[dnavas]

Originally Posted by Ailuros

Seriously those rough 2W difference should make what and by how much in $ or Euros a difference in an electricity bill?

Wrong focus. The differences between boards are marginal. A few performance percentage points, a few watts, etc. Different people will focus on different marginal benefits. Arguing about the objective importance of those differences is pointless.

All things discussed so far, I'd buy a Radeon. But my next purchase will be an nvidia card. I don't know that it'll be an improvement, but my current card flashes across monitors when starting/stopping movies. That's a problem because I use computer mainly for video editing (if I were using Adobe products, CUDA would be a selling point as well, but, I don't). This bug has been present in every driver drop I've ever gotten, though it used to be much worse. At some point, scrolling content in one monitor caused the content on my other monitor to tear and otherwise freak out. Well, it's mostly better now, so it's just another marginal benefit -- but it's driving me absolutely bat$%^* crazy. Objectively silly. Subjectively very important.

Of course, I don't know when my next purchase will be, because the current crop of cards is overpriced from my perspective (I don't play games, I "need" the card for my titling software and transition effects).

It's just as likely that the card after that will be a Radeon. I can see myself getting upset about the lack of DirectCompute performance, the poor quality of nvenc, the lack of extreme low power idles (though not likely the 1-2 watt difference in 2D mode :shrug. But right now, those aren't the marginal features that make the difference.

YMWV
-Dave

[tviceman]

Originally Posted by Ailuros

Probably yes. But until they announce anything themselves it's all speculation.

And much like a sporting event, the speculation leading up to the event is often more entertaining than the event itself!

[Ailuros]

Seems we can't keep this thread on topic no matter what Typically an economy light bulb uses less than 1/3rd the wattage of a conventional one, where one of the downside is that it takes a bit of time until they're properly heated up and reach their maximum output. It depends how many lightbulbs someone is using at a time and how many are going to get exchanged really. For lighting exclusively if you'd exchange them all in a household the difference isn't just 1% at all and there's no difference in location since light bulbs are light bulbs. Compared to the total electricity consumption the difference might be small, but it's not much difference for any other appliance or device if you don't isolate the consumption just for that particular case.

Besides heating and air conditioning is highly relative to how "power conscious" the devices/installation really are, how well maintained and how good isolated the apartment or building really is etc.

[ninelven]

Originally Posted by jimbo75

It's simple brand.

Well, yes and no. Yes, Nvidia's brand is stronger but WHY?

(Hint: It isn't marketing)

[UniversalTruth]

Originally Posted by ninelven

Yes, Nvidia's brand is stronger but WHY?

Sorry that I have to quote myself but perhaps you missed it:

Originally Posted by UniversalTruth

Because those rich milksops (or cissies) don't have a damn clue what they buy. Perhaps geforce sounds better to them than Radeon, which is actually kind of true.
Truth is that people who are deep enough in hardware field are willing to buy Radeons better.

Apple or nvidia addiction , it is the same. People in the know would prefer to go for HTC or Samsung.

So, why do people prefer iphone, istuff in general.
It is a simple matter of what people like more. It's more like a question of why person A has a lot of fans despite being what-ever-what, while person B doesn't look worse but people don't like him so much.

Simple psychology, very weird.

[Ailuros]

Originally Posted by tviceman

And much like a sporting event, the speculation leading up to the event is often more entertaining than the event itself!

There's always some sort of frustration when you find out that you've been dead wrong with it though, irrelevant if we'd like to admit it or not.

That said either I'm missing something or the difference in power consumption between a K10@225W and a GTX690@300W seems quite big. Of course do I realize that there are sizeable frequency differences and the K10 probably lacks the turbo mode, but if those early K20 specs are for real (13 SMX, 705MHz, 320bit bus, 5GB, 225W TDP), then I'm rather safe than sorry. Going up to say 850MHz+turbo for example even with 14 SMXs, 3GB 1250MHz GDDR5@384bit, doesn't sound at this point "that" encouraging for a ~250W TDP under today's conditions.

We'll find out eventually next year I guess.

