PowerColor Radeon HD 5850: Worth the Wait

Want to talk about one of the articles in SPCR? Here's the forum for you.
mczak
Posts: 147
Joined: Wed Jul 21, 2004 6:13 pm

Post by mczak » Tue Jan 12, 2010 7:16 pm

DAve_M wrote:
mczak wrote:rpsgc is sort of correct. It's true the card won't downclock the ram at all (and the gpu less) if more than one monitor is attached.
The reason for this afaik is that it is only possible to adjust memory clock during vblank, and since multiple monitors don't run synchronized it is impossible to get this to work.
What about if the 2 monitors both had the same refresh rate? Can't the monitors then be kept in sync?
Since the clocks are still separate (with DVI/VGA at least) they still don't run synchronized. That said, I was wondering about that too - the chip has some facilities to synchronize displays (e.g. it can trigger vblank on one display based on the other and stuff like that, this may screw up timing a bit but I guess if it's reasonably close it might work). But in any case, even if it might technically be possible, the driver doesn't support it.
Display port is not clock based like DVI/HDMI/VGA. So you may not get any problems with 2 monitors attached as long as one is display port.
Hmm I thought it didn't help. The link clock may be the same for all DP displays, but I don't think it changes anything wrt display timings.

DAve_M
Posts: 198
Joined: Sun Sep 02, 2007 9:58 am
Location: UK

Post by DAve_M » Wed Jan 13, 2010 7:18 am

My understanding is that the computer monitor is a slave to the clock in the graphics card. So I thought if the graphics card could trigger vblank on all the displays at once, then all the displays could be kept in sync. Even if all the displays get turned off/go blank for a couple of seconds while it is being set up, it would still be worth it.

Sounds like display port overcomes a lot of limitations. I will wait and see if the situation is improved with 100% displayport eyefinity setups. I have a feeling it will work properly.

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

Post by mczak » Wed Jan 13, 2010 9:17 am

DAve_M wrote:Even if all the displays get turned off/go blank for a couple of seconds while it is being set up, it would still be worth it.
I think they'd only blank for a split second. But I think these clock changes happen quite dynamically, so with for instance Aero moving a window might constantly flicker :-). Even if it's not quite that dynamic, it was apparently considered not acceptable, and I sort of agree - it could be quite annoying, and some unknowing user would probably think the graphic card is defective...

irusun
Posts: 8
Joined: Sat Jan 22, 2005 10:23 am
Location: seattle

Post by irusun » Wed Jan 13, 2010 11:02 am

I don't see how this video card gets a "SPCR Recommended" rating?

Aside from being energy efficient (does that in any way translate to a quieter overall system in the real world?), it appears from the review to be relatively loud at idle (when it most counts) and difficult to make quiet.

While I think it's great for SPCR to review such a product so potential buyers will know what they're in for and how it compares to other products, it just seems incongruous to suggest that this product caters to those looking to build a quiet computer (or even a quiet "gaming" computer).

Please remove the "SPCR Recommended" rating from the review or add a prominent qualification to why it's been marked as "SPCR Recommended", because it's certainly not obvious from the review.

flippant
Posts: 7
Joined: Wed Jan 13, 2010 10:41 am
Location: Norway

Post by flippant » Wed Jan 13, 2010 11:07 am

I wonder why it can't stay quiet during idle? It's obviously not warm enough to warrant that amount of air circulation.

Could this be handled with a bios/driver update?

lm
Friend of SPCR
Posts: 1251
Joined: Wed Dec 17, 2003 6:14 am
Location: Finland

Post by lm » Wed Jan 13, 2010 11:29 am

Jordan wrote:It seems strange to me when you guys are given higher end products to review that clearly don't cater towards the low-noise expectations or requirements of the average SPCR reader :? Does the name of the site not give it away?
Some of us need our PC to be quiet AND still have high performance to be able to do things a low performance hippie PC can't do. I for example might replace my 8800GTS 512MB with a 5850 or a 5870. Though I have to admit things became much easier in the silent department when I moved my computer to the closet next to my study.

The most quiet PC is a non-existing PC. But while it is silent, it also does nothing.

