Thursday Mar 27, 2008
Excellent performance with open-source software! Some of you may have missed this, so I'll post the summary here (see Vanga's blog below for more info). Sun's MySQL running on a Sun SPARC Enterprise T5220 server supports 700 concurrent SugarCRM users using the SugarCRM workload. SugarCRM is a popular open-source CRM application.
On this benchmark Sun consolidated both the open-source MySQL Database
and the open-source SugarCRM Application on the same UltraSPARC T2.
In addition the UltraSPARC T2 hardware design is also open-sourced.
This open-source combination of SugarCRM and MySQL dramatically reduces
the price of deploying CRM on Sun SPARC Enterprise T5220 server, and
industry-leading free and open Solaris 10 OS.
This open-source combination of SugarCRM and MySQL supports 700 concurrent
SugarCRM users on Sun's SPARC Enterprise T5220 server on top of Solaris 10
consumed only 260 watts in a 2U rack space (16GB memory)
See Also:
SugarCRM users (bigger is better)
| System |
Processors |
Users |
Database |
| Type |
GHz |
Chips |
Cores |
Threads |
| Sun T5220 |
US-T2 |
1.4 |
1 |
8 |
64 |
700 |
MySQL |
SugarCRM does not allow vendors to compare to other vendors in public, you'll have to do you own digging 
.
Benchmark Description
SugarCRM is a popular open source CRM application, that runs on Apache, MySQL
and PHP also known as AMP stack. There are close to 1200 customers and the
community is growing much faster. SugarCRM supports only MySQL and Oracle DB.
Results Summary
| Certified Results |
|
700 users |
| Reference Date: |
|
February 26, 2008 |
| Systems: |
|
1 x Sun SPARC Enterprise T5220 16GB
|
| Total Number Processors: |
|
1 |
| Processor/GHz of Server: |
|
UltraSPARC T2 1.4 GHz |
| Operating System: |
|
Solaris 10 8/07 |
| Software: |
|
SugarCRM v4.5.1i
|
Wednesday Oct 24, 2007
Packaged water seems to abound in the states and given that tap water is generally good here, it makes you think about the potential for over-spending and waste given
40% of bottled US water is filtered tap.
Ira Flatow had a blog posting about a week ago talking about "Bottled water waste"
Here are some tidbits (see http://www.sciencefriday.com/blog/index.php?archives/235-Bottled-Water-Waste.html for more):
- ...Bottled water costs as much as $10 per gallon vs. penny per gallon of tap
- ...3 liters of water to produce each liter of bottled water
- ...in the US, less than 20% of these bottles are recycled
To find out the carbon footprint of your packaged water you'll have to go
to his blog... SHOCKING!
Friday Sep 28, 2007
Large memory is used on many of the standard benchmarks on systems using Xeon, Opteron,
and IBM Power. Actually this is fine, Sun systems can and do use large memory as well.
However, given the fact that memory can draw much more power than CPUs and that every system implementation is different (memory types: DDR2, FB-Dimms, very different ASICs) we really need power to be measured at these memory configurations. I have seen to many wattages only specifying TDP of a single processor, and far too many system wattage claims based on 2GB or 4GB of low-speed memory. We need the vendors to specify measured system wattages on configurations that they benchmark.
Price. Another common thing done by many vendors is to show the online prices for slow GHz and small memory configurations as their starting price. If you look at getting the latest GHz CPU and memory sizes like listed below, I've seen prices explode to 4 times to 10 times more than their starting price! Dear.
This suggests we should really list systems by memory size first and then processor count.
Processor count is mattering less than memory size for watts & price.
SPECweb2005:
64GB: HP DL580 G5 (4 quad-core Xeon 2.933GHz)
64GB: HP DL585 G2 (4 dual-core Opteron 3GHz)
32GB: HP DL380 G5 (2 quad-core Xeon 2.66GHz)
SAP-SD 2-Tier ECC6.0:
48GB: HP rx6600 (4 dual-core Itanium2 1.6GHz)
32GB: HP DL380 (2 quad-core Xeon 3GHz)
32GB: HP BL685c (4 dual-core Opteron 3GHz)
32GB: HP BL460c (2 quad-core Xeon 3GHz)
32GB: IBM p570 (2 dual-core Power6 4.7GHz)
SPECjbb2005:
32GB: Dell PE6950 (4 dual-core Opteron 2.8GHz)
SPECjAppServer2004:
32GB: HP rx6600 (4 dual-core Itanium2 1.6GHz)
Lotus:
32GB: IBM p550Q (2 quad-core Power5+ 1.5GHz)
Disclosure Statement (all data as of Sept 27, 2007)
SPEC, SPECjbb, SPECjAppServer, SPECweb, reg tm of
Standard Performance Evaluation Corporation, Info on www.spec.org.