[Alexko]

Originally Posted by Ailuros

Seriously those rough 2W difference should make what and by how much in $ or Euros a difference in an electricity bill? If you use common light bulbs for example exchanging them with economy bulbs will save you a lot more electricity and they last longer too.

Not much, however it's about as significant as the 20~60W difference under full load, but for much shorter periods—unless you really game a lot.

And that was my point: for usage patterns I believe to be typical, the energy consumption contest is roughly a wash.

[ninelven]

@UniversalTruth I'm sure that is what AMD has been telling themselves for years now. How is that working out?

IOW, you are wrong.

[tviceman]

Originally Posted by Ailuros

There's always some sort of frustration when you find out that you've been dead wrong with it though, irrelevant if we'd like to admit it or not.

That said either I'm missing something or the difference in power consumption between a K10@225W and a GTX690@300W seems quite big.

Nvidia's funny TDP rating system. http://www.techpowerup.com/reviews/N...TX_690/26.html GTX690 draws less average power, the same peak power, and less maximum power than the gtx480. Yet the gtx480 was listed as a 250 watt card, while the gtx690 is 300 watts.

[3dcgi]

Originally Posted by silent_guy

It's great that GPUs have come a long way in this field, but it amuses me how review website suddenly start measuring long idle too. More 'content' is more clicks, I guess.

If this GPU generation is an indication, the laptop makers don't seem to be too concerned about long idle power either...

I don't see the humor. It's a useful data point and previously there was no difference between monitor on and off idle power.

[Chalnoth]

Originally Posted by UniversalTruth

Sorry that I have to quote myself but perhaps you missed it:



So, why do people prefer iphone, istuff in general.
It is a simple matter of what people like more. It's more like a question of why person A has a lot of fans despite being what-ever-what, while person B doesn't look worse but people don't like him so much.

Simple psychology, very weird.

nVidia had much better parts than ATI by most measures for many years back in the early days of 3D. At the time, ATI parts had very widespread adoption, but very poor 3D capabilities. People would swap out an ATI card to put in an nVidia card. I don't think this perception of ATI's poor quality ever quite left them.

Now, nVidia has had its setbacks. The GeForce FX was a huge one, with ATI's parts of the time dramatically outperforming them (at least in most games...I still remember being supremely annoyed when City of Heroes played better on my GeForce Ti4200 than a Radeon 9700 Pro...given the 9700's capabilities that should never have happened). But they bounced back relatively quickly with the GeForce 6 series, and have maintained parts that were highly competitive with ATI/AMD's parts in one way or another ever since. Even when AMD/ATI has been ahead in one way or another since the FX debacle, there has never been a time when AMD/ATI was a knock-down clear win for consumers, but a few times when nVidia's parts have been. For example, both the GeForce 8 series and the GeForce 6xx series in particular have been quite dramatic wins for nVidia, though the 6xx series less so.

This isn't to say that they always won all of the benchmarks, but the GeForce 8 series combined dramatic performance with incredible computing power and tremendous improvements in visual quality, while the GeForce 6xx series pushed a dramatic increase in performance per watt, with the possibility of some truly incredible performance in the possible upcoming "Big Kepler" part.

nVidia's consistent delivery and attention to consumer interests has earned them a big following. It really shouldn't be much of a surprise.

[jimbo75]

I would say that Cypress and Juniper were as clear wins for AMD as anything Nvidia had done since the 8 series. The problem is they didn't build on it and they've just slowly slipped back into Nvidia's grasp.

[UniversalTruth]

Originally Posted by Chalnoth

nVidia had much better parts than ATI by most measures for many years back in the early days of 3D. At the time, ATI parts had very widespread adoption, but very poor 3D capabilities. People would swap out an ATI card to put in an nVidia card. I don't think this perception of ATI's poor quality ever quite left them

I don't think it's fair. You are talking about pure framerates. But as far as I know in those years it had been exactly ATi and Matrox delivering 'much' superior image quality. And I don't like nvidia because once I owned Matrox G200, changed it to geforce 2 and image quality was dramataically worse. From that moment on no respect whatsoever to that company.

Originally Posted by Chalnoth

the GeForce 8 series combined dramatic performance with incredible computing power and tremendous improvements in visual quality

I had the chance to compare X1600 Pro with 8500GT and ATi card offered superior 3D image quality.

