Dont understand...hd4850 10w idle @ 68c

They make noise, too.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Thu Jun 19, 2008 1:17 pm

ryboto wrote:
drjunk wrote: Jeez....Even the gtx280 burns less power when idle...
tis a bit odd.
It’s built on the same 55nm process as the HD 38x0 series and is a lot more complex so it’s not surprising that it consumes more power. But weren’t AMD talking about 10W at idle for the HD 48x0 series! Either they screwed up or the current drivers aren’t supporting the down-clocking at idle yet; fingers crossed that it’s the latter.

ryboto
Friend of SPCR
Posts: 1439
Joined: Tue Dec 14, 2004 4:06 pm
Location: New Hampshire, US
Contact:

Post by ryboto » Thu Jun 19, 2008 3:33 pm

smilingcrow wrote:
ryboto wrote:
drjunk wrote: Jeez....Even the gtx280 burns less power when idle...
tis a bit odd.
It’s built on the same 55nm process as the HD 38x0 series and is a lot more complex so it’s not surprising that it consumes more power. But weren’t AMD talking about 10W at idle for the HD 48x0 series! Either they screwed up or the current drivers aren’t supporting the down-clocking at idle yet; fingers crossed that it’s the latter.
That's what I was referring to when I called it odd, it should be throttling down a bit lower than what the reviews are reporting.

chahahc
Posts: 75
Joined: Tue Oct 23, 2007 7:52 pm
Location: The Most Mexican Place in the U.S.A

Post by chahahc » Thu Jun 19, 2008 8:04 pm

Well here's a review of the 4850 by Anandtech showing around 27 watts higher than the 3870 at load, but with a large increase in performance IMO. It seems like ATI has a pretty decent card on their hands, stepping up the competition, especially in the prices...I love capitalism. :mrgreen:
http://www.anandtech.com/video/showdoc.aspx?i=3338

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Fri Jun 20, 2008 12:44 am

ryboto wrote:That's what I was referring to when I called it odd, it should be throttling down a bit lower than what the reviews are reporting.
What were they reporting the idle clock speeds to be? The voltage also needs dropping for a significant reduction in power to be achieved I assume!
chahahc wrote:Well here's a review of the 4850 by Anandtech showing around 27 watts higher than the 3870 at load, but with a large increase in performance IMO. It seems like ATI has a pretty decent card on their hands, stepping up the competition, especially in the prices.
Performance is good for the price and the wattage so it’s only the idle power draw that disappoints.
nVIDIA has announced the 9800 GTX+ which is a 55nm GPU so will likely compete closely with the HD 4850 in the performance per watt stakes. The idle power draw of the GTX 200 series is surprisingly low so it’ll be interesting to see if they manage to make the 9800 GTX+ more competitive in that department.

Modo
Posts: 486
Joined: Wed Apr 16, 2008 3:32 am
Location: Poland

Post by Modo » Fri Jun 20, 2008 1:58 am

Well, the 9800GTX+ will cost more. At the same price point, the HD 4850 looks better than the (now cheaper, guess why) 9800GTX. It's still a crap performer when power consumption is considered, but hey, let's wait for the lower-echelon cards. :)

FartingBob
Patron of SPCR
Posts: 744
Joined: Tue Mar 04, 2008 4:05 am
Location: London
Contact:

Post by FartingBob » Fri Jun 20, 2008 2:13 am

I wonder/hope. Will the 4850 have mounting holes for my Accelero? Or is this just a day dream? This cards power consumption makes it a possibility for passive, or low speed fan at most.
If not, i hope AC get around to making a new version!

rpsgc
Friend of SPCR
Posts: 1630
Joined: Tue Oct 05, 2004 1:59 am
Location: Portugal

Post by rpsgc » Fri Jun 20, 2008 2:19 am

FartingBob wrote:I wonder/hope. Will the 4850 have mounting holes for my Accelero? Or is this just a day dream? This cards power consumption makes it a possibility for passive, or low speed fan at most.
If not, i hope AC get around to making a new version!
Mounting holes are the same as the HD38x0 series.

