that 3U Xserve would only be accepted if it was 8 way though.
If they had them I would be in the market for a couple They will produce WAY too much heat. I just got a few Dell 1750's in with Dual Xeon 2.8's and they produce WAY more heat then the Dell 2600's and 2650's I have with dual Xeon 3.06's. I will have to post a picture tomorrow. I just got them in the rack last night. BTW, this is a database cluster and I am testing a dual G5 2.0 to see how well it performs in comparison to the other machines.
People underestimate the heat a 1U produces when you have high-end processors.
The size of the case does not impact heat production!!!!!!!!!
Dave
Quote:
Originally posted by Mike
If they had them I would be in the market for a couple They will produce WAY too much heat. I just got a few Dell 1750's in with Dual Xeon 2.8's and they produce WAY more heat then the Dell 2600's and 2650's I have with dual Xeon 3.06's. I will have to post a picture tomorrow. I just got them in the rack last night. BTW, this is a database cluster and I am testing a dual G5 2.0 to see how well it performs in comparison to the other machines.
People underestimate the heat a 1U produces when you have high-end processors.
"People underestimate the heat a 1U produces when you have high-end processors."
It may feel like the unit is dissipating more heat, but they only thing really happening is that the heat is removed through a smaller vent, usually at a high velocity. given the same chip set and operating speed the amount of heat given off remains the same in all chassis.
Quote:
Originally posted by Mike
I never said it did...however, it does impact heat dissipation...but I am sure you knew that
It may feel like the unit is dissipating more heat, but they only thing really happening is that the heat is removed through a smaller vent, usually at a high velocity. given the same chip set and operating speed the amount of heat given off remains the same in all chassis.
You are wrong. By your logic the exact same heater would keep a 100' x 100' room the same temperature as a 5' x 5' room.
Can't speak for the others, but my company switched to an Xserve for hosting static and php/SQL dynamic pages from a linux box. It's been phenominal. The remote management tools provided by Apple are far superior than Webmin or anything else we've used on the Linux side. Of course, you can always use Webmin on the Xserve as well. We've only had to restart it once in a year's time and it paged us in advance to let us know...super cool.
Can't speak for the others, but my company switched to an Xserve for hosting static and php/SQL dynamic pages from a linux box. It's been phenominal. The remote management tools provided by Apple are far superior than Webmin or anything else we've used on the Linux side. Of course, you can always use Webmin on the Xserve as well. We've only had to restart it once in a year's time and it paged us in advance to let us know...super cool.
I am interested and impressed in Apple's administrative functionality. However, Dell has much of this as well if you need to use it. However, we do not install anything on our servers that we do not need. There is no xwindows or any gui...just our custom builds of our own software. We recently tested RedHat Ent AS to see how it compared to our custom builds...we were VERY disappointed in the performance of the RedHat build.
Basically, we have our Linux machines setup to monitor themselves and if there is something wrong they take themselves offline (or they are re-booted by another machine).
I am EXTREMELY interested to see where Apple is going in the next couple months with their server hardware. I did hear something that I can not share with anyone yet as to the direction of Apple server hardware in the near future Something that will remove the primary hesitation of anyone looking at buying an Apple server that is running X86 and Linux.
Have you been happy with the servers? How about the storage array? Are you running a switched fiber array?
We're quite happy with them (we have two more than you see). The Xserve RAID is only connected to one of the servers which acts as an image server processing lots of images all day long.
No; I'm very much correct. Given your example, the same 100 watt heater generates equal amounts of heat in both rooms.
How the rooms deal with that generated heat is dependant on the design of the rooms not the heater. A PC enclosure is much the same the, temperature of the enclusure is very much dependant on the design of the enclosure. Look at it this way, we can now get a 970, running at 1.6GHz, in both G5 and blade forms; given equal software loads each processor will produce the same amount of heat. The difference is that the two machines have contrasting approaches to deal with the removal of that heat.
Thanks
Dave
Quote:
Originally posted by Mike
You are wrong. By your logic the exact same heater would keep a 100' x 100' room the same temperature as a 5' x 5' room.
No; I'm very much correct. Given your example, the same 100 watt heater generates equal amounts of heat in both rooms.
How the rooms deal with that generated heat is dependant on the design of the rooms not the heater. A PC enclosure is much the same the, temperature of the enclusure is very much dependant on the design of the enclosure. Look at it this way, we can now get a 970, running at 1.6GHz, in both G5 and blade forms; given equal software loads each processor will produce the same amount of heat. The difference is that the two machines have contrasting approaches to deal with the removal of that heat.
Thanks
Dave
You're still wrong! The issues was not the amount of heat a processor creates by itself...it was about the heat a MACHINE will create or dissipate.
