Again with 1000w+ power supplies on systems designed to use 50-100 watts...
No power supply manufacturer will optimize its power supplies to have good efficiency below below the 20%, which is what 50-100 watts will be on this 1250w psu.
As you can see on techpowerup.com/reviews/OCZ/ZX_1250W/5.html this power supply has 66% efficiency at 65 watts and 75% efficiency at 90 watts.
"These are the real world values that consumers may expect from a typical system (minus the monitor) using this motherboard."
No, they absolutely are not. Even if customers would buy cheap 350-500w power supplies, those would do 80% efficiency at even 60 watts load.
And who the hell is going to buy a 130$+ video card such a factory overclocked Radeon 5850 when the motherboard is supposed to targeted to value users?
Value users are going to either use the integrated video card (why no power consumption values WITHOUT a video card?) which is good enough to watch movies or even play games like World of Warcraft or they're going to get something like a sub 100$ 5750 or a 6750.
It doesn't make a big difference at idle, but it sure does at high loads, and some people are just concerned enough to compare power consumptions and lean to a solution or another based just on this.
Yeah, this is standard, as the "anand" testbed and it has to be as similar as possible across all the systems they test from i7 3960X to an E-350.
I do not accuse anand of foul play, but it is not really a fit setup for mid to low performance systems.
If you want a proper look at Llano, Brazos or similar and what they offer in a "daily" scenario, you can try looking at the SPCR reviews of these platforms.
Well, they say it's to make things comparable. But then they have power consumption charts where some systems use different components than others. And they don't put it in the chart. So one system was tested with a Radeon 5870 (low idle) and Windows 7 and another system was tested with a GeForce GTX 285 (~20-30w higher idle) and Windows XP and they don't mark it on the charts. Typically they misrepresent AMD that way. Has to be said.
I haven't used AMD (on the cpu side) myself in a long time, but I feel they are getting an undeservedly bad rep on tech sites.
I don't buy it. I understand the concept but I don't buy it.
Such a popular website like Anandtech with reasonable income from advertising can very well keep 2-3 power supplies for tests, in order to provide accurate, correct reviews.
A 600w power supply is good enough to run benchmarks on virtually ANY processor and single video card combination, even when overclocked.
With 2 or more video cards, you can very well change the power supply to a 1000w+ one.
Changing power supplies takes minutes on a test bed.
Anandtech could also just buy a batch of power supplies from the same series (to have a few as backup in case one or several break) and put them through a load tester, to get the efficiency graph, one similar to the one in the Tech Power Up article I linked.
From that point, as they have the efficiency graph, they can just measure the power consumption at the wall and estimate how much the system actually uses.
It won't be 100% exact as the ratio between 3.3v, 5v and 12v rail usage will vary between systems, but as 12v is a large percent of usage nowadays, it will be within a few percents. Close enough.
I do not like it either, but remember that Anandtech is not written by people in on physical location.
I know some of the contributors are located in different parts of Europe, and I am pretty sure they have some people scattered across the US as well. They cannot ship the same equipment between all these locations and it looks like they have purchased somewhat similar equipment for use in their test.
This 1000 watt supply is then a part of the "testbench" and you will have to visit other sites for for more realistic power/efficiency numbers when reading a review of a part that is not high end.
why not just purchase some DC power supply and take the read out directly? what the system will suck out of the socket can be computed by the efficiency of your PSU at that particular load. (say the system uses 100W, and you know your PSU has 80% efficiency at 100W, then you can compute that you system will suck out 120W from the wall)
speaking of which, when will anandtech actually buy some product for testing instead of taking what manufacture 'supplied' parts? because we all know there is absolutely no any conflict of interest when a manufacture 'send' you their 'regular' product for 'free' for this website to write an great review to 'speak truth' about their product.
Actually a 5850 is a smart choice. Maybe not this particular version, but a generic 5850. One might find such a beast for under $100 on ebay at some point in the next 6 months. That would be a great deal for a value gamer.
