AMD Beats Intel in Power-Efficiency Study

Ted Samson writes "AMD Opteron servers proved up to 15.2 percent more energy-efficient than those running Intel Xeon in a server-power-efficiency test performed by Neal Nelson and Associates, InfoWorld reports. That translates to annual electricity savings between $20.29 per server and $36.04 per server, depending on the workload, the study concluded. The benchmark tests were conducted on similarly configured 3GHz systems running Novell SUSE Linux, Apache2, and MySQL."

Multiple OS-es

I would have liked to seen them test it with the 'big three' OSes of Linux (RH and SUSE), VMware and WIndows. It would have been nice to see if the power management of the operating systems would have come into play some above and beyond just the single OS. Besides the OS the applications used run on any of those platforms.

GHz != Performance

A 3.0GHz Core 2 is more than 15% faster than a 3GHz Opteron for many tasks.

AMD is doing better at idle speeds (Intel definitely needs to crank Penryn down more when it's not in user) but if this survey compared equivalent performance processors, the differ

3.0Ghz May Not Meen Equal Performance

Both systems had 3.0Ghz CPUs and similar amounts of RAM. But did they offer the same performance? If both servers were being pushed 100% would one be able to server more users than the other? If the servers were never pushed to 100% then the test is not really a like-with-like comparison. I imagine that one CPU performs better than the other (and I'd expect right now that's the Intel one). Perhaps a 2.66Ghz vs 3.0Ghz test is closer to the same performance?

Re:

If you read page 6 of the test description, under Test Design, they say this:

This test is not intended to measure the maximum throughput that a server can deliver

The test simulates credit card transactions coming in at a controlled rate. So this test wo

Re:

The test simulates credit card transactions coming in at a controlled rate. So this test would let someone get an idea of their operating costs.

No, for just reason the parent stated - it doesn't tell you how many computers of each type you'd need to han

Re:

"This is a very important special case (unless you somehow have a steady workload 24/7, which I think would be highly atypical)"
Yes for most servers but for say Rendering farms or HPC clusters "I will not use b word" it could be very typical.

In the low end

Yeah, 'cos render farms are sooo common these days

Render farms, HPC, etc. are a tiny percentage of all the "servers" out there.

I don't think we should throw away the test results because of a few render farms.

Re:

The savings from consolidating a number of server workloads on a smaller number of machines using something like Xen...

The problem with that is you often don't have a steady load 24/7. At 3am, you need 1 server; at 8:30am, you need 40. Since virtualiza

Re:3.0Ghz May Not Meen Equal Performance

Anandtech recently did that kind of power efficiency vs. performance test actually: (2.6Ghz vs. 2.33ghz), with AMD coming out on top: http://www.anandtech.com/IT/showdoc.aspx?i=3039 [anandtech.com]

Re:

The opteron is actually faster, noticeably faster in fact. Or atleast it's slower equivalent was faster so I can only imagine the 3ghz model is even more powerful. http://www.tomshardware.com/2003/04/22/duel_of_the _titans/page18.html#database_test_mysql_3 [tomshardware.com]

Re:

I just want you to know...

That Page is over 4 years old..

*sigh*

Saveing are also higher

If add that ram for AMD severs costs less then FB-DIMMS.

Also FB-DIMMS and the intel chipset need a lot more power then amd chipsets and DDR2 ECC / DDR1 ECC ram.

Intel CPU is only 1 part that uses a lot of power

We see similar when we build systems.
The Intel CPUs are competitive with the Opterons on power consumption.
But: The whole system uses more with Intel.

Why? the northbridge memory controller is a separate chip with Intel, and it is very power hungry.
In the AMD chips the memory controller is a part of the CPU.
In the case of a similar dual XEON compared to a dual Opteron,
the XEON machine uses about 80W more power.

What a lot of these studies do not even get into is cooling cost.
for every watt of power , which ends up heat, we have to expend at least 1.5 watts, on air conditioning.

As for the comment about the size of the power supplies, that is irrelevant.
The maximum rated output of a supply has nothing to do with the power consumed.

Bottom line:
Assuming an Intel XEON server uses about 80 watts more than an equivalent AMD one,
which is what we see when we build them:
80w x 24 hours/day x 365 days is 700KWh. @ 9c/kWh costs $63/year.
Add aircon costs for that extra 80W:
120w x 24 hours/day x 365 days is 1050KWh. @ 9c/kWh costs $96/year.