NotesBench Domino[R] R6iNotes, more info www.notesbench.org.
Two-tier SAP Standard Sales and Distribution (SD) standard
SAP ERP 2004/2005 application benchmark: SAP, R/3, mySAP reg TM of
SAP AG in Germany and other countries. More info www.sap.com/benchmark.
Tuesday Sep 25, 2007
The BBC reports that "Large majorities in many countries now believe human activity is causing global warming, a BBC World Service poll suggests."
It goes on to say, "A sizeable majority of people agreed that major steps needed to be taken soon to address global warming."
More at: http://news.bbc.co.uk/1/hi/world/7010522.stm
Friday Sep 14, 2007
I know AMD & Intel love to focus on the power used by CPUs
in their high-stakes battle to gain server chip dominance. Both
started talking TDP (Thermal Design Power) and getting people to judge
systems based on TDP.
...wait a minute, buying a system based on the power of a CPU is
a bit of misdirection unless the CPU is most of the power. This isn't
the case any more. So CPU TDP should be ignored, unless you are
designing your own product and are just buying CPUs.
First of all system power is what datacenters care about, so Intel and
AMD should be talking about system power in realistic memory configurations.
Memory draws lots of power these days.
Second,TDP was created so the manufactures of servers manufacturers
would know much power the chip consumes in worst-case maximum-power cases
so they could design power supplies, cooling, etc. That just isn't
useful to datacenter managers.
AMD's marketing only slightly improved the situation by telling
customers of SYSTEMS to look at the processor's ACP (Average CPU
Power).
Two problems:
- Focus on CPU power to avoid talking system power, but system power is
what one needs to know, in average case to estimate electrical bills.
- Focus on average CPU power not server maximums that datacenters need
to design cooling on (see: http://blogs.sun.com/bmseer/entry/watts_a_matter_with_their.
ACP of a CPU ...hmmm, do you know what a pain it is to just measure a CPU. AMD
in their whitepaper , has to isolate the power consumed by the processor
and that consumed by the motherboard -- this requires motherboard modifications and special instrumented server platforms!.
way to much work to get a marketing message, all we want is server power on a variety of memory configuations and full-speed CPUs and actually running at good datacenter utilizations!
WARNING: Everyone loves to talk performance of high-GHZ CPUs and low-power of low-GHz CPUs, so watch for the confusing marketing and much worse "bait&switch." Also watts per core is useless marketing, it is the watt/perf for a system that counts. Also any vendor trying to sell power-efficiency on high-performance
systems should report watt/performance along with their world record performance on that system.
Thursday Sep 13, 2007
The register really needs to start asking tougher questions about power.
A new article "Researchers: AMD less power-hungry than Intel"
by Austin Modine.
This article just takes the results and ignores the really tough
questions facing datacenters and power. I'll help here...
- What is the power draw of large memory configurations 16GB, 32GB, 64GB.
The 2GB to 8GB used in this report are tiny for those making decisions in many datacenters.
- If you have racks and racks of idle servers (even servers at
30% utilisation) you are major problems and should be counting on
new servers to save power!!!
- It is Watt/Performance (like $/perf) not perf/watt, as this highlights
efficiency.
- CPU utilisation needs to be reported, if you are measuring at
less than 60% you are wasting 2 TIMES to 5 TIMES more power per unit of work!
- Everyone needs to realize the variability of power measurements and not report differences as ##.#% - what? one simply doesn't have one-tenth of
1% reproducibility in power measurements of servers.
Focusing on small memory tries to focus on CPU power which is fine for chip vendors, but it is frankly no what customers I talk to care about.
Customers care about full system ("car
not piston performance").
The results show that under certain configurations and load levels, the Intel server was 2% to 12% per cent more power efficient. But in a majority of cases, the AMD server was 9% to 23% per cent more efficient.
These are tiny differences. Not enough testing was done to show if these
results are consistent (power measurement has big run-to-run variations due to
the complexity of most systems.
AMD server was 30% to 53% more power efficient. If accurate,
it's a noteworthy figure, considering many servers spend the most of their time waiting for work.
WHAT?!?!? Datacenters idle? this is insane, maybe in extremely poorly run datacenters.
Actually, if this the case, those people running the datacenter should find other jobs and
not run datacenters.
On the whole, NN&A's tests showed that Intel's power efficiency decreases as memory size increases. Conversely, AMD's power efficiency increases as the memory is upped.
Intel, of course, disputes the results.
"The report doesn't measure our latest Xeons, or quad cores," said Intel rep Nick Knupffer in an email. We have 2 GHz quad cores in the market at 50 watts, 12.5 per core!"