I'm in the process of decision making for a low cost, small form factor system for my wife to use. I'm comparing Intel i3 Vs Llano and finding I can get an i3 2125 or a AMD A8 38xx for about the same price and to me it comes down to graphics performance. My take:
1. If you want to get the best processing power, go with i3 2. If you want the best graphics you can get without the need for a pci-e card, and processing power is not that important to you get llano. 3. If you want processing power and better graphics too, get i3 with a pci-e card 4. If you want better graphics than Llano offers, once again its i3 with pci-e card
So the only scenario (at least for me) where AMD makes sense is #2 which actually happens to be what we are looking for. Testing it with an add on card doesn't help me. And if you wanted to use AMD with an add on card, you could just go with a cheap Athlon X4 - as you wouldn't need the graphics on the Llano cpu anyway.
Yes, I'd like to see power consumption figures without a discrete VGA. I'm fine with 5850 given it's a requirement of formula across motherboard reviews, but something like integrated VGA would be welcomed.
Btw, does Anandtech consider using smaller PSU in the future review? It seems ridiculous to use 1250W for such an entry-level product. It's unrealistic for mass deployment and inefficient at such low load.
It's good to see other people complaining about this, too. Power consumption is, especially compared to the high quality of their other reviews and articles, an underdeveloped appendage at Anandtech.
A lot of sentences and general flow of this article are in my opinion sub-par for a site like AnandTech. It seems like this article was first written in Dutch and then translated straight up sentence-by-sentence into English. It would really help readability if the article were edited by a native speaker.
You may wish to provide examples as I don't see what you're referring to. In my eyes, it's a very comprehensible article and I haven't had to read something twice to decipher what the author is trying to get across.
I felt the same way as ssj3gohan when I read the article, here's 3 awkward sentences in a row:
"The blue PCB which the GA-A55M-S2V is built on is a typical sighting from the current Gigabyte budget segment. Keen eyes will notice the lack of VRM heatsinks, despite this board being certified to run 100W processors. This will undoubtledly lead to little room in terms of pushing the platform."
Don't get me wrong however, this is a good review and bravo to Anandtech for reviewing more entry-level components. Just peer-review them before putting it live.
There's absolutely nothing wrong with any of those sentences. The closest to "awkward" one could get would be the use of "a typical sighting". Even that, however, is a somewhat common turn of phrase.
If I wrote tech reviews for a popular website, I would get very tired of writing essentially the same thing over and over again.
In that circumstance, I would certainly try to inject a little variety by doing exactly what Brendan appears to have done.
The power supply would ruin any board efficiency and stating the power consumption with a dedicated card makes little sense. Maybe Car reviewer now too, will show the mileage of a truck, towing a trailer. Funny but not a useful review.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
21 Comments
Back to Article
DMisner - Wednesday, March 21, 2012 - link
Isn't FM1 on the way out with Trinity's upcoming release?HW_mee - Thursday, March 22, 2012 - link
Yes it is, FM2 is the next socket for desktop Trinity, but I have not ben able to find any information on compatibility between FM1 and2.Some rumors claim that the current chipsets will work with the FM2 socket, but no info an the processors :-(
Feel free to add more rumors/info :-)
MonkeyPaw - Friday, March 23, 2012 - link
I thought I read that AMD had one last upgrade for FM1. It was assumed (hoped) they meant Trinity.Not sure what FM2 will offer that's different. Maybe add more memory channels?
mariush - Thursday, March 22, 2012 - link
Again with 1000w+ power supplies on systems designed to use 50-100 watts...No power supply manufacturer will optimize its power supplies to have good efficiency below below the 20%, which is what 50-100 watts will be on this 1250w psu.
As you can see on techpowerup.com/reviews/OCZ/ZX_1250W/5.html this power supply has 66% efficiency at 65 watts and 75% efficiency at 90 watts.