[Dave Baumann]

Originally Posted by silent_guy

It's great that GPUs have come a long way in this field, but it amuses me how review website suddenly start measuring long idle too. More 'content' is more clicks, I guess.

For anyone that uses their PC for games a few hours a day but leaves it on for the rest of the time its a very important factor - likewise for businesses, icafes etc. It is also a key metric in some regulatory spec.

[silent_guy]

Originally Posted by UniversalTruth

. And I don't like nvidia because once I owned Matrox G200, changed it to geforce 2 and image quality was dramataically worse. From that moment on no respect whatsoever to that company.

That's some very sensible and rational thinking when making 2012 purchasing decision. Let's also not forget the uselessness of NV1 and 3DMark-early-this-century.

[Chalnoth]

Originally Posted by UniversalTruth

I don't think it's fair. You are talking about pure framerates. But as far as I know in those years it had been exactly ATi and Matrox delivering 'much' superior image quality. And I don't like nvidia because once I owned Matrox G200, changed it to geforce 2 and image quality was dramataically worse. From that moment on no respect whatsoever to that company.

Framerates weren't the issue nearly as much as compatibility. nVidia, at the time, worked pretty well with just about everything. But the ATI Rage series often didn't. This was espcially true before the Rage 128 series.

ATI really didn't come into its own until the release of the 9700 Pro.

[iMacmatician]

Originally Posted by Ailuros

That said either I'm missing something or the difference in power consumption between a K10@225W and a GTX690@300W seems quite big. Of course do I realize that there are sizeable frequency differences and the K10 probably lacks the turbo mode, but if those early K20 specs are for real (13 SMX, 705MHz, 320bit bus, 5GB, 225W TDP), then I'm rather safe than sorry. Going up to say 850MHz+turbo for example even with 14 SMXs, 3GB 1250MHz GDDR5@384bit, doesn't sound at this point "that" encouraging for a ~250W TDP under today's conditions.

Well in terms of (SP) GFLOPS/W, the K10 has 20.3 GF/W and the 690 has 18.8 GF/W (base clock), a 8.6% difference. There's a similar but somewhat larger difference between the Quadro K5000 and the 680, maybe the lower clocks on the non-GeForces have something to do with it (the clock difference is also larger between the K5000 and the 680)?

14 SMXes, 250 W, and the same GFLOPS/W as the K20 would give 727 MHz. Of course, that's ignoring memory differences, etc. That's < 10% above a 1189 MHz 8 SMX GK114 or a 1006 MHz 10 SMX GK114, at least in terms of FLOPS.

I'm not sure if I shouldn't expect NVIDIA to have at least one GK110 GeForce with something like a 260-275 W TDP (275 W instead of 250 W above gives 800 MHz).

[lanek]

Originally Posted by UniversalTruth

Sorry that I have to quote myself but perhaps you missed it:



So, why do people prefer iphone, istuff in general.
It is a simple matter of what people like more. It's more like a question of why person A has a lot of fans despite being what-ever-what, while person B doesn't look worse but people don't like him so much.

Simple psychology, very weird.

No, you have the feeling peoples prefer Iphones... Apple just own 15% of the mobile phone/device market... when Samsung own 37% of it with their galaxy line .. But Apple have a larger visibility on public news and global large public audience magazine have Samsung. With Apple you have a pure marketing effect who give the feeling, Apple is number one and there's only Apple then the other.

If we take global mobile device market: Samsung is number 1, then come Nokia, and far of them you find the other brands with Apple ( both Nokia and Samsung own 75% of the market, other are in the last 25% )
( with the bad popularity of Nokia thoses last years, it can be surprising to see them on the second place with 3.5x more devices sold of Apple, outside marketing publicity, you can see this due to the revival of the financial market in large public audience, and Apple with his incredible financial have there again a large publicity )


AMD GPU over the last years, are really stable: if we take IGP in consideration they sell 2x more of Nvidia ( Intel number one followed by AMD and Nvidia have 2x less gpu sold in this case of AMD ).. if we take only retail gpu... they hover around 50/50 ( AMD could be at 47% and Nvidia at 53%, last year or 2010 AMD was even higher with 51% vs 49% for Nvidia ).. offcourse the sell change, whatever it is when AMD release is full line up during the first 3 months of the year and Nvidia during the second half of the year, etc .


