Donations for Shawns coffee pot...

[we_are_borg]

Yeah... Definitely uses less server resources.

Can you give a percentage?

 

[digitalpoint]

Nice, Shawn just be careful of heat output in such a small space as the higher end E5-2690 are pushing 130W per cpu 12x130w = 1,560w in 6U space

Yeah, had a talk with power supply engineer at SuperMicro yesterday because I was worried about exactly that. BTW it's 24 CPUs in 6U (not 12)..

Anyway, with the number of drives we are using, we are going to go with the CPU one step down... Still 8 core though. The E5-2680.

 

[digitalpoint]

Can you give a percentage?

Probably about half.

 

[we_are_borg]

Probably about half.

That's a lot is it all explainable why it runs smoother, say better programming or less features in the software etc etc.

Those servers look really nice but why not the once or there big brothers that you wanted to have one or two years back thought they where more sexy.

 

[p4guru]

Yeah, had a talk with power supply engineer at SuperMicro yesterday because I was worried about exactly that. BTW it's 24 CPUs in 6U (not 12)..

Anyway, with the number of drives we are using, we are going to go with the CPU one step down... Still 8 core though. The E5-2680.

oh 24 cpus - yikes 3,120 watts of heat - E5-2660 weighs in at 95W though so save 35W per cpu http://www.cpu-world.com/Compare/9/Intel_Xeon_E5-2660_vs_Intel_Xeon_E5-2680.html - 2,280 watts vs 3,120 watts

Either way will be interesting to see how that 6U cluster fairs

 

[digitalpoint]

Yeah, it's a lot of heat, but it's specced out to be fine up to 30 degrees C ambient temp (87 degrees F), which is fine in the data center they are going in.

 

[p4guru]

Well let us know how it goes. FYI, my Dual E5-2650, 32GB ddr3-1600 1.35v rated memory and 6x 256GB SSD rig at home pulls in around ~245-255 watts at the wall usually but using 93% efficiency based Antec power supply

 

[TheLaw]

Well let us know how it goes. FYI, my Dual E5-2650, 32GB ddr3-1600 1.35v rated memory and 6x 256GB SSD rig at home pulls in around ~245-255 watts at the wall usually but using 93% efficiency based Antec power supply

I don't envy your electric bill. I prefer to let someone else deal with the heat in the kitchen as well. Nice to see how much has progressed here. So much so that I was confused in following conversations and who you were talking to given some dramatic avatar changes, lol. I'm just about ready to be moving 3 sites off VB as well after a great deal of planning.

 

[p4guru]

I don't envy your electric bill. I prefer to let someone else deal with the heat in the kitchen as well. Nice to see how much has progressed here. So much so that I was confused in following conversations and who you were talking to given some dramatic avatar changes, lol. I'm just about ready to be moving 3 sites off VB as well after a great deal of planning.

well better that i am using 93% efficient power supply versus 80-85% efficient one

 

[Digital Doctor]

Those are some impressive start render times for a site with

Online now: 1,919 (members: 203, guests: 1,716)
Discussions:2,200,538
Messages:16,664,233
Members:669,400

http://www.webpagetest.org/result/130208_3M_GT5/

The waterfall diagrams are the smallest I have ever seen.

 

[Ditto]

You're running the test against a thread and not forum home landing page which gives a very different result, and what most would land on first. Why did you use a thread for the test?

http://www.webpagetest.org/result/130208_AW_HJQ/

 

[Digital Doctor]

Because when I visit forums I read threads, not forum lists

 

[Deebs]

Heh, that test is flawed. My first byte time is 0.562 yet it is rated an F. Go figure

 

[p4guru]

Heh, that test is flawed. My first byte time is 0.562 yet it is rated an F. Go figure

interpretation is flawed not the test.. DO NOT look at the score A-F but the actual figures themselves

 

[Deebs]

interpretation is flawed not the test.. DO NOT look at the score A-F but the actual figures themselves

The scoring is flawed. How can a lower time to first byte score worse than a higher one especially as the score is specific to that metric.

 

[RobParker]

Because when I visit forums I read threads, not forum lists

But isn't it likely that as you've already visited forumhome you'd have a lot of the stuff for the threadview already cached?

 

[digitalpoint]

I want it to be (much) faster... Our bottleneck is 7 year old servers.

 

[Slavik]

well better that i am using 93% efficient power supply versus 80-85% efficient one

Do you know under-utilising a PSU can make it more wasteful than using it to its max specification?

Eg if you have a 1000 watt Platinum rated PSU, but only 600 watts of power is being drawn from it, it will waste more power than using a bronze rated 650 watt psu.

 

[MattW]

I want it to be (much) faster... Our bottleneck is 7 year old servers.

How long before you get your super spec new ones?

 

[p4guru]

Do you know under-utilising a PSU can make it more wasteful than using it to its max specification?

Eg if you have a 1000 watt Platinum rated PSU, but only 600 watts of power is being drawn from it, it will waste more power than using a bronze rated 650 watt psu.

ah but can 600w psu handle a 7x GPU blender render setup