Therefore, a machine using an extra 80W costs an extra $160 to run in an air conditioned room.

Source of power rates:
http://www.neo.ne.gov/statshtml/115.htm [ne.gov]

Re:Intel CPU is only 1 part that uses a lot of pow

Your idea is right, but your math is a little off.

You should be able to get down below 1.5 kW per Ton of A/C. (efficient systems can get down below 1.0 kW/T, even including all the pumps and fans)
That works out to close to 0.4 kW of A/C power used per 1.0

Re:

That translates to annual electricity savings between $20.29 per server and $36.04 per server

Over what time period? Or are they using the prime interest rate to figure out what the one-time savings are? :)

They're talking about those very special monthly an

energy $$$ savings yea right

AMD 8222 cpu = $2149
Intel 5160 cpu = $851

The AMD system will be obsolete before you realize any "cost savings".

Also you don't buy these top dog chips if you're going to let them sit idle all day.

Re:Sponsorship?

If you read the 6th paragraph of the article, AMD wasn't aware of this test, according to "John Fruehe, manager of worldwide market development for server/workstation products at AMD"

Re:

If you read the PDF, you'll see that the AMD system was tested with a 500W power supply while the Intel one was tested with a 600W one. I wonder how much of the different can be associated with that.

None. They would draw exactly the same power whether the

Re:Different Power Supplies

however differing power supplies do have different efficiencies of conversion. So we're really comparing top-to bottom solutions, and the processor may actually be a small part of the energy savings.

Re:Different Power Supplies

Actually the power capacity of a supply WILL affect the power measured as even same amount of power output will be on different part of the efficiency curves even if component losses are identical. Power supplies tends to be most efficient from 1/3 to 2/3 of their power ratings.

For a more scientific study, they should use the same power supply.

Re:Different Power Supplies

IANAEE but I found this thing [dell.com] (pdf) from DELL that has a "typical" efficiency curve (fig A, on the third page of the pdf, page # 64) that shows efficiency is pretty flat from 35% up to max load. Within maybe 5%.

Re:Different Power Supplies

I'd like to know why they compared a Woodcrest Xeon, circa June 2006 to the latest and greatest Opteron of today.

Re:

because they're trying to be biased without being too obvious...
-nB

Re:Different Power Supplies

The reason why is because the Woodcrest Xeon is the only 3GHz Xeon that Intel made, and for some reason they decided to standardize this test on "3.0 GHz". Since everybody knows that AMD outperforms Intel on a per-GHz basis, it does lead one to wonder why they chose that particular metric, but honestly no matter what metric they chose people would complain.

For one, AMD and Intel don't release their new chips on the same date, so one side can always complain "that's not our newest stuff" or "yeah, but just wait until our next generation". If you wait for same generation, same CPU frequency chips from both manufacturers before you do a benchmark, you're going to be waiting a while - it'll never happen. And if you pick a "performance class" to set your benchmark on, somebody will complain "yeah but XXXX's chip is .5GHz slower/faster than XXXX's". It's a lose-lose situation for the tester.

Also above there is a discussion about chipsets / power supplies / etc. Again nearly impossible to standardize on this stuff as well. Obviously there is no motherboard that is identical in every regard except the processor that it accepts. Another thread talks about the memory controller for Intel being off-chip vs. on-chip for AMD - so right there you have to go beyond the CPU and include more platform to make a "fair" comparison. Even if they standardized on a power supply, people can argue that the system that pulls less power doesn't need the larger power supply and could save more power (less loss to inefficiencies) on a smaller unit. So do you run the recommended unit for the server or run the same, possibly wrong power supply for both?

My overall point being that in for somebody to do any kind of test like this, they need to setup some base rules. I don't know why people complain so much - they provide all the criteria they chose and did a comparison based on that. If that doesn't answer a question you had, do it yourself or go to another benchmark. Don't complain that the test is invalid because your chip of choice didn't win. For this benchmark, power consumption for 3.0 GHz servers under "real world" conditions (not idle, not pinned, running various applications from databases to web servers), AMD won. Get over it.

Re:Different Power Supplies

600 is a 20% increase from 500, and 500 is a 17% decrease from 600. It's equally correct to say difference between 500 and 600 is 17% and 20%.