Source: Tom's Hardware UK – Keywords: radeon, hd, 5770
Categories: Graphics
Power And Noise

ATI rates the Radeon HD 5770 for 18W of power consumption at idle and 108W maximum board power at load. The Radeon HD 5750 is even lower at 16W idle and 86W under load. The delta there checks out almost exactly, as our Radeon HD 5750 sample used two watts less than our 5770 at idle, and 22W less (instead of the expected 21W) with FurMark running at full tilt.
The only card in this lineup able to use less power at load is the Radeon HD 4770—though it consumes more power at idle than either of the 5700-series boards.
Nvidia’s G92-based GeForce GTS 250 and GT200b-based GeForce GTX 260 lay down impressive idle numbers, given the fact they’re manufactured using 55nm technology. They can’t compete with the power savings of ATI’s 40nm boards under load, though (the exception being the power-hungry 5850).

We were much more concerned with noise last month. The flagship cards push the envelope, generating the most heat under duress. When the same technology is derived down into the mid-range, it’s rarely as hot or loud.
Indeed, all of these boards are able to idle along fairly quietly with complete systems generating less than 40 decibels. With a FurMark stress test running, fans are forced to keep temperatures under control, but never really need to operate at full tilt. All of ATI’s 40nm GPUs turn in fairly similar results, and even the CrossFire configurations aren’t intrusively obnoxious under load.
- Previous page Benchmark Results: Batman: Arkham Asylum
- Next page The CPU Scaling Story: From 2.66 GHz...

on the test setup page;
"Corsair Dominator 4GB (3 x 2GB)"
shouldn't it be 2 x 2GB?
shouldn't these cards be able to overclock well, considering how energy efficient they are?
Where's my 5830?
The procesor bottle neck test you presented is just horrible!
For reasons readers will never know is why you left out AA in some of the bottle neck tests but added AA in some the previous benchmarks. For example on page 8 testsing farcry2 had test with AAx4 but not on the CPU bottle neck test, which is what you SHOULD be tsting for!
AA/AF uses the CPU where the graphics card can't handle it and for whatever reason you will come up with to answer me, i think you are simply trying to advertise the Core i5-750 CPU as some sort of incredible cpu that doesn't need overclocking. Are you being paid to do this on purpose are are you just really that shit at making decent review?
umm AA/AF is done purely on the graphics card.
So you're saying that someone forking out small money for a mid range gfx card sees the value in supporting three monitors? Sorry but I really can't see that proposition!
And the DX11 support to me seems a red herring too. Bit-Tech gave a more luke-warm reception and I have to say that I agree with them. DX11 is hardly a sell if the card can only just keep up with the upper-mid range in DX10? Its not like DX11 is going to increase FPS.....
Bitstreaming? Don't know enough to comment!
Re: Your Rap song...
I see you're new to the Nerdcore Hip-hop scene. May I recommend you try listening to a little MC++
- NO, really, it exists! ;-)
Thanks Chris.
Another good article.
Although Intel won't want to hear your last sentence I am sure AMD wins either way.
I think that was eskimo's point .. .wasn't it? Be happy with the Q3 earnings report.
We have long know that money spent on better graphics (or a second graphics card in SLI / CF) once you have a quad core runnng at around 3Ghz (Intel i7 or AMD Phenom II) gives much better value return for gaming than anything else.
wheres the overclocking?
Its not like DX11 is going to increase FPS.....Bitstreaming? Don't know enough to comment!
There's already battleforge out there that renders using dx11 subroutines, etc thats showing increase in fps over the dx10 version. So technically, its a game that's taking advantage of a dx11 feature to get higher fps.
http://www.fudzilla.com/content/view/15721/1/
@ Anonymous 13/10/2009 16:37
Perhaps some of the people "forking out small money for a mid range gfx card" are in the market purely for a 3-monitor setup that previously they'd have to use multiple cards for? I don't know.
I would buy a 5770 for an HTPC (the idle power rating / core clock look pretty good to me) but not for anything else.
shouldn't these cards be able to overclock well, considering how energy efficient they are?
Yup. They should.
Worth noting ATI cards seem to have a lot of problems with decoding video streams - in particular there is an open problem with Win7 MCE and BBC HD over satellite.
@ anon
im pretty sure one of the big selling points of dx11 is how much more efficent it is, and how it can offer incresed performance with a higher level of visual fidelity.
so having a dx11 card will most likely prove to allow for higher framerates than a game run with a higher spec dx 10 only card.
I bought this card. My x1950 bit the dust in my last move, and I need a replacement under $200. My mobo only supports a single card, and this is probably the only upgrade I'll do for 2 years or more. At that point, my CPU will likely be a bottleneck (core 2 duo 2.1ghz, overclockable to 2.8 or so).
My thoughts: Of single card setups under $200, this one seems like the best long term value. Especially overclocked, it's as powerful as anything else at this price point, and it has DX11 support. It killed me that I bought my x1950, only to have DX10 come out and instantly obsolete my card... the 5770 won't have that problem. I recognize that I'm giving up a few fps in the short term, but in the long term... this card will be competitive for a longer time in its field.
i have a 5750 and i have oc it to gpu clock 750 and memory clock to 1250 when the card is at 97 % activity the card gets up to 64C! is that dangerous and can destroy my card?