"These are the real world values that consumers may expect from a typical system (minus the monitor) using this motherboard."
No, they absolutely are not. Even if customers would buy cheap 350-500w power supplies, those would do 80% efficiency at even 60 watts load.
mariush - Thursday, March 22, 2012 - link
And who the hell is going to buy a 130$+ video card such a factory overclocked Radeon 5850 when the motherboard is supposed to targeted to value users?Value users are going to either use the integrated video card (why no power consumption values WITHOUT a video card?) which is good enough to watch movies or even play games like World of Warcraft or they're going to get something like a sub 100$ 5750 or a 6750.
It doesn't make a big difference at idle, but it sure does at high loads, and some people are just concerned enough to compare power consumptions and lean to a solution or another based just on this.
HW_mee - Thursday, March 22, 2012 - link
Yeah, this is standard, as the "anand" testbed and it has to be as similar as possible across all the systems they test from i7 3960X to an E-350.I do not accuse anand of foul play, but it is not really a fit setup for mid to low performance systems.
If you want a proper look at Llano, Brazos or similar and what they offer in a "daily" scenario, you can try looking at the SPCR reviews of these platforms.
tim851 - Thursday, March 22, 2012 - link
Well, they say it's to make things comparable. But then they have power consumption charts where some systems use different components than others. And they don't put it in the chart.So one system was tested with a Radeon 5870 (low idle) and Windows 7 and another system was tested with a GeForce GTX 285 (~20-30w higher idle) and Windows XP and they don't mark it on the charts.
Typically they misrepresent AMD that way.
Has to be said.
I haven't used AMD (on the cpu side) myself in a long time, but I feel they are getting an undeservedly bad rep on tech sites.
mariush - Thursday, March 22, 2012 - link
I don't buy it. I understand the concept but I don't buy it.Such a popular website like Anandtech with reasonable income from advertising can very well keep 2-3 power supplies for tests, in order to provide accurate, correct reviews.
A 600w power supply is good enough to run benchmarks on virtually ANY processor and single video card combination, even when overclocked.
With 2 or more video cards, you can very well change the power supply to a 1000w+ one.
Changing power supplies takes minutes on a test bed.
Anandtech could also just buy a batch of power supplies from the same series (to have a few as backup in case one or several break) and put them through a load tester, to get the efficiency graph, one similar to the one in the Tech Power Up article I linked.
From that point, as they have the efficiency graph, they can just measure the power consumption at the wall and estimate how much the system actually uses.
It won't be 100% exact as the ratio between 3.3v, 5v and 12v rail usage will vary between systems, but as 12v is a large percent of usage nowadays, it will be within a few percents. Close enough.
HW_mee - Thursday, March 22, 2012 - link
I do not like it either, but remember that Anandtech is not written by people in on physical location.I know some of the contributors are located in different parts of Europe, and I am pretty sure they have some people scattered across the US as well. They cannot ship the same equipment between all these locations and it looks like they have purchased somewhat similar equipment for use in their test.
This 1000 watt supply is then a part of the "testbench" and you will have to visit other sites for for more realistic power/efficiency numbers when reading a review of a part that is not high end.
mariush - Thursday, March 22, 2012 - link
ಠ_ಠThis is just another point for computing the efficiency graph of the power supply and report the estimated real usage of the system
We all know efficiency of a power supply changes when running from 110-120v (us) compared to running from 230v.
The 65w idle on 230v in Europe can very well be 70w in US, due to less efficiency when running on US mains power.
It would be more honest to report the actual consumption, after factoring losses in psu.
seanleeforever - Thursday, March 22, 2012 - link
why not just purchase some DC power supply and take the read out directly? what the system will suck out of the socket can be computed by the efficiency of your PSU at that particular load. (say the system uses 100W, and you know your PSU has 80% efficiency at 100W, then you can compute that you system will suck out 120W from the wall)speaking of which, when will anandtech actually buy some product for testing instead of taking what manufacture 'supplied' parts? because we all know there is absolutely no any conflict of interest when a manufacture 'send' you their 'regular' product for 'free' for this website to write an great review to 'speak truth' about their product.