Numbers do not lie. Our 1U Dell 1750's produce more heat then our 6U Dell 2600's and our 2U 2650's with the same CPU's in all three. Our tests back-up this information. I can't tell you much more...you can continue to think I am wrong...it really doesn't matter. The machines are not in your server room and you do not have to worry about cooling them...I do
During our tests we found that the 1750's will produce twice the heat of a 2600. We were actually shocked at the difference!
We're quite happy with them (we have two more than you see). The Xserve RAID is only connected to one of the servers which acts as an image server processing lots of images all day long.
Are you processing images for use on the web or print? We process over 2 million a day for use on the web We are using ImageMagic to process them along with jpegtran and PHP functions.
You have me really interested in how you came up with this conclusion. Was it by measuring AC power input into the units?
The machine will not create heat on its own, any heat dissipated is the result of the electrical loads inside the case. The only way you will have greater heat production, given equal hardware, is if you have more components generating heat. Now the exhaust air temperature may be higher but that is not heat that is temperature.
This is a very curious issue, if the 1U machines are truely producing that much extra heat it would indicate that some very ineffiecent parts where used in there construction.
Dave
Quote:
Originally posted by Mike
You're still wrong! The issues was not the amount of heat a processor creates by itself...it was about the heat a MACHINE will create or dissipate.
Numbers do not lie. Our 1U Dell 1750's produce more heat then our 6U Dell 2600's and our 2U 2650's with the same CPU's in all three. Our tests back-up this information. I can't tell you much more...you can continue to think I am wrong...it really doesn't matter. The machines are not in your server room and you do not have to worry about cooling them...I do
During our tests we found that the 1750's will produce twice the heat of a 2600. We were actually shocked at the difference!
You have me really interested in how you came up with this conclusion. Was it by measuring AC power input into the units?
The machine will not create heat on its own, any heat dissipated is the result of the electrical loads inside the case. The only way you will have greater heat production, given equal hardware, is if you have more components generating heat. Now the exhaust air temperature may be higher but that is not heat that is temperature.
This is a very curious issue, if the 1U machines are truely producing that much extra heat it would indicate that some very ineffiecent parts where used in there construction.
Dave
Actually, they (the machines) are not the same. The 1750's were single power supplies and single drives while every 2600 and 2650 had redundant power supplies, multiple drives and a RAID card. Otherwise, the machines had the same CPU's. We also did not measure the power going into the machines as we were only concerned about the heat generated not the power consumption.
Again, my statements towards the beginning of this thread were in relation to the heat generated by a 1U versus a MultipleU.
Comments
Originally posted by applenut
that 3U Xserve would only be accepted if it was 8 way though.
People underestimate the heat a 1U produces when you have high-end processors.
My poop is brown...and it is known to generate frequencies, while incubating, at over 29Ghz!
You gotta love the 'confirmed!!' trustworthiness.
-walloo.
Dave
Originally posted by Mike
People underestimate the heat a 1U produces when you have high-end processors.
Originally posted by wizard69
The size of the case does not impact heat production!!!!!!!!!
Dave
I never said it did...however, it does impact heat dissipation...but I am sure you knew that
(Even though each release makes my G5 more obsolete, I don't care. I care not about my computer, but Apple's state as a company)
Originally posted by Mike
Maybe we will have some G5's in this cabinet very soon
I would prefer to se them added to this:
"People underestimate the heat a 1U produces when you have high-end processors."
It may feel like the unit is dissipating more heat, but they only thing really happening is that the heat is removed through a smaller vent, usually at a high velocity. given the same chip set and operating speed the amount of heat given off remains the same in all chassis.
Originally posted by Mike
I never said it did...however, it does impact heat dissipation...but I am sure you knew that
Originally posted by wizard69
Sorry I must have misunderstood this line:
It may feel like the unit is dissipating more heat, but they only thing really happening is that the heat is removed through a smaller vent, usually at a high velocity. given the same chip set and operating speed the amount of heat given off remains the same in all chassis.
You are wrong. By your logic the exact same heater would keep a 100' x 100' room the same temperature as a 5' x 5' room.
Originally posted by JLL
I would prefer to se them added to this:
Have you been happy with the servers? How about the storage array? Are you running a switched fiber array?
Originally posted by shawk
Pixar wants them.
Yeah right...
Since when does Apple make Blades?
Originally posted by Michael Wilkie
Can't speak for the others, but my company switched to an Xserve for hosting static and php/SQL dynamic pages from a linux box. It's been phenominal. The remote management tools provided by Apple are far superior than Webmin or anything else we've used on the Linux side. Of course, you can always use Webmin on the Xserve as well. We've only had to restart it once in a year's time and it paged us in advance to let us know...super cool.
