Our Dozens Of Linux Benchmark Systems Still Burning Around 3000 kWh A Month

[profoundWHALE]

I use about 600KWh/month (no AC usage), but maybe 1/3 of that is an 30-year-old freezer I need to replace with a more efficient & less reliable unit.

It's a sad thing.

[by.peroux]

It's a sad thing.

Everyone on Phoronix knows that proprietary fridge firmware are the cancer.

[Isedonde]

I use about 600KWh/month (no AC usage), but maybe 1/3 of that is an 30-year-old freezer I need to replace with a more efficient & less reliable unit. Most of the rest is fridge, fans, PCs and a big-ass stereo power amp with high idle current..

200kWh per month for a freezer? That's ~277W. My fridge+freezer combination uses 33W. That's 290kWh per year, or 24kWh per month. (And that's not even very energy efficient compared to other new fridge/freezer combos.)

200kWh equals about 50 Euros where I live, so a new freezer would pay off *very* quickly.

Also, we (household of 2) use about 3000kWh per year. About 1000kWh for computers, probably a little more for kitchen stuff (incl. dish washer). Then <1000kWh for lights, washing machine, tumble dryer, etc.

[duby229]

Electricty is reasonably cheap in America. Most people don't turn off their TV's and leave 100w light bulbs running all the time. It's probably not a good thing on the grand scheme of things, but I'd rather have cheap electricity than what you guys in Europe have to pay. You guys are getting raped. Plus a lot of people don't realize this, but more than half of the electricity generated gets lost in transmission, which Europeans pay for, while Americans don't. And then there's the stupidly high taxes Europeans have to pay on top all that.

The cost of living in Europe is redonkulious

[ossuser]

Are the non-benchmarking systems already virtualized ?

[bridgman]

200kWh per month for a freezer? That's ~277W. My fridge+freezer combination uses 33W. That's 290kWh per year, or 24kWh per month. (And that's not even very energy efficient compared to other new fridge/freezer combos.)

Yeah, although it was pretty energy efficient when I bought it. When you say "fridge+freezer" do you just mean a regular fridge with a small freezer section or are you talking about a big chest freezer that can handle a few deer plus the usual "I live >30km from the nearest food store" frozen foods ?

200kWh equals about 50 Euros where I live, so a new freezer would pay off *very* quickly.

Electricity costs have gone up a lot here recently now the local government is larding their accumulated debts onto our bills rather than saying "see how smart we are ? look at how low your bills are" and running up huge debts in off-the-budget crown corporations.

200 KWh is the equivalent of ~35 Euros here now, but a few years back it was maybe half that. Time for an upgrade, although for some reason large freezers don't seem to be anywhere as efficient as small freezers. I would have expected the opposite since surface area relative to volume is lower for a large freezer than a small one.

[Michael_S]

Just to contrast the other numbers here, we use about 1500 kwh per month all year round.
- Family of six, so we run the washer and the dryer at least six times a week. It's not practical to hang that much laundry up to dry. We also use a lot of hot water for showers for the older people and baths for the younger kids. The electric stove and microwave are used a lot.
- L shaped house, 2050 square feet. The home size isn't that large for a family of six, but a home shaped like a cube (or even better, a sphere) would be far more efficient for trapping heat in the winter and cool in the summer.
- Most of the exterior walls are stone, which is nice for keeping out termites but poor for insulation.
- We have one television, two laptops, two desktops, a Playstation 3, and every lightbulb is LED or CFL.
- We have two dogs, so the front door is opened a lot to take them out and back in.
- Air conditioning temperature is set to 70-75F (21 - 24 C) in the warmer months and the heat is set to 65-70F (18-21 C) in the cooler ones.
- My basements get wet when it's rainy, so I have to run dehumidifiers a lot to prevent mold.

[sdack]

For when finding out about any regressions, it's much easier to track down the change / bisect than every three days.

You do not seem to have thought this through. Unless you are directly working with the compiler projects, have discussed your efforts with them and get some funding out of them is there no need to find out about regressions each and every day.

It is in fact part of the development life cycle of each release and quite normal for changes to cause regressions. The GNU C Compiler Collection has a development phase before each new release specifically dedicated to finding and resolving regressions. This has led the developers to ignore any regressions during the normal development until towards the end of a release cycle and it has become a widely accepted practise to do so.

Even if your daily benchmarks do show a regression will nobody seriously care for them. At least not on a daily basis. Reporting them here on Phoronix can only lead to some kind of sensational journalism when in fact all developers and users are expected to accept some regressions during the development. It is then even possible due to the common practise that the responsible developers already know about a regression before a patch has reached the github.

In case of a regression do the developers then often know where it is coming from. If not do they use a binary search through the patches until a culprit is found, which is far more efficient than a linear search, which is what this daily benchmarking is. Say there was like 1 regression per year, then it would take you 365 benchmarks on average to find it, whereas a binary search through the patches would need only 9 benchmarks.

In short: Trying to find regressions in daily development snapshots is pointless and thanks to the article is now also officially proven to be a waste of energy.

[ArthurBorsboom]

Hi Michael,

I do not mind to pay for the use of your website, but I do mind to pay for this benchmarking.

What percentage of the articles on your website are actually using the results of all these benchmarks?

In my opinion, turn off all the automated benchmarking and perform only manual benchmarks when there is actually an interesting topic to report, as you do, once in a while.

Is saves you a lot cash, it saves the environment, and it saves you the time to complain about the costs of the electrical bill.

[smitty3268]

what the hell is the average american household doing that uses 1000kWh/month? I use 200kWh. That's just under average for a 2 bed house here (uk) according to my electric company. My mates 5 bed, 4 people house is only using 350 ish.

Are you guys running the a/c 24/7 on 17c? how is it even possible?.

There's a lot that goes into it. For a start, the climate in England is much milder. Energy use in the US skyrockets in the summer and winters. Plus houses on average are just bigger. And although most houses are still hooked up to gas, the US has a higher percentage of all electric - meaning that heating, stoves, etc. all run on electricity. But a lot just has to do with the price. If people here were paying the same prices, they'd work a lot harder to make sure less was getting wasted.

The avg UK house uses about 390, too, so you are under the avg even there.