Efficiency and its importance in gaming - updated GPU cost effectiveness list

[kalrith]

Except, if you looked at the data presented for comparable setups, the electricity cost differences between NV and AMD tend to hover at about $5-10 (+/- $2-3). So it's nowhere near $40; and therefore hardly material, despite being projected at higher electricity cost rates than the avg in US already.

I read your posts and concede to your point. At this point they're doesn't seem to be a large difference in costs for cards of comparable performance. As you've already mentioned, the only recent outliers in energy usage are the GTX470 and GTX480. So, this spreadsheet possibly would've been useful in evaluating those vs. another comparable video card.

It seems to me that the real impact of this spreadsheet could be in deciding to choose a lower-tier card over a higher-tier card in order to save additional money. While someone looking to spend $300 on a video card probably isn't worried about the extra energy costs, someone who's scraping pennies to save up for a card might appreciate the future energy costs of a lower-tier card.

Another applicable area would be in something like an HTPC or a media server, especially if they're left on 24/7. Someone might have a spare 4870 lying around (or buy a used energy hog) and think they should use it in their HTPC/media server, when they would be better served by a card that uses a lot less energy. One of the reasons I put a 4550 in my HTPC is because it uses very little energy.

 

[pandemonium]

It would be difficult to spend $.21 per KWh in the USA, where residential electricity costs $.11 per KWh average across the country:

http://www.eia.gov/cneaf/electricity/epm/table5_6_a.html

So I say "HD6990 buyers, enjoy."

Thanks for sharing that, FredGamer. This certainly gives some perspective to actual amounts around the U.S. Unfortunately that average isn't based on population; it's averaged by state. There's obviously more people living in California than there are in South Dakota, so the average will dramatically increase for the U.S. overall.

You are missing my point, because of the statement you made. For instance, sure GTX480 is available NOW for $300. But hardly anyone would buy that card when you can get a GTX570 for same price.

I got your point the first time. But, there's no way I'm going to create spreadsheets retroactively just to satisfy instances in time. Over time, however, you will see the spreadsheets grow to suit your instanced needs.

My point is in North America, electricity cost is a very small fraction of the total ownership cost (bar the extremely power hungry GTX480). From your own graphs it's evident it amounts to $5-15 annually between a worst and best case scenario (it's even smaller than that).

If you're looking at it like that, then of course electricity costs won't be considerable. You can dispute how insignificant electricity costs are for someone that has cheap energy bills and actually uses energy saving practices (computer slept or turned off when not in use; as per 6 hours use/2 hours idle), but that's not the primary focus of the spreadsheets. The primary focus is to see how effective the money you're spending is being returned in performance, IE: total cost/FPS.

Again, the purpose is to baseline total costs/FPS and see how each card stacks up to each other. Comparing energy consumption alone doesn't tell you much. Neither does FPS and initial cost standing alone.

Except, if you looked at the data presented for comparable setups, the electricity cost differences between NV and AMD tend to hover at about $5-10 (+/- $2-3).

As much as you may think this is a "side" war, never once did I promote a certain team. The data is as it is. I'm simply here advocating efficient spending and the fact that initial cost shouldn't be the only factor when comparing it to performance.

So it's nowhere near $40; and therefore hardly material, despite being projected at higher electricity cost rates than the avg in US already.

This has been greatly removed from context. The $40/month figure is from my original post and in regards to having an inefficient computer, leaving your computer on all day at idle and not using any power saving utilities, as well as paying more for your electricity.

 

[pandemonium]

So, this spreadsheet possibly would've been useful in evaluating those vs. another comparable video card.

And that's what total cost/FPS does. It brings all cards to a baseline. Find what FPS requirements you have, look for cards in that range, compare by rank in a certain kWh and ownership range and you'll understand which reference cards are best for your needs.

It seems to me that the real impact of this spreadsheet could be in deciding to choose a lower-tier card over a higher-tier card in order to save additional money. While someone looking to spend $300 on a video card probably isn't worried about the extra energy costs, someone who's scraping pennies to save up for a card might appreciate the future energy costs of a lower-tier card.

Indeed the real impact will readily show that 2 cheaper cards can ultimately be better than 1 more expensive card, in overall costs.

 

[FredGamer]

Thanks for sharing that, FredGamer. This certainly gives some perspective to actual amounts around the U.S. Unfortunately that average isn't based on population; it's averaged by state. There's obviously more people living in California than there are in South Dakota, so the average will dramatically increase for the U.S. overall.

You're welcome, but I think my point still stands.

If you look at the 8 most populous states in the USA (CA,TX,NY,FL,IL,PA,OH, MI) you have 78m people paying 10-12 cents per KWh offsetting 36m in CA paying 14 cents per KWh and 18m in NY paying 17 cents per KWh.
(populations per Wikipedia)

And CA is only 3 cents above the 11 cent average themselves.