Jordan
Posts: 557
Joined: Wed Apr 28, 2004 8:21 pm
Location: Scotland, UK

Post by Jordan » Thu Jan 14, 2010 1:46 am

lm wrote:
Jordan wrote:It seems strange to me when you guys are given higher end products to review that clearly don't cater towards the low-noise expectations or requirements of the average SPCR reader :? Does the name of the site not give it away?
Some of us need our PC to be quiet AND still have high performance to be able to do things a low performance hippie PC can't do. I for example might replace my 8800GTS 512MB with a 5850 or a 5870. Though I have to admit things became much easier in the silent department when I moved my computer to the closet next to my study.

The most quiet PC is a non-existing PC. But while it is silent, it also does nothing.
I was expecting at least one response like this... I should expanded a bit on what I meant :)

Obviously I realise we all have different needs (I used to be in your boat!), but I'm talking about the general interest of the majority of SPCRers. It's just the nature of the niché that a much smaller percentage of the SPCR audience is interested in things that make a fair bit of noise vs things that don't.

I'm not complaining about high-performance parts, just bog-standard parts that have little relevance to the site's theme. At the very least I'd expect to see cards fitted with aftermarket coolers submitted instead. This isn't just another Tom's Hardware clone! :)

If you were intending on fitting an aftermarket cooler or water cooling block anyway then you can largely ignore the noise part however so I'm not saying it's not worth SPCR's time, by any means.

DAve_M
Posts: 198
Joined: Sun Sep 02, 2007 9:58 am
Location: UK

Post by DAve_M » Thu Jan 14, 2010 6:17 am

Jordan wrote:It seems strange to me when you guys are given higher end products to review that clearly don't cater towards the low-noise expectations or requirements of the average SPCR reader :? Does the name of the site not give it away?
I requested a follow up article with noise performance for 5850 + accelero S1. I think after that, the card could earn a SPCR recommended award. But right now it seems that the awards are given away far too easily.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Thu Jan 14, 2010 8:19 am

Most review sites wouldn't bother with any justification about handing out awards...

If we look at this card only from a noise perspective, it's good but not exceptional... though 18/24 dBA for idle/max is amazing by gaming standards. But consider the embedded technology, the 21W idle, max power scaled back from previous gen cards, and the big jump in performance/features. I suppose we'd be inclined to give just about any 5850 a recommendation, a recognition of the work ATI/AMD has put in to finally curb GPU power while still making significant gains in performance. In that sense, our recommendation is less for this card per se and more for the whole 5800 series. This goes against the norm for consumer product awards... but it's sort of like wanting to award gold to all the olympic competitors who finish within 0.1s of each other.

Call our position quirky, idiocyncratic or even inconsistent if you like -- we simply think it deserves to be recommended... if you're looking for a high end gaming graphics card. This recommendation is for the PowerColor 5850, but we're quite willing to extend it to every 5850 variant with the same acoustics.

Naturally if we had half a dozen 5850 variants and found substantial acoustic differences between them, only the best would get recommended. We don't know at this point whether they would all sound the same, having only had this one in hand, but since every 5850 from every brand looks like it uses the reference cooler, we suspect they do -- sound the same. In an ideal world, we'd have got that big sampling... but it's not an ideal world, and many video card makers don't even think it worthwhile to send any samples to us -- our testing doesn't highlight enough the differences they think are important.

lm
Friend of SPCR
Posts: 1251
Joined: Wed Dec 17, 2003 6:14 am
Location: Finland

Post by lm » Thu Jan 14, 2010 9:06 am

Jordan: Well I actually agree with you now that you elaborated. Although this review does indeed at least tell you if you can consider the reference-cooled model or not for your silent rig, but I'd love to see some non-reference-cooled models in comparison!

BDBrot
Posts: 2
Joined: Thu Jan 14, 2010 11:36 pm
Location: Australia

Post by BDBrot » Fri Jan 15, 2010 12:03 am

Hi, I agree that seeing some reviews of non reference models would be excellent. However, I think many of us would like to see each and every product which is even remotely noise relevant tested here... Wouldn't it be nice to just look up exactly that particular item which you were interested in?

Unfortunately, I don't expect that SPCR has the kind of resources to test much more than we are already seeing. After all, a card like this costs a fair chunk of money, so, as companies don't seem to be too generous in giving samples to SPCR, I imagine that the problem is not so much SPCRs willingness to do the reviewing.