I am interested and impressed in Apple's administrative functionality. However, Dell has much of this as well if you need to use it. However, we do not install anything on our servers that we do not need. There is no xwindows or any gui...just our custom builds of our own software. We recently tested RedHat Ent AS to see how it compared to our custom builds...we were VERY disappointed in the performance of the RedHat build.
Basically, we have our Linux machines setup to monitor themselves and if there is something wrong they take themselves offline (or they are re-booted by another machine).
I am EXTREMELY interested to see where Apple is going in the next couple months with their server hardware. I did hear something that I can not share with anyone yet as to the direction of Apple server hardware in the near future
Originally posted by Mike
Have you been happy with the servers? How about the storage array? Are you running a switched fiber array?
We're quite happy with them (we have two more than you see). The Xserve RAID is only connected to one of the servers which acts as an image server processing lots of images all day long.
How the rooms deal with that generated heat is dependant on the design of the rooms not the heater. A PC enclosure is much the same the, temperature of the enclusure is very much dependant on the design of the enclosure. Look at it this way, we can now get a 970, running at 1.6GHz, in both G5 and blade forms; given equal software loads each processor will produce the same amount of heat. The difference is that the two machines have contrasting approaches to deal with the removal of that heat.
Thanks
Dave
Originally posted by Mike
You are wrong. By your logic the exact same heater would keep a 100' x 100' room the same temperature as a 5' x 5' room.
Originally posted by wizard69
No; I'm very much correct. Given your example, the same 100 watt heater generates equal amounts of heat in both rooms.
How the rooms deal with that generated heat is dependant on the design of the rooms not the heater. A PC enclosure is much the same the, temperature of the enclusure is very much dependant on the design of the enclosure. Look at it this way, we can now get a 970, running at 1.6GHz, in both G5 and blade forms; given equal software loads each processor will produce the same amount of heat. The difference is that the two machines have contrasting approaches to deal with the removal of that heat.
Thanks
Dave
You're still wrong! The issues was not the amount of heat a processor creates by itself...it was about the heat a MACHINE will create or dissipate.
Numbers do not lie. Our 1U Dell 1750's produce more heat then our 6U Dell 2600's and our 2U 2650's with the same CPU's in all three. Our tests back-up this information. I can't tell you much more...you can continue to think I am wrong...it really doesn't matter. The machines are not in your server room and you do not have to worry about cooling them...I do
During our tests we found that the 1750's will produce twice the heat of a 2600. We were actually shocked at the difference!
Originally posted by JLL
We're quite happy with them (we have two more than you see). The Xserve RAID is only connected to one of the servers which acts as an image server processing lots of images all day long.
Are you processing images for use on the web or print? We process over 2 million a day for use on the web
The machine will not create heat on its own, any heat dissipated is the result of the electrical loads inside the case. The only way you will have greater heat production, given equal hardware, is if you have more components generating heat. Now the exhaust air temperature may be higher but that is not heat that is temperature.
This is a very curious issue, if the 1U machines are truely producing that much extra heat it would indicate that some very ineffiecent parts where used in there construction.
Dave
Originally posted by Mike
You're still wrong! The issues was not the amount of heat a processor creates by itself...it was about the heat a MACHINE will create or dissipate.
Numbers do not lie. Our 1U Dell 1750's produce more heat then our 6U Dell 2600's and our 2U 2650's with the same CPU's in all three. Our tests back-up this information. I can't tell you much more...you can continue to think I am wrong...it really doesn't matter. The machines are not in your server room and you do not have to worry about cooling them...I do
During our tests we found that the 1750's will produce twice the heat of a 2600. We were actually shocked at the difference!
Originally posted by wizard69
You have me really interested in how you came up with this conclusion. Was it by measuring AC power input into the units?
The machine will not create heat on its own, any heat dissipated is the result of the electrical loads inside the case. The only way you will have greater heat production, given equal hardware, is if you have more components generating heat. Now the exhaust air temperature may be higher but that is not heat that is temperature.
This is a very curious issue, if the 1U machines are truely producing that much extra heat it would indicate that some very ineffiecent parts where used in there construction.
Dave
Actually, they (the machines) are not the same. The 1750's were single power supplies and single drives while every 2600 and 2650 had redundant power supplies, multiple drives and a RAID card. Otherwise, the machines had the same CPU's. We also did not measure the power going into the machines as we were only concerned about the heat generated not the power consumption.
Again, my statements towards the beginning of this thread were in relation to the heat generated by a 1U versus a MultipleU.