FartingBob
Patron of SPCR
Posts: 744
Joined: Tue Mar 04, 2008 4:05 am
Location: London
Contact:

Post by FartingBob » Fri Jun 20, 2008 3:10 am

rpsgc wrote:
FartingBob wrote:I wonder/hope. Will the 4850 have mounting holes for my Accelero? Or is this just a day dream? This cards power consumption makes it a possibility for passive, or low speed fan at most.
If not, i hope AC get around to making a new version!
Mounting holes are the same as the HD38x0 series.
Fantastic, well thats pretty much sealed the deal for me then!

mczak
Posts: 147
Joined: Wed Jul 21, 2004 6:13 pm

Post by mczak » Fri Jun 20, 2008 4:26 am

smilingcrow wrote:What were they reporting the idle clock speeds to be? The voltage also needs dropping for a significant reduction in power to be achieved I assume!
The current drivers do not seem to lower neither voltage nor clock for the HD4850 by a really significant amount: http://www.techpowerup.com/reviews/Powe ... 50/21.html
HD3850/HD3870 had both lower idle voltage and frequency (and higher load voltage with higher load frequency).
Seems a bit odd that the "low" and "medium" power settings would be the same, so I guess there's some room for improvement. No idea though if this will actually materialize...

cd8uk
Patron of SPCR
Posts: 47
Joined: Wed Apr 18, 2007 12:55 pm

Post by cd8uk » Fri Jun 20, 2008 4:56 am

[b]rpsgc[/b] wrote:
[b]FartingBob[/b] wrote: I wonder/hope. Will the 4850 have mounting holes for my Accelero? Or is this just a day dream? This cards power consumption makes it a possibility for passive, or low speed fan at most.
If not, i hope AC get around to making a new version!
Mounting holes are the same as the HD38x0 series.
So, does that mean that a ZM-GWB3850/3870 will fit on a reference 4850??? If so that'd be perfect for my Reserator 2! I just don't want to take the risk of buying the card and cooler if it's not going to work.
Google hasn't been much help so does anyone know if the 3850 and 4850 layouts are effectively identical?

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Fri Jun 20, 2008 7:40 am

Matija wrote:Or not, if this preview is correct:

http://www.hardware.fr/articles/724-2/p ... -4850.html

Image

But who knows, maybe it was a different system...
how the hell is this french moron getting that much wattage draw in general???? his 3870 system shouldnt be breaking 280 watts.

eh?

thejamppa
Posts: 3142
Joined: Mon Feb 26, 2007 9:20 am
Location: Missing in Finnish wilderness, howling to moon with wolf brethren and walking with brother bears
Contact:

Post by thejamppa » Fri Jun 20, 2008 8:52 am

Maybe his PSU is bad or has really, really bad effeciency...

Scoop
Posts: 121
Joined: Sat Dec 08, 2007 1:28 am

Post by Scoop » Fri Jun 20, 2008 8:53 am

loimlo wrote:3850's reference single slot cooler had the similar situation, 65'c, at idle. It looks like ATi implement the same temperature-controlling mechanism on 4850 as well. Anyway, I wonder whether 4850 enjoy the same quiet cooler design as 3850 that has been verified by silent authority, SPCR.
Don't know where you got that but my vanilla manually OC'd 3850 idles at 52C and the fan spins at the minimum.

I'm confused like many others but of the fact that the 4850 manages to draw almost as much as a 9800GTX under load. 9800GTX has TWO pcie connectors and the 4850 only one. Something is definately off here. Unless the other pcie connector on the 9800GTX is useless. I don't think AMD would risk it on one connector if the draw could peak higher than it's capable of.

rpsgc
Friend of SPCR
Posts: 1630
Joined: Tue Oct 05, 2004 1:59 am
Location: Portugal

Post by rpsgc » Fri Jun 20, 2008 9:08 am

Or maybe the HD4850 isn't downclocking/undervolting like it should?

drjunk
Posts: 51
Joined: Tue Apr 25, 2006 8:46 am
Location: Scotland

Post by drjunk » Fri Jun 20, 2008 10:17 am

Is the 40nm shrink still on for the end of the year? I think im going to hold on till then and get a 40nm version.

ryboto
Friend of SPCR
Posts: 1439
Joined: Tue Dec 14, 2004 4:06 pm
Location: New Hampshire, US
Contact:

Post by ryboto » Fri Jun 20, 2008 10:40 am

drjunk wrote:Is the 40nm shrink still on for the end of the year? I think im going to hold on till then and get a 40nm version.
It's probably going to be 45nm....right?