For those who would like to see more reviews, I believe I remember seeing a hardware donation page somewhere. I imagine that lending would also be an option - if it is feasible getting the hardware to SPCR labs and pack. But of course who would be willing to part with their shiny, new (and not to mention expensive) toy?

Conclusion? I'm sure SPCR will test it, if they manage to get their hands on it somehow, or other.

So, getting back from wishful thinking to the topic... I agree that, going by the temperature, the fan should be able to run slower in idle... I'd be really keen to hear anyone's experience with trying to reduce the idle RPM! The bios option sounds very promising! Has anyone at all tried this? Does anyone know what minimum percentage the fan will physically start to run at?

Cheers,

BDB

Jordan
Posts: 557
Joined: Wed Apr 28, 2004 8:21 pm
Location: Scotland, UK

Post by Jordan » Fri Jan 15, 2010 1:25 am

I agree it would be madness for SPCR to turn down reviewing a hot product such as this! It's good for publicity, if nothing else. I just question the manufacturer's thinking in this case is all, other than free brand publicity, which you can't really argue with. Do SPCR have to return the card in this case or do they get to keep it? It's no secret that SPCR reviews focus on silencing and they're not just another generic hardware review site.

Conversely, if a company has put special consideration and effort into making a product that's quieter than the competition and they're not just making unfounded claims (Thermaltake spring to mind here!) then they'd be crazy not to submit to SPCR for free publicity to a market they're directly targeting :)

Re: Mike's comments - I agree. Though I think it's long overdue and I can't see what the delay has been in getting technology that's been incorporated into desktop CPUs for years and laptop CPUs for even longer, I still praise AMD for getting the jump on their main competitor in this respect.

From my personal perspective though, cutting down on heat/power consumption isn't much of a benefit if this potential isn't used to then cut down noise to the same degree. This is why I can't wait to see what come companies to with after-market coolers or what enthusiasts can achieve fitting them themselves.

What I really want to see is a semi-passive design on any of the 5xxx series :D But I'm not holding my breath...

EDIT: Oh, and while I'm still a regular forum member I don't feel I'm in the position to criticise SPCR's methods in any way, so I wouldn't have been complaining about them anyway :oops: Have always said I'd pay for my upgraded membership when I am in the cash-positive position I'd like to be in, especially as Mike has always answered my questions, but until that day it just sounds like a cliché :lol: I also didn't visit the site around the time they were doing the lab upgrade as I had lost my interest in hardware for a while.

Telamon
*Lifetime Patron*
Posts: 18
Joined: Wed Jul 20, 2005 8:34 am
Location: Seattle

Post by Telamon » Fri Jan 15, 2010 10:22 am

I couldn't wait for this review and already made my 5850 purchase back in November but I'm glad it's here for all my gaming SPCR comrades. The stock cooler is of course insufficient but I wanted to point out this thread where falcon26 put a Twin Turbo Pro on his 5870 and has already spawned a discussion about the 5850 in turn. The Scythe Musashi was also touched on in page 3 and I can qualitatively state that it works well on my Sapphire variant. I'm seeing less than 60C at load with the two fans running at about 40%, however, I too would like to see how it stacks up to the competition using SPCRs rigorous testing.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Fri Jan 15, 2010 1:19 pm

Jordan wrote: What I really want to see is a semi-passive design on any of the 5xxx series :D But I'm not holding my breath...
Have you not seen the passive 5750 models?

viewtopic.php?p=494520

Powercolor has one with and one without the 6 pin power connector. The one without is called:

PowerColor Go! Green HD5750 1GB GDDR5 AX5750 1GBD5-NS3DH

http://www.powercolor.com/Global/produc ... uctID=6838 shows specs

PowerColor Go! Green HD5750 = http://www.powercolor.com/Global/NewsInfo.asp?id=929

I'm still waiting for the Green version to show up at a retail site and for someone to review the Green vs the non green Powercolor 5750 so we can have an apples to apples power comparison.

