The Fermi Thread - Part 3

**spursindonesia** · 03-23-2010, 03:09 AM

Hmm, that 130 w more, that's the total system consumptions comparison, right ? not just the comparison beetween the power consumption of the cards themselves. The system with Fermi in it might work a bit harder so not all of the power difference goes to the card itself. And it still has to account the efficiency of the PSU.

But i think, and it seems, ATi is more conservative with its TDPs on HD 58xx series compared to nVidia Fermi's TDPs.

**mindfury** · 03-23-2010, 03:40 AM

Originally Posted by neliz

Right.

there's an explanation to all the wild power usage / SP numbers.

It seems the 512CC will be reserved for the B1 part, which is currently slated for Q3.

295W is the power consumption of the part with 512CC and 725+Mhz Core (not going to see these for a while, I guess not everyone liked them.)
275W is the power consumption of the A3 part, with 480CC and 725+Mhz Core and 1050Mem i.e. OC GTX480 models
250W is the power consumption of the A3 part, with 480CC and 700Mhz Core and ~950Mem, i.e. GTX480.

So well.. everyone was right.

I also heard that power consumption is still an issue, especially idle power usage. Tomorrow should bring a new driver release as NV engineers are trying to get it down as much as possible.

http://forum.beyond3d.com/showpost.p...postcount=4500

**Florinmocanu** · 03-23-2010, 03:48 AM

Originally Posted by spursindonesia

Hmm, that 130 w more, that's the total system consumptions comparison, right ? not just the comparison beetween the power consumption of the cards themselves. The system with Fermi in it might work a bit harder so not all of the power difference goes to the card itself. And it still has to account the efficiency of the PSU.

But i think, and it seems, ATi is more conservative with its TDPs on HD 58xx series compared to nVidia Fermi's TDPs.

well, if you use the same test-setup and the same benchmark and you get 130W more power consumption for the Fermi run vs the 5870 run, than it's pretty conclusive that Fermi consumed 130W more during that test.

**mindfury** · 03-23-2010, 03:57 AM

Originally Posted by Florinmocanu

well, if you use the same test-setup and the same benchmark and you get 130W more power consumption for the Fermi run vs the 5870 run, than it's pretty conclusive that Fermi consumed 130W more during that test.

It is lower than 130W because of PSU efficiency.nApoleon's PSU is PC Power & Cooling Turbo Cool 1KW-SR 1000W,efficiency=83%,so GTX480 should consume
about 100W more than HD5870.

**zalbard** · 03-23-2010, 04:04 AM

Originally Posted by Chickenfeed

If it is true at anything less than 1920x1200 8xAA very high settings, I'll be impressed. Otherwise

Exactly.
Hope this is true, though, would be worth the wait and increased power consumption at least.

Originally Posted by Chumbucket843

since when did all of the 200 series reach eol?

A few months ago or so...

Originally Posted by alcachofa

Definitely, but there is an exception with NV cards : 3D Vision.

ATI has this as well.

Originally Posted by mindfury

GTX480 machine power consumption at full load is nearly 130W more than HD5870 machine，full load temperature is 92℃.

I don't care much about extra 130W, but another melting card after my 8800 Ultra... No, thanks.

Originally Posted by RPGWiZaRD

"Delta electronics inc" says it all. Performance before silence.

This fan has to be very loud at 100% I bet (knowing Delta this will probably run up to like 5500rpm if it is 70/72mm or whatever), maybe closed headphones starts becoming a recommendation soon.

Luckily Delta fans usually scale well noise wise if you lower speeds and they have a tendency to last like forever.

5870's fan @ 100% is crazy, also, though.

**Florinmocanu** · 03-23-2010, 04:15 AM

Originally Posted by mindfury

It is lower than 130W because of PSU efficiency.nApoleon's PSU is PC Power & Cooling Turbo Cool 1KW-SR 1000W,efficiency=83%,so GTX480 should consume
about 100W more than HD5870.

but the 5870 results also have the same 17% off in efficiency so the difference is not 100W, it's a bit more.

Look at it this way.

Let's give some theoretical numbers.

5870 = 400W total power consumption x 83% = 332 real power consumption.
GTX 480 = 530W total power consumption x 83% = 439.9 real power consumption.

difference would be 107.9 W, with these theoretical numbers.

Anandtech in their 5870 review said a 5870 consumed 401W in full load, so my numbers are pretty ok i guess
http://www.anandtech.com/video/showdoc.aspx?i=3643&p=26

But, we do not know if that guy took into account this lack of efficiency from the PSU, when stating that GTX 480 consumes 130W more than the 5870.

We need to wait 3 more days to really find out what is going on.

**mindfury** · 03-23-2010, 04:41 AM

我9241分
nApoleon 发表于 2010-3-23 19:00

GTX480 Vantage X=9241

**Andrea deluxe** · 03-23-2010, 04:56 AM

Gtx 480

core: 700mhz shader:1400mhz memory:3696mhz

gtx 470

core: 607mhz shader:1214mhz memory:3048mhz

100% FINAL FREQ.