Shadout
Posts: 117
Joined: Tue Apr 01, 2008 6:04 pm
Location: Denmark

Post by Shadout » Fri Jun 20, 2008 2:17 pm

drjunk wrote:Is the 40nm shrink still on for the end of the year? I think im going to hold on till then and get a 40nm version.
still?
Have they ever mentioned a 45 (or 40) nm shrink?

ryboto
Friend of SPCR
Posts: 1439
Joined: Tue Dec 14, 2004 4:06 pm
Location: New Hampshire, US
Contact:

Post by ryboto » Fri Jun 20, 2008 2:43 pm

Shadout wrote:
drjunk wrote:Is the 40nm shrink still on for the end of the year? I think im going to hold on till then and get a 40nm version.
still?
Have they ever mentioned a 45 (or 40) nm shrink?
I remember reading news of TSMC finalizing 45nm technology....or that they do have the capacity to do it, but it will probably take some time to validate it for a complex GPU.

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Fri Jun 20, 2008 4:15 pm

cd8uk wrote:Google hasn't been much help so does anyone know if the 3850 and 4850 layouts are effectively identical?
Everything I’ve read points to them being extremely similar although in one photo one of the RAM chips looked as if it was spaced slightly wider than on my HD 3850. Not sure if that would be a problem anyway.

Scoop wrote:I'm confused like many others but of the fact that the 4850 manages to draw almost as much as a 9800GTX under load. 9800GTX has TWO pcie connectors and the 4850 only one. Something is definately off here. Unless the other pcie connector on the 9800GTX is useless. I don't think AMD would risk it on one connector if the draw could peak higher than it's capable of.
Even if the 9800 GTX is only 10W over what a single 6 pin connector + PCIe slot can provide it would still need the second connector (or possibly a single 8 pin connector) so you can’t use that as an accurate yardstick. The second connector also gives headroom for over-clocking.

Shadout
Posts: 117
Joined: Tue Apr 01, 2008 6:04 pm
Location: Denmark

Post by Shadout » Sat Jun 21, 2008 3:54 am

ryboto wrote: I remember reading news of TSMC finalizing 45nm technology....or that they do have the capacity to do it, but it will probably take some time to validate it for a complex GPU.
Ye, they do, but I think they have said it wouldn't be before 2009.
Maybe they surprise us, but tbh, I believe those waiting for die shrinks might end up waiting for quite a while.

ryboto
Friend of SPCR
Posts: 1439
Joined: Tue Dec 14, 2004 4:06 pm
Location: New Hampshire, US
Contact:

Post by ryboto » Sat Jun 21, 2008 7:56 am

Shadout wrote: Ye, they do, but I think they have said it wouldn't be before 2009.
Maybe they surprise us, but tbh, I believe those waiting for die shrinks might end up waiting for quite a while.
There's almost always a video card refresh around Dec/Jan. I am going to bet that's when the process switch will occur. Either then, or maybe Oct/Nov. Depends on when Nvidia decides to attempt the 280 on 55nm. Maybe Ati would react and do the same, only with 45nm? They'll do it, and they'll do it quick if the yields are better and cost is lower.

nutball
*Lifetime Patron*
Posts: 1304
Joined: Thu Apr 10, 2003 7:16 am
Location: en.gb.uk

Post by nutball » Sat Jun 21, 2008 8:57 am

ryboto wrote:There's almost always a video card refresh around Dec/Jan. I am going to bet that's when the process switch will occur. Either then, or maybe Oct/Nov. Depends on when Nvidia decides to attempt the 280 on 55nm. Maybe Ati would react and do the same, only with 45nm? They'll do it, and they'll do it quick if the yields are better and cost is lower.
Thing is that NVIDIA desperately need to shrink GT200 from stupidly large to plain large, whereas ATI don't need to shrink RV770 given how small it is already.