Jordan
Posts: 557
Joined: Wed Apr 28, 2004 8:21 pm
Location: Scotland, UK

Post by Jordan » Sat Jan 16, 2010 9:06 am

I hadn't actually, thanks for the link. The Green looks fantastic, any idea what they've done to the card to reduce power consumption?

I did have something slightly higher up the line in mind though (5850 for instance), and a semi-passive solution to allow for a faster card. I really don't mind noise once the game is up and running and I have the speakers on :D It's when the PC is idle and being used as a web browser or office work that I can't stand any noise.

The 5750 should be plenty for me though when I upgrade from my 7300GT :lol:

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Thu Jan 21, 2010 3:21 pm

There are theories (or are they hypotheses?). Any way this is fact + conjecture:

1. Many video cards either "factory overclock" the ram, GPU, or both
2. Many video card makers will assume that if the card isn't already overclocked that you'll do it if the fan is in place.
3. The fan on many video cards may draw enough power at full load to be significant.
4. There are bins of any mass produced chip. Generally giving you Fast+High Power, Fast+Medium Power, Slower+Medium Power, Slower+High Power. The presumption is that Fast+High go with active cooling, Fast+Medium will go in Green product like this, Slower+medium will go in a budget card with active cooling to keep costs low, and slower+high power gets tossed in the trash or sold depending on the state of the manufacturing process.

My conjecture is that the Green Passive 5750 will not factory overclock, saves power on the non existent fan, has a GPU and RAM chosen from that favorable bin that ensure it makes it below the 75W barrier so they can avoid the 6pin power connector and comfortably be cooled by a passive solution.

If you are the type that takes a passive video card (2 slot solution) and adds a very slow 100mm or 120mm fan to it (home brew 3 slot solution) you might get a slight overclock on it but most of us SPCR types will be happy as long as it works at stock speeds with no fan and will only break out the low speed fan if temps at stock are an issue. Although I'm the type to strap a fan on there just to measure the temp change.

javitxi
Posts: 64
Joined: Thu Jan 01, 2009 7:30 am
Location: Madrid (Spain)

Post by javitxi » Fri Jan 22, 2010 6:25 am

MikeC wrote:Naturally if we had half a dozen 5850 variants and found substantial acoustic differences between them, only the best would get recommended. We don't know at this point whether they would all sound the same, having only had this one in hand, but since every 5850 from every brand looks like it uses the reference cooler, we suspect they do -- sound the same. In an ideal world, we'd have got that big sampling... but it's not an ideal world, and many video card makers don't even think it worthwhile to send any samples to us -- our testing doesn't highlight enough the differences they think are important.
Mike, just a short question about the wide range models using the stock cooler. The little experience I have with them by changing stock cooler and other issues, if the most part of 5850 i.e. come with stock ATI cooler, can I assume that they have the same PCB revision of the card? And if my assumption is correct, so then, what is the point buying one card of X brand (Sapphire i.e) if it is the same (same PCB model, same stock cooler) as Y brand (i.e Power Color)? Does the assembler make any significant difference in quality (maybe capacitors installed?) between them (no OC models, no added-RAM models)?

Many thanks, because is something I have been wondering long time ago

Javier

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Fri Jan 22, 2010 7:10 am

javitxi wrote:Mike, just a short question about the wide range models using the stock cooler. The little experience I have with them by changing stock cooler and other issues, if the most part of 5850 i.e. come with stock ATI cooler, can I assume that they have the same PCB revision of the card? And if my assumption is correct, so then, what is the point buying one card of X brand (Sapphire i.e) if it is the same (same PCB model, same stock cooler) as Y brand (i.e Power Color)? Does the assembler make any significant difference in quality (maybe capacitors installed?) between them (no OC models, no added-RAM models)?

Many thanks, because is something I have been wondering long time ago

Javier
My conjectures on this matter are probably no better or worse than yours. :lol:

I have visited some mobo/vga card makers' factories in Taiwan (and even Toronto) and always came away impressed at the level of mfg technology they all seem to incorporate. But how this translates to quality differences between products made from the same prototypes I've never been certain. I do think there are differences in things like capacitors, PCBs, or fan bearings but how to ascertain these differences without examining the boards all on hand, side by side??? And even then the differences may not be apparent. I suppose the products that sell for higher prices may reflect high quality, production cost -- or conversely, merely the marketers' desire to make you believe these things.