Shadowmaster625 - Thursday, March 22, 2012 - link
Actually a 5850 is a smart choice. Maybe not this particular version, but a generic 5850. One might find such a beast for under $100 on ebay at some point in the next 6 months. That would be a great deal for a value gamer.Ratman6161 - Thursday, March 22, 2012 - link
I'm in the process of decision making for a low cost, small form factor system for my wife to use. I'm comparing Intel i3 Vs Llano and finding I can get an i3 2125 or a AMD A8 38xx for about the same price and to me it comes down to graphics performance. My take:1. If you want to get the best processing power, go with i3
2. If you want the best graphics you can get without the need for a pci-e card, and processing power is not that important to you get llano.
3. If you want processing power and better graphics too, get i3 with a pci-e card
4. If you want better graphics than Llano offers, once again its i3 with pci-e card
So the only scenario (at least for me) where AMD makes sense is #2 which actually happens to be what we are looking for. Testing it with an add on card doesn't help me. And if you wanted to use AMD with an add on card, you could just go with a cheap Athlon X4 - as you wouldn't need the graphics on the Llano cpu anyway.
loimlo - Saturday, March 24, 2012 - link
Yes, I'd like to see power consumption figures without a discrete VGA. I'm fine with 5850 given it's a requirement of formula across motherboard reviews, but something like integrated VGA would be welcomed.Btw, does Anandtech consider using smaller PSU in the future review? It seems ridiculous to use 1250W for such an entry-level product. It's unrealistic for mass deployment and inefficient at such low load.
ssj3gohan - Thursday, March 22, 2012 - link
It's good to see other people complaining about this, too. Power consumption is, especially compared to the high quality of their other reviews and articles, an underdeveloped appendage at Anandtech.ssj3gohan - Thursday, March 22, 2012 - link
A lot of sentences and general flow of this article are in my opinion sub-par for a site like AnandTech. It seems like this article was first written in Dutch and then translated straight up sentence-by-sentence into English. It would really help readability if the article were edited by a native speaker.silverblue - Thursday, March 22, 2012 - link
You may wish to provide examples as I don't see what you're referring to. In my eyes, it's a very comprehensible article and I haven't had to read something twice to decipher what the author is trying to get across.Taft12 - Thursday, March 22, 2012 - link
I felt the same way as ssj3gohan when I read the article, here's 3 awkward sentences in a row:"The blue PCB which the GA-A55M-S2V is built on is a typical sighting from the current Gigabyte budget segment. Keen eyes will notice the lack of VRM heatsinks, despite this board being certified to run 100W processors. This will undoubtledly lead to little room in terms of pushing the platform."
Don't get me wrong however, this is a good review and bravo to Anandtech for reviewing more entry-level components. Just peer-review them before putting it live.
ggathagan - Thursday, March 22, 2012 - link
There's absolutely nothing wrong with any of those sentences.The closest to "awkward" one could get would be the use of "a typical sighting".
Even that, however, is a somewhat common turn of phrase.
If I wrote tech reviews for a popular website, I would get very tired of writing essentially the same thing over and over again.
In that circumstance, I would certainly try to inject a little variety by doing exactly what Brendan appears to have done.
silverblue - Friday, March 23, 2012 - link
The only thing I could recommend here is adding a small amount of punctuation, but that's it. For example:"The blue PCB - which the GA-A55M-S2V is built on - is a typical sighting from the current Gigabyte budget segment."
The Clownspotter - Sunday, July 1, 2012 - link
The power supply would ruin any board efficiency and stating the power consumption with a dedicated card makes little sense. Maybe Car reviewer now too, will show the mileage of a truck, towing a trailer. Funny but not a useful review.