I'll agree with Intel's criticism about latest CPUs and QCs, but even in
the Intel statement they quickly talk about 2GHz QC and not the full
GHz ones. Everyone loves to talk performance of high-GHZ CPUs and low-power of low-GHz CPUs, so watch for the confusing marketing and much worse "bait&switch." Also watts per core is
useless marketing, it is the watt/perf for a system that counts.
NN&A's white paper http://www.worlds-fastest.com/d.pdf/wfw991.pdf
Also take a look at: http://blogs.sun.com/bmseer/entry/saving_the_planet_one_datacenter
If you want to see big power savings (not the small percentages talked about above!) take a look at:
http://blogs.sun.com/ValdisFilks/entry/another_win_for_ecological_computing
...it is far too late, enough writing for today.
Tuesday Sep 11, 2007
Huge reductions can be made in datacenters with existing technology by just changing a few datacenter practices & metrics!
Things that make a HUGE difference (factors = big percentages):
- Turn off un-used servers, figure out how to turn more off
- Drive up utilisation up to 60% utilisation on all servers (this can decrease power by a factor of THREE TIMES (even if you have the latest power throttling chip features!)
- If you need virtualisation software to drive up the utilization get it, but make sure it is very efficient no overhead virtualisation like Solaris zones.
Remember CPU overhead means your are burning watts, or at least measure the CPU overhead on your virtualisation software to pick the most efficient alternative.
- Put the datacenters utility bill in the IT department! Motivate people!
- ...spread the word, as this knowlege virally spreads larger and larger number of people get this happening in their datacenters!
Remember if you save watts at the server you save watts to cool it, and
also save the inefficiences of getting the electricity from the powerplant to the datacenter! http://blogs.sun.com/bmseer/entry/power_from_utility_to_servers
Things that make some Difference (percentages):
- datacenter layout (5-10%)
- More efficient airflow & air conditioning (maybe 10-20%)
- tune your application performance, the faster it goes the more efficiently it goes
What to measure:
server-watt/perf (just like $/perf), perf/watt is misleading!
judge servers by full configuration power utilisation
compare servers of the same memory size (more important that processor count)
IT budget improvements (HW, SW, & Utility bills).
more details at: http://blogs.sun.com/bmseer/entry/eco_actual_fancy_power_saving
...everything above this was above the line!
Below the line.
What NOT to measure:
- Do NOT judge which server to buy by the TDP of the processor --- it is the whole server st*pid
- Do NOT judge which server to buy by the in-lab measurements of the CPU-only --- it is the whole server st*pid
- perf/cpu-watt is misleading, use server-watt/perf (just like $/perf)
- Do NOT assume perf/watt is server metric, you must ask!
- Do NOT compare servers at 30% utilization, that is like judging which
commuter car to buy by looking at vendor stats that compare overloaded SUV uses less gas going up a 20% road grade?!? -- you need to look at the full-memory config/CPU MHz you are buying at the good utilization 60% or more that you should be running your datacenter at!
- Do NOT judge which server to buy by looking at performance of fastest GHz CPUs at full utilization & full memory and then judging energy efficiency of low-GHz CPUs at 30% utilization and small memory. All benchmarks should have measured server watt/perf on every SPEC & TPC benchmark.
I think the above hints are much more useful than the following story:
http://www.cnn.com/2007/TECH/biztech/09/08/tech.green.credentials.ap/
final note:
Now go home and buy some CFLs, and turn off your lights when not in use...
Monday Aug 27, 2007
Is there a "Revolution in Solar Hydrogen on the Horizon?"
At Penn State University, a research group headed by professor of electrical engineering Craig Grimes in the Materials Research Institute is "only a couple of problems away" from developing an inexpensive and easily scalable technique for water photoelectrolysis - the splitting of water into hydrogen and oxygen using light energy - that could help power the proposed hydrogen economy.
read more at:
http://www.mri.psu.edu/articles/Revolution/index.asp
Wednesday Aug 22, 2007
Disingenuous Dell vs. HP comparison? Dell has a power comparison with measured data, cool. But they try to compare three 16GB HP Opteron systems versus two Dell Xeon 32GB systems. Quite the easy way to it appear that Dell draws less power, by comparing a total of 64GB for HP and 48GB for Dell. Source:
http://www.dell.com/downloads/global/power/dell2socket_vs_hp4socket_vmware.pdf
- 32GB 4-chip Dual-core 2.8GHz Opteron = 730 watts
- 16GB 2-chip Quad-core 2.66GHz Xeon = 445 watts
Really memory is taking lots more than CPUs, I think memory size
is more important than chip count. Prove me wrong, with data please!