So, I still say "Buy the HD6990 if you want it, electricity is cheap.".

(unless you live in Hawaii or Europe)

 

[pandemonium]

I'm working on updating the spreadsheet again since prices have changed quite a bit on a few cards. I'll probably finish it for tomorrow. Nothing new has been releasing so I haven't bothered updating it much lately.

In the meantime, I've added a custom input sheet for everyone to use at their own discretion. All of the formulas are already placed. All you have to do is put in your numbers, or the numbers of a card you're looking at. (Download the file in the format you wish then modify: File > Download As >)

>>>Enjoy!<<<

 

[cmdrdredd]

It's hard to say for me. I would usually upgrade only when the card I have is not sufficient for what I'm expecting. Example...I just got a HD 6950 and it replaced a GTX 295. You might say "but the 295 isn't slow at 1920x1200 and matches the 6950, even unlocked and overclocked in most DX10 benches." That would be a 100% true statement. However, I could not use DX11 at all which I really did want to do. SO it was a worthy buy for me, for now.

 

[frozentundra123456]

Crazy how cheap electricity is for Americans.

It's around 1/6 - 1/10th the price compared to a lot of EU countries.

Wow, how can you live with that. Granted we are not very energy effficient in our house because we use a lot of A/C in the summer and a space heater in the winter, but our bill runs from 100.00 to 200.00 per month. Even 5 times that would take up to 1000.00 per month.

 

[frozentundra123456]

I have another question. How much energy would a basic system use, such as a dual core conroe with one HDD and maybe a low end video card. Just wondering how much that would use at idle and under load. I would think the idle power would be under 100 watts easily, no? And under load, I would think 200 watts or less.

And thanks to the OP for all the work and bringing this topic to our attention. I dont mean to belittle your contribution in any way. However, I would agree with some other posters that energy costs for a computer are a relatively low expense unless you are running an overclocked CPU with one or more high end cards for many hours a day. I spend over 100.00 per month for phone and DSL internet, over 100.00 for cable and probably 100.00 for gas just to go to work. Not to mention mortgage payments, movies, eating out etc. So the energy costs of running my computer are minimal compared to these. (Yea, I know we need to bundle or cut costs on the phone,internet, and cable, but my wife does not want to change what we have, so you know how that goes!!)

But it never hurts to save even a little bit whenever possible, not to mention the environmental impact of saving energy.

 

[toyota]

Wow, how can you live with that. Granted we are not very energy effficient in our house because we use a lot of A/C in the summer and a space heater in the winter, but our bill runs from 100.00 to 200.00 per month. Even 5 times that would take up to 1000.00 per month.

where I live in Alabama, $350 or more is fairly common during the summer months and about $100 during winter since we have gas heat.

 

[Concillian]

I have another question. How much energy would a basic system use, such as a dual core conroe with one HDD and maybe a low end video card. Just wondering how much that would use at idle and under load. I would think the idle power would be under 100 watts easily, no? And under load, I would think 200 watts or less.

Dunno about conroe, but my i3-540 @ 4GHz with OCed 5770 is about 80 watts idle and max of about 150 in games, 130-140 is more typical. Furmark pops it over 200W.
That's from the Kill-a-watt, so at the wall. Before I got the 80+ gold PSU, it was about 5W more or so on the Kill-a-watt.

Conroe is probably a little more at idle, similar at load.

 

[happy medium]

Its important to gamers who can afford a high end power hungry card IN THE FIRST PLACE...............which also means you cant afford the games that need a high end card anyway.

I'm sorry if the truth hurts.

 

[Leyawiin]

First time I've seen this thread. I keep seeing "California" mentioned by the O.P. as such a high utility cost state. I live in the Sacramento Valley (90s and 100s all summer) and the first two tiers are $.12 and $.14 per kwh. I have a smaller 1,200 sq ft, 3/2 house, well insulated with central heat/air that I run whenever I feel the need. My typical monthly bill during the summer (with two people in the home) is $75 for gas/electric. I just got my October bill and it was $45 (yes). When I lived on the Central Coast and San Francisco I never had air conditioning in any place I lived because the climate is so mild they don't even put them in most homes and apartments. You just open the windows when its 65-75 degrees all summer.

Just some observations from someone who lives in this supposedly expensive place. Excellent compilation by the O.P., I'm not dismissing all the work he did. We have a PC on all day (waking hours that is) and I game about two hours each evening. Hardware efficiency just isn't that big of a concern given the small size of my P.G.&E. bill.

 

[GotNoRice]

Its important to gamers who can afford a high end power hungry card IN THE FIRST PLACE...............which also means you cant afford the games that need a high end card anyway.

I'm sorry if the truth hurts.