I agree w/ you -- it's not easy.

Having said all that, there seems little question that cheap budget motherboards by lesser brands (maybe ECS, Jetway....) don't last as long as nearly identical boards from better brands (like Asus, Gigabyte...) that cost more, and they often have more niggly problems. I just don't have as much experience w/ VGA cards, but I suspect similar trends may exist, at least among cheaper VGA cards.

javitxi
Posts: 64
Joined: Thu Jan 01, 2009 7:30 am
Location: Madrid (Spain)

Post by javitxi » Fri Jan 22, 2010 8:04 am

Well, I'll probably have bad luck with Asus on VGA:

I have lots of trouble with one 4850: I have put on an Accelero Twin Turbo and could play at 1680x1050 (as I had expected and w/ help of 3dmark & Everest, the "Glaciator" at 95% would not be enough). Then, installing w7 x64 in someway the VGA crashed and it shows red tiny spots all over the presentation (8 months of working VGA). I had put the original heatsink, went to the shop and 7 weeks later, I had what it was supposed to be my VGA fixed as the repair sleeve stated. Then I looked to the serial number of the VGA and surprise surprise, it wasn't the same :roll: Well, after exhaustive search, I noticed that on RAM chip it was like someone had put some kind of resin and at the bottom of the Glaciator heatsink, it hadn't got any rubber black circles for the screws to prevent harming VGA's PCB.

Well... I thought it was ok because I had only changed the heatsink and get back a ""new"" card,... and the past week while doing some prime95 tests over CPU running passive without the fan of Noctua NHC-12P, suddenly the VGA crashed and there was no image. Then the motherboard (which is an Asus P5Q-E) told me by beeps that the VGA was failing. Many trials proved and the same results: the VGA was dead.

So, another time I have to go to the shop to validate warranty.... although Asus states 3-year warranty for their VGA products :lol: (I couldn't find the link in English, sorry)

So, I don't think Asus would be a great assembler of ATI's VGA, although with a budget Asus nVidia passive 8400GS no prob at all (after ~2 years ussage on my father's Pc), which I have today on my Pc for testing.

Then, over the Internet for ATI's VGA many recommend Sapphire and HIS/MSI/XFX meanwhile Power Color have striked back hardly with the new 5xxx series overwhelming the others (except for the Sapphire Vapor-X series).

By all this things, every time I look over a VGA, I search for an extensive review which shows me mainly the heatsink of the VGA (which resembles airflow), and with my little experience on mid-top VGAs, I'll say on ATI VGAs:

- Asus is no good for ATI VGAs
- Sapphire probably make the difference on the components they assemble although maybe it use stock ATI heatsink
- They say Power Color uses low quality products, but with 5xxx it has done a great job by innovating on passive models (5750), so I would get a point of trust to the brand
- Gigabyte is known to have quality prodcuts at competitive prices, and they have a 4850 totatly passive cooled. Probably it would be a choice like Power Color, but I don't trust on Gigabyte at all on VGAs as well as it has very good quality motherboards
- MSI/HIS/XFX uses the stock coolers or the different stock coolers for the different PCB revisions (except on one HIS heatsink), so what would be the differeneces over the 3?

So, my final thoughs are:

Sapphire really makes the difference (or at least the capacitors seem so apart of Vapor-X and another things) although it is a little more expensive than the others.
Power Color on 5xxx I probably say they make a great difference, but what about quality? I don't know

Looking for quiet & eco-friendly -> Power coolor
Looking for peace of mind & no problems & quality -> Sapphire

What is your opinion Mike?

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Fri Jan 22, 2010 8:23 am

My opinion -- as I already said, I really don't feel I have enough experience to have one.

swinster
Posts: 19
Joined: Wed Mar 30, 2005 4:28 am

Post by swinster » Tue Apr 27, 2010 11:47 am

I don't suppose there is any chance of getting hold of a SAPPHIRE VAPOR-X HD5850 1GB GDDR5 PCIE so that you can run a comparison with this card. It would appear that the Powercooler and Sapphire are registering as the lowest noise cards of this type on the market.

What would also be useful would be if you could also test the same cards with after market coolers.

Post Reply