When i read some comment i have the feeling AMD own 15% of the gpu market and Nvidia 75%... it have never happend. CPU wise vs Intel this is another story..

Originally Posted by Chalnoth

ATI really didn't come into its own until the release of the 9700 Pro.

Haaa the 9700Pro Maya edition, i still have 2 of thoses sitting here ( one is surely still working, the other is part of the cards i have kill on overclocking and benching ) this was a so magic card. ( then i have got 2x 9800Pro ( killed too on overclocking ), then 2x 6600GT SLI and 2x 6800 ( killed too, but after have down many scores with the 6600 GT's ( damn good overclocking cards in SLI with subzero cooling ), then a x1800XT PE 512mb edition, 2x x1900XTX + 1 X1950XTX GDDR4 ( still working ).. ( dont laugh but i have nearly 15-20 gpu's unused who are sitting here ( need said i allways build SLI and CFX system till the 6600 period ).. ( its easy, i have absolutely no more any place for put my old motherboards box and gpu's in this room, I even have some old AtariST 1024 + Atari SH205 HDD who are still working ... If i make a list of all hardware who is sitting here...


(offtopic, i try to set for a friend 2 monitors with different resolution: a 12" and a 22" with cloned mode, if someone know how to do it on a GF104 ... ( i dont know why it allways switch to 800/600 whatever is the primary setup, on amd driver it keep both frequency and resolution in this mode for the different screen )

[Sxotty]

Originally Posted by UniversalTruth

I don't think it's fair. You are talking about pure framerates. But as far as I know in those years it had been exactly ATi and Matrox delivering 'much' superior image quality. And I don't like nvidia because once I owned Matrox G200, changed it to geforce 2 and image quality was dramataically worse. From that moment on no respect whatsoever to that company.

You are just misremembering. The ATI cards back then and openGL were death. And OpenGL was common then. The leadtek Nvidia cards had superb image quality. Nvidia cards were just more inconsistent since a variety of manufacturers made them.

[DavidGraham]

Originally Posted by lanek

if we take only retail gpu... they hover around 50/50 ( AMD could be at 47% and Nvidia at 53%, last year or 2010 AMD was even higher with 51% vs 49% for Nvidia )..

nope , NVIDIA has about 65% of the discrete GPU market share ..

[Ailuros]

lanek,

When you're done with the one sided Apple/Samsung comparison, you might want to look into profit margin figures as it might help understand why it's not so crucial for Apple to have the biggest market share in terms of units sold as long as they're still shoveling profits by the ton.

And while we're there to insert about anything irrelevant to the actual topic here, let's also do some console market share analysis, which for upcoming next generation consoles is equal to 0 for NVIDIA.

At the very least I'd love to hear some equally stupid notion that AMD got all 3 of those desigh wins simply because of any marketing stunts. When any company has N success in X markets its definitely because of marketing, black money circulating under the table or one company having a better Feng Sui in its offices than the other, nor are consumers complete idiots to consistently over and over again to fall through marketing alone for the same brand of hypothetical inferior products.

[Kaotik]

Originally Posted by DavidGraham

nope , NVIDIA has about 65% of the discrete GPU market share ..

Actually tad under 60% http://jonpeddie.com/press-releases/...-last-quarter/

[UniversalTruth]

This is partially due to laptop manufacturers like MSI... They don't even know what Radeon in their laptops whatsoever is.

http://datacomp.sk/notebooky-17-inte...pe=1&scaid=154

[Kaotik]

Originally Posted by UniversalTruth

This is partially due to laptop manufacturers like MSI... They don't even know what Radeon in their laptops whatsoever is.

http://datacomp.sk/notebooky-17-inte...pe=1&scaid=154

Laptops have nothing to do with the discrete GPU marketshare (aka add-in boards)

[UniversalTruth]

Originally Posted by Kaotik

Laptops have nothing to do with the discrete GPU marketshare (aka add-in boards)

lol

And where do all these discrete (add-in) videocards for laptops belong to?