atoncelli
asked on
Compare UltraSPARC performance to Intel Dual Core
Hello,
I have 2 Solaris servers that I would like to upgrade. I am thinking to move to Linux and buy Dell servers with the new Intel Dual Core processor.
This is my issue. I don't know how to compare the speed and performance of these 2 servers vs. new Intel servers.
Is there any data that can help me make my decision easier?
The first server is qute old, the second one is newer and quite fast.
It would be nice to have at least an idea of what from the Intel/AMD world is comparable to these machines. In this way I can plan my purchuse based on getting the same performances or buying something better depending on the budget available.
In details...
My first server is:
System Configuration: Sun Microsystems sun4u Sun Enterprise 450 (4 X UltraSPARC-II 400MHz)
System clock frequency: 100 MHz
Memory size: 2048 Megabytes
========================= CPUs =========================
Run Ecache CPU CPU
Brd CPU Module MHz MB Impl. Mask
--- --- ------- ----- ------ ------ ----
SYS 0 0 400 4.0 US-II 9.0
SYS 1 1 400 4.0 US-II 9.0
SYS 2 2 400 4.0 US-II 9.0
SYS 3 3 400 4.0 US-II 9.0
2nd server:
System Configuration: Sun Microsystems sun4u Sun Fire V440
System clock frequency: 177 MHZ
Memory size: 16GB
==========================
E$ CPU CPU Temperature
CPU Freq Size Implementation Mask Die Amb. Status Location
--- -------- ---------- ------------------- ----- ---- ---- ------ --------
0 1593 MHz 1MB SUNW,UltraSPARC-IIIi 3.3 - - online -
1 1593 MHz 1MB SUNW,UltraSPARC-IIIi 3.3 - - online -
2 1593 MHz 1MB SUNW,UltraSPARC-IIIi 3.3 - - online -
3 1593 MHz 1MB SUNW,UltraSPARC-IIIi 3.3 - - online -
Thanks
I have 2 Solaris servers that I would like to upgrade. I am thinking to move to Linux and buy Dell servers with the new Intel Dual Core processor.
This is my issue. I don't know how to compare the speed and performance of these 2 servers vs. new Intel servers.
Is there any data that can help me make my decision easier?
The first server is qute old, the second one is newer and quite fast.
It would be nice to have at least an idea of what from the Intel/AMD world is comparable to these machines. In this way I can plan my purchuse based on getting the same performances or buying something better depending on the budget available.
In details...
My first server is:
System Configuration: Sun Microsystems sun4u Sun Enterprise 450 (4 X UltraSPARC-II 400MHz)
System clock frequency: 100 MHz
Memory size: 2048 Megabytes
========================= CPUs =========================
Run Ecache CPU CPU
Brd CPU Module MHz MB Impl. Mask
--- --- ------- ----- ------ ------ ----
SYS 0 0 400 4.0 US-II 9.0
SYS 1 1 400 4.0 US-II 9.0
SYS 2 2 400 4.0 US-II 9.0
SYS 3 3 400 4.0 US-II 9.0
2nd server:
System Configuration: Sun Microsystems sun4u Sun Fire V440
System clock frequency: 177 MHZ
Memory size: 16GB
==========================
E$ CPU CPU Temperature
CPU Freq Size Implementation Mask Die Amb. Status Location
--- -------- ---------- ------------------- ----- ---- ---- ------ --------
0 1593 MHz 1MB SUNW,UltraSPARC-IIIi 3.3 - - online -
1 1593 MHz 1MB SUNW,UltraSPARC-IIIi 3.3 - - online -
2 1593 MHz 1MB SUNW,UltraSPARC-IIIi 3.3 - - online -
3 1593 MHz 1MB SUNW,UltraSPARC-IIIi 3.3 - - online -
Thanks
Sun have a name for reliability, Dell doesn't. Wouldn't it be more useful to ask for a comparison of a Sun Sparc based box and a Sun Intel based box? Or even a Sun Sparc box Vs an IBM or HP box.
ASKER
I just would like to contemplate the idea to move from a Solaris to a Linux machine. Dell was just annexample of a platform with INTEL/AMD technology.
What ever can be found to compare a SPARC to an x86 computer would be fine for me.
Thanks
What ever can be found to compare a SPARC to an x86 computer would be fine for me.
Thanks
Well first thing is what are you using them for? Web servers that serve up static pages have different performance needs from Web servers that are doing dynamic HTML, which have different needs from database servers.
Sun sells AMD based servers.
Sun sells AMD based servers.
ASKER
The servers are used for source control. They run Telelogic software (Informix database).
ASKER
Just to clarify, the information I am looking for is benchmarks that can compare Solaris system (with UltraSPARC) to intel systems (Dell, HP ot others).
Thanks
Thanks
You may want to look at some of of the TPC benchmarks. They do not directly compare Intel to Sparc, but if you look at some of the test you can find different configurations for each.
http://www.tpc.org/tpch/results/tpch_results.asp?orderby=hardware
This is a list of the TPC-H by hardware vendor by database size. Of couse the first thing that you should do (hopefully you have) is to make sure the software you are running will run under Linux on x86.
Generally newer Intel will perform better than older Sparc, but newer Sparc will also perform better than older Sparc. I could not find any new comparsisions, all were from the late 90's comparing Intel's running between 450-1000 Mhz to Sparc running at 100-200 Mhz and the Sparcs performed about 2-4.
As Intel/AMD are now int the 3GHz range for some functions they may equal today's fastest Sparcs.
http://www.tpc.org/tpch/results/tpch_results.asp?orderby=hardware
This is a list of the TPC-H by hardware vendor by database size. Of couse the first thing that you should do (hopefully you have) is to make sure the software you are running will run under Linux on x86.
Generally newer Intel will perform better than older Sparc, but newer Sparc will also perform better than older Sparc. I could not find any new comparsisions, all were from the late 90's comparing Intel's running between 450-1000 Mhz to Sparc running at 100-200 Mhz and the Sparcs performed about 2-4.
As Intel/AMD are now int the 3GHz range for some functions they may equal today's fastest Sparcs.
ASKER
I reviewed some that data but it's not very friendly.
Is there a free tool that can be run on the systems to get some CPU/Memory/HD benchmark to have some basic data I can use to compare the computers?
Is there a free tool that can be run on the systems to get some CPU/Memory/HD benchmark to have some basic data I can use to compare the computers?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
I am not looking to upgrade necessarely. Same performance would be fine. I have to change computer though because the new version of Telelogic doesn't support Solaris 9 which is what I have at the moment. Therefore, while I am purchusing a new machine ( I have to leave the other one on line during the upgrade) I was thinking to also investigate the possibility to move to Linux. Some Telelogic users found they were able to get better performance and save money.
Upgrading to Solaris 10 is not an option I guess?
I would strongly suggest that you be very carefull about how a change of an enviroment "saves" money. My background is in IBM mainframe enviroments. What I call small, but most would call "large" (1500-2000 mainframe MIPS). I call is small because we are a service originazation for envirments that have 5000-10000 mainframe MIPS. It is amazing how many shops "saved money" by migrating off the mainframe. They save money in hardware and in SOME instances software. However their support costs go through the roof. They don't even count their enviromental costs. The largest mainframe today fully loaded only draws about 18 Kw. The same is about 5 fully loaded blade servers from most vendors. The problem is to replace the largest fully loaded