Regardless, we have a few more wattage datapoints.
If Dell compared systems with the same size memories I would not have used the term "disingenuous."
Wednesday Aug 22, 2007
I could hear the cries in the blogosphere as I posted
the entry earlier today... "these graphs are based on what?"
...actually there were based on a real system, but I can't share the data.
I searched and found the following public data(only 8 months old):
http://www.anandtech.com/printarticle.aspx?i=2892
they have graphs that show:
- workload1: two 2.6GHz Opt 2218 8GB DDR2 that go from
260watts(100% util) down to 165 watts (20% util)
- workload2: two 2.6GHz Opt 2218 8GB DDR2 that go from
275watts(100% util) down to 190 watts (20% util)
- workload1: two 3GHz Xeon 5150 8GB FB-DIMMS that go from
325watts(100% util) down to 250 watts (20% util)
- workload2: two 3GHz Xeon 5150 8GB FB-DIMMS that go from
325watts(100% util) down to 270 watts (20% util)
...Not hard to imagine going from 400watts with the latest power
savings to 200 watts at 0% util, is it?
They did make a mistake by graphing perf/watt instead of watt/perf.
looking at their graphs we can make a table of the watt/perf:
watt/performance workload1="DVD Store", 8GB servers
| %Util | 3 GHz Xeon
2chip woodcrest | 2.6GHz Opteron
2chip 2218 |
| 20% | 37 watt/perf | 33 watt/perf |
| 40% | 22 watt/perf | 18 watt/perf |
| 60% | 14 watt/perf | 13 watt/perf |
| 80% | 11 watt/perf | 11 watt/perf |
| 100% | 9 watt/perf | 11 watt/perf |
Not a huge difference between systems, but a huge difference between
the same system running at 20% util vs. 60% util for example. In
either case it is 2.6x more wasteful running at 20% util vs. 60% util.
That means you use 2.6x more watts!!!
This data is actually worse than what I showed.
Get eco smart! Sure, turn on the power saving features. But if you
really want save $ and be a hero you must
put effort into driving up utilisation.
I've been talking about single servers, but this is not about going one by one through all of your servers, any datacenter has RACKS and RACKS of servers, which should make it easier to drive up the utilisation en masse.
Wednesday Aug 22, 2007
You will be mislead if you look at graphs of watts that do not show
"zero watts" to "maximum watts". The graph below
makes it appear that you start at max watts and go down all the
way as your utilisation decreases.
Below is a graph that is more like what modern servers can do with
the latest power saving features.
Poor graphs to make changes look bigger than they are.
Finding where to save money actually requires a different graph. One
needs to look at "watts per unit of work" which equates to
"watts per performance". Looking at the graph below it now becomes
quit quite clear how to save watts, you need to have your server at over
50% utilization even if you have the latest power-saving technology.
This was discussed yesterday in this blog.
You can see in this graph that five servers running at 10% utilisation uses 1200 watts/unit-of-work versus one server running at 50%
utilisation only uses 400 watts/unit-of-work. The 10% case requires
3 times more power to do the same amount of work!
Get eco smart, drive up your utilisation through good policy, or the
correct use of consolidation and/or virtualisation.
Tuesday Aug 21, 2007
Truly being Eco is about looking at the right things and highlighting
them in a way so you can make real judgments that really make a difference.
A fellow Sun employee gave me a great analogy, if you want to save
power switch off the light! Obvious. But too often we try to
use fancy technology to save energy when proper policy is more important.
If you leave incandescent lights on all day, switching to CFL
(compact florescent lights = better technology) can save energy.
but if you "switch it off during the daylight hours" (better policy) you
can save a lot more.
How does this apply to datacenters? Everyone needs to change
their policies to operate servers at not an insane 20-30% utilisation
but to 50% or more.
There are all kinds of virtualisation technologies that can be used
if for some reason you can't simply drive up the load. Take two
servers at 25% and consolidate them on to one server running at 50%
utilization and "turn the other one off".
There are lots of virtualisation technologies VMware, Xen, Solaris 10 Zones,... Another thing to consider is that Solaris 10 zones have no overhead for CPU, IO, and network. Overheads reduced performance but they also require one to burn more CPU watts. But use whatever technology suits your needs.
...more on using fancy datacenter power saving technology tomorrow...
Monday Apr 09, 2007
Just changed to the Eco theme. It should be a lot easier to read
how Sun provides leading performance now. See the previous posting today and last week.
The only thing is I don't know is why the table font is increased beyond the
size of the body text. I'll figure it out, or someone can post a comment.
|
|
|
|
To get good quality water (e.g. for coffee/tea) he...
Speaking of bottled water - try spelling evian bac...