It's difficult to decipher your incoherent ramblings, but it sounds like you're saying that people who can afford a "power hungry" card can't also afford to buy games? It seems to me that the opposite would be the case, if anything. Perhaps if you proofread your post, the point you were attempting to make might come across more clearly.

First time I've seen this thread. I keep seeing "California" mentioned by the O.P. as such a high utility cost state. I live in the Sacramento Valley (90s and 100s all summer) and the first two tiers are $.12 and $.14 per kwh.

The first two tiers aren't the problem

 

[pandemonium]

-Added the following to the original post:

Power Supply Cost Effectiveness breakdown:

From the following price comparisons, it looks as though the 700W range is your best bet for getting your money's worth out of a Gold rated PSU. Here, between 650, 700, and 750W PSUs the average difference will be $56.75. I was going to include 600W, but there was only 1 Gold rated+ PSU available and that wouldn't be very fair.

Here are the average prices of the PSUs listed on Newegg per Maximum Watt listing. [All prices taken from Newegg. Only available PSUs were used and prices after rebate and shipping were included.]

650W Gold+ average: $132.24 (4)
650W Bronze+ average: $73.99 (8)
Difference: ............. $58.25

700W Gold+ average: $140.71 (5)
700W Bronze+ average: $95.81 (6)
Difference: ............. $44.90

750W Gold+ average: $154.90 (8)
750W Bronze+ average: $87.81 (13)
Difference: ............. $67.09

Here's what I was able to figure out:

• At $0.07 kWh, a Bronze rated PSU that uses 325W full load (85% per Bronze standard) 6 hours/day and 160W at idle (82% per Bronze standard) 18 hours a day will cost: $127.87/year

• At $0.07 kWh, a Gold rated PSU that uses 325W full load (90% per Gold standard) 6 hours/day and 160W at idle (87% per Gold standard) 18 hours a day will cost: $120.56/year

As you can see, the difference is pretty negligible. It would take 7.8 years to cover the initial cost difference. It really won't matter unless you're spending a lot more than $0.07 on each kWh or are running your computer at full load more often (e.g. folding@home).

-Also, corrected a calculation mistake in the custom spreadsheet. It wasn't properly calculating the power costs per year.

 

[Leyawiin]

The first two tiers aren't the problem

Especially when you rarely get out of them like we do (and I have a dreaded Smart Meter).

 

[Grooveriding]

Wow you take it dry on power charges.

Our pricing is based on time of day usage. Cheaper in the late evening and weekends, more expensive in the daytime through the week.

Overall though our costs are much cheaper than what your graph shows.

Likely because Canada has an abundance of natural resources and electricity. Those are some nasty power prices, glad we're not paying that.

 

[pandemonium]

Spreadsheet is finally updated.

 

[pandemonium]

I figured today would be a good day to get an update to the comparison sheet since pricing gets thrown around this time of year. Hopefully I'll have it all sorted out in the next couple of hours.

You guys won't believe these cliffs though:

• The HD5970 is on sale for -$200.00 off it's original price (from Newegg)! Huge steal.

• AMD cards seem to be mostly in stock, while some NVIDIA cards are very difficult to find (in some cases I had to resort to getting prices from Amazon; don't worry, I only took available and new prices to remain fair)

• AMD cards all dropped in price, and by an average of -$24.76; medium of -$6.07 I originally wasn't able to find an HD5870 but found one on Amazon (list new from an external seller). This is now the only exception to AMD cards increasing in price from last update: average -$12.44; medium -$4.05
• NVIDIA cards all increased in price (exception being GTX 570), and by an average of $70.33; medium of $5.43

 

[LiuKangBakinPie]

Our power prices is based on the area you live in. If its a posh area you pay your arse off and its extremely expensive.
If you live in a scruffy area eg like the Cape Flats you pay like 10usd a month for power even running your heater 24/7 lol

 

[pandemonium]

New sheet is up. All links have been updated.

 

[olo]

Is there any chance we could get an update of the sheet with current GPUs?

 

[pandemonium]

I certainly could. It takes considerable time and it didn't seem like anyone was demanding it any more, so I stopped. (I think the metric I was pushing here wasn't valued.)

When I get some downtime I'll get an update done.

Edit: I'll resume my interests in learning to lay out some pretty graphs instead of just the spreadsheet since I know that'll make it much easier to look at.

 

[pandemonium]

I'm waiting for AT to update their data on Bench. Also, the old spreadsheet has now been made completely defunct since 1920x1200 is no longer provided in result (as well as none of the cards on there showing in the GPU2013 list).

For the new data I'll take 1920x1080 and possibly 2560x1440. Manually inputting all of these numbers, then ordering them by kWh and ownership is tremendously time consuming. I'll need to trim this down quite a bit to make it easier just to create.

Suggestions welcomed.