In any case they're playing in different markets - the NVIDIA shrink of interest wrt RV770 is probably G92b (AKA 9800GTX+) which has already happened ... in the sense that it exists even if you can't quite buy it yet.

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Sat Jun 21, 2008 9:31 am

The problem with the 9800 GTX+ is that it doesn’t have the power saving features of the GTX 200 series which means its idle power consumption is higher than that of the HD 4850. At least that’s what the early reviews are showing although it’s possible that new drivers might improve the situation for both platforms.

It’s good to see nVIDIA finally addressing their high idle power consumption with the new GTX 200 series. With a board that supports Hybrid Power you could feasibly run an SLI GTX 280 system with a low power draw at idle. Good job as the power draw at load is still fairly monstrous.

nutball
*Lifetime Patron*
Posts: 1304
Joined: Thu Apr 10, 2003 7:16 am
Location: en.gb.uk

Post by nutball » Sat Jun 21, 2008 12:47 pm

9800GTX+ should support Hybridpower, shouldn't it? 9800GTX did so what's different? Drivers? Were the tests on HP-capable motherboards?

TBH I think that for both IHVs performance is 98% and idle power 2% of their attention span, and with the recent releases the caca has hit the fan on the performance front so it wouldn't surprise me at all if these vaunted power-saving features don't work quite right yet.

smilingcrow
*Lifetime Patron*
Posts: 1809
Joined: Sat Apr 24, 2004 1:45 am
Location: At Home

Post by smilingcrow » Sat Jun 21, 2008 1:25 pm

nutball wrote:9800GTX+ should support Hybridpower, shouldn't it? 9800GTX did so what's different? Drivers? Were the tests on HP-capable motherboards?
It’s very likely that they tested with an Intel platform and I’m not sure if Hybrid Power motherboards have been released for Intel systems yet; they haven’t for non SLI systems anyway.
But regardless you still need to know the normal power draw as most people don’t have a motherboard that supports Hybrid Power.

johnniecache7
Posts: 332
Joined: Wed Jan 17, 2007 8:10 pm
Location: Toronto - Ontario - Canada

Post by johnniecache7 » Sat Jun 21, 2008 10:41 pm

:o Problem with the HD 4850 is it runs way to hot loads in the 90s. I doubt anything cool it passively but Thermalright's T-Rad with slient 120mm fan will cool it silently I bet. I wouldn't count on any the older coolers like the S1 doing the job very well.

rpsgc
Friend of SPCR
Posts: 1630
Joined: Tue Oct 05, 2004 1:59 am
Location: Portugal

Post by rpsgc » Sun Jun 22, 2008 12:25 am

johnniecache7 wrote::o Problem with the HD 4850 is it runs way to hot loads in the 90s. I doubt anything cool it passively but Thermalright's T-Rad with slient 120mm fan will cool it silently I bet. I wouldn't count on any the older coolers like the S1 doing the job very well.
Replace stock thermal goo. There.
And what exactly has the stock cooler's temperature to do with the cooling performance of the S1?

mkk
Posts: 687
Joined: Sun Sep 05, 2004 1:51 pm
Location: Gefle, Sweden
Contact:

Post by mkk » Sun Jun 22, 2008 12:26 am

There's a "hotfix" driver version 8.6 out for the 48xx range, in case you haven't tried it already.
http://support.ati.com/ics/support/defa ... onID=35298

Tephras
Posts: 1140
Joined: Tue Sep 07, 2004 11:03 am
Location: Europe

Post by Tephras » Sun Jun 22, 2008 1:01 am

johnniecache7 wrote:I wouldn't count on any the older coolers like the S1 doing the job very well.
4850 stock cooler VS Accelero S1

Sendorm
Posts: 56
Joined: Wed Jun 08, 2005 10:21 pm

Post by Sendorm » Sun Jun 22, 2008 1:21 am

Tephras, that was really funny and to the point, thanks.

Can you get a reading of the wattage?

Post Reply