**jaredpace** · 03-23-2010, 04:58 AM

Originally Posted by mindfury

ouch if true

Mk · 03-23-2010, 05:00 AM

Originally Posted by mindfury

GTX480 Vantage X=9241

fail if true

**annihilat0r** · 03-23-2010, 05:03 AM

I don't trust Chiphell for some reason (that I don't remember) but.... well.....

**Arxagelos** · 03-23-2010, 05:05 AM

yeah right!!! 5870 has min fps 20 in crysis!!!

2fps at least...

**annihilat0r** · 03-23-2010, 05:09 AM

Is 5870's and 5970's min FPS really the same in Crysis? Probably some "loading" situation...

**trinibwoy** · 03-23-2010, 05:19 AM

Ouch indeed. Especially if they needed to bump clocks and power consumption so high just to reach parity.

**jaredpace** · 03-23-2010, 05:20 AM

Originally Posted by mindfury

kinda matches phk's gtx470 benchmarks:

-------- min / avg fps
gtx470: 18.0 / 24.0 fps
hd5870: 20.5 / 30.5 fps

gtx470: 18.6 / 25.5 fps
hd5870: 19.6 / 30.2 fps

**Blacky** · 03-23-2010, 05:31 AM

each time I browse this thread makes me more confident that Fermi is the new R600...

**GoldenTiger** · 03-23-2010, 05:35 AM

Originally Posted by Blacky

each time I browse this thread makes me more confident that Fermi is the new R600...

That confidence makes you feel good, why? Less viable competition = worse for everyone here.

**eric66** · 03-23-2010, 05:36 AM

Originally Posted by annihilat0r

Is 5870's and 5970's min FPS really the same in Crysis? Probably some "loading" situation...

nah tahts problem of dual gpu cards they mostly have min of single gpu

if this is true that explains 10 percent difference between 470 and 5870 in previous bench seems they used 2560 res

this one is more realistic i think

**Blacky** · 03-23-2010, 05:38 AM

Originally Posted by GoldenTiger

That confidence makes you feel good, why? Less viable competition = worse for everyone here.

I didn't say makes me feel good, its just sad, the hope I had on nvidia within the release of this new gen is currently at zero...I guess ATi will not drop prices after fermi launch seems

**eric66** · 03-23-2010, 05:41 AM

Originally Posted by mindfury

http://forum.beyond3d.com/showpost.p...postcount=4500

thats more like joke which 480 we are gonna see on stores ? the one with 250 or the one that has 275 power cons.

**GoldenTiger** · 03-23-2010, 05:45 AM

Originally Posted by Blacky

I didn't say makes me feel good, its just sad, the hope I had on nvidia within the release of this new gen is currently at zero...I guess ATi will not drop prices after fermi launch seems

My bad, it sounded like it made you glad

. Yeah, I still am holding a little hope but not a lot of confidence but for the B3D quote from Felix:

OK, some first hand info, but my source is still under NDA, so not much in-depth details.

GTX470 is ~10% faster than stock HD5870, benched in Crysis.
The noise level under load is similar to GTX285 -- noticeable but not irritating.

**japamd** · 03-23-2010, 05:54 AM

Originally Posted by annihilat0r

Is 5870's and 5970's min FPS really the same in Crysis? Probably some "loading" situation...

Did a quick test using my program and ambush map, that's what I got
http://i41.tinypic.com/30if1uo.png

//////////// Summary \\\\\\\\\\\\\

03/23/2010 10:48:57 - Microsoft Windows 7 Ultimate

DirectX 10 ENTHUSIAST 2X @ Map: ambush @ 0 1920 x 1200 AA 4x
==> Framerate [ Min: 25.94 Max: 55.75 Avg: 42.85 ]

HD 5970 data seems to be ok, maybe they are running a lower clocked CPU (mine is at 4.2 ghz) hence the lower min fps.

**labs23** · 03-23-2010, 05:54 AM

Originally Posted by Blacky

I didn't say makes me feel good, its just sad, the hope I had on nvidia within the release of this new gen is currently at zero...I guess ATi will not drop prices after fermi launch seems

If your sig is gonna be right, then I might be back askin myself again when can I have my first DX11 videocard.

I thought might have one by April-May TBH, ATI or nVidia no preference there, as long as it wont cost me >390$, currency considered since I'm not from the US.

Epic fail n very effn sad if true,

have to wait 3 more days though.

**jaredpace** · 03-23-2010, 06:12 AM

According to our Czech buddy, 470 & 480 overclock decently, and there is a 512sp bios floating around? - hehe can't wait to see this one

http://www.evga.com/forums/tm.aspx?m=249263

**kromosto** · 03-23-2010, 06:13 AM

true or not we are not seeing any good info for gtx4x0 series in 3 days i dont think that will change. my opinion nvidia doesnt bother about gtx4x0 series because they understand nothing good will be from them. they are producing these just for saying yes "we have dx11 on the market too". improved cards will come in a short time.

Thread: The Fermi Thread - Part 3

Thread Tools

Search Thread

Display

Tags for this Thread

Bookmarks

Bookmarks

Posting Permissions