2 the AMDs use less than C2D at idle when everything is left at stock (CnQ/EIST enabled), Mikael's measurements are undervolted.jimmyfergus wrote:But isn't one of the big things about the AMD, that it uses so much less power at idle than the Intel?Mikael wrote:X2 3800+: 144-108 = 36W
E6600: 130-101 = 29W
http://www.xbitlabs.com/articles/cpu/di ... ent_6.html
However the new E4XXX CPUs from Intel may eradicate AMDs lead in this area.
-
- *Lifetime Patron*
- Posts: 4284
- Joined: Fri Apr 04, 2003 6:21 pm
- Location: Undisclosed but sober in US
Well if you got the NewEgg combo deal then you must have been one of the last ones because that offer is no longer available.frostedflakes wrote:Well, I bought one. Should be able to set everything up later this week, will let you guys know how it does.
I asked them to combo it with the T6100. No such luck.
-
- Posts: 1608
- Joined: Tue Jan 04, 2005 4:02 pm
- Location: United States
-
- Posts: 1608
- Joined: Tue Jan 04, 2005 4:02 pm
- Location: United States
-
- Posts: 189
- Joined: Sat Jun 05, 2004 12:56 pm
- Location: Houston, Texas
- Contact:
-
- Posts: 1608
- Joined: Tue Jan 04, 2005 4:02 pm
- Location: United States
UPDATE: Car problems turned out to be nothing serious, so I redid my order. X2 3600+ HERE I COME!!!
I have been using a single-core laptop, but since getting broadband I've started to do a lot more multimedia stuff. This dual core should really help out with that. Will let you all know how it undervolts. I also have a wattmeter so I can measure power consumption.
I have been using a single-core laptop, but since getting broadband I've started to do a lot more multimedia stuff. This dual core should really help out with that. Will let you all know how it undervolts. I also have a wattmeter so I can measure power consumption.
-
- *Lifetime Patron*
- Posts: 618
- Joined: Mon May 23, 2005 7:05 am
- Location: State College, PA
-
- Posts: 189
- Joined: Sat Jun 05, 2004 12:56 pm
- Location: Houston, Texas
- Contact:
AMDZone's first look at the 3600+ Brisbane, for what it's worth. They basically tell you they could overclock the chip to 3.1GHz with a core voltage boost. That's it.
New Egg is selling this OEM at under $120.
I'd think that how Cool/Quiet this runs would be a BIG thing here. Undervolt a bit-stick a Ninja on it...you could run a low restriction case with a lone low speed 120 mm case fan + whatever the PSU contributes.
Yeah,Gamers may care about overclocking,I'd see this as a competitor to the Semprons as the heart of a SILENT system for pretty low cost.
Yeah,could go Intel,but I like the options for AMD mobos and like too that the whole package adds up to less $.
We talk about Watts out of the wall,the claimed worst case watts for the Brisbane series-what I WANT to see is how FEW RPM's I could get with this,A Core 2,a Venice or Semp....in other words-whats the cool-quiet factor.
I'd think that how Cool/Quiet this runs would be a BIG thing here. Undervolt a bit-stick a Ninja on it...you could run a low restriction case with a lone low speed 120 mm case fan + whatever the PSU contributes.
Yeah,Gamers may care about overclocking,I'd see this as a competitor to the Semprons as the heart of a SILENT system for pretty low cost.
Yeah,could go Intel,but I like the options for AMD mobos and like too that the whole package adds up to less $.
We talk about Watts out of the wall,the claimed worst case watts for the Brisbane series-what I WANT to see is how FEW RPM's I could get with this,A Core 2,a Venice or Semp....in other words-whats the cool-quiet factor.
-
- *Lifetime Patron*
- Posts: 4284
- Joined: Fri Apr 04, 2003 6:21 pm
- Location: Undisclosed but sober in US
Before I decided against an AMD (for efficiency and bang-for-the-buch reasons) I saw this as a super cooler for an OEM chip.
http://www.newegg.com/Product/Product.a ... 6835200014
A nice sold chuck of copper on the bottom and it takes a 92mm fan, for $13. At this price odds are good you'll have to chuck the fan. Has a generic speed control as well.
I'd put a 120mm fan on it with those plastic ties and cool the board as well.
http://www.newegg.com/Product/Product.a ... 6835200014
A nice sold chuck of copper on the bottom and it takes a 92mm fan, for $13. At this price odds are good you'll have to chuck the fan. Has a generic speed control as well.
I'd put a 120mm fan on it with those plastic ties and cool the board as well.
-
- Posts: 189
- Joined: Sat Jun 05, 2004 12:56 pm
- Location: Houston, Texas
- Contact:
I saw this article yesterday but didn't have a chance to read it until today: Josh Walrath's testing of the new 3600+ at Penstar Systems. Again, no thermal or power consumption data; just an overclocking endeavor. He got close to 2.8Ghz but cited the motherboard as the limiting factor. The chip scales nicely with clock speed, so no surprises there.
-
- Posts: 1608
- Joined: Tue Jan 04, 2005 4:02 pm
- Location: United States
All the parts showed up today. After getting everything put together and spending a few hours pulling my hair out, I finally got it working. Had to disable CnQ on my board to get it to complete POST. The motherboard shipped with an old BIOS, and I suspect Brisbane's odd half multipliers just didn't play well with CnQ. After updating to a more recent BIOS with Brisbane support all seems to be good w/CnQ enabled. Just a little tip for those who might be trying to put together a similar setup.
Am getting ready to download RMClock and start testing undervolting. Will be sure to post the results and power tests coming up soon.
Am getting ready to download RMClock and start testing undervolting. Will be sure to post the results and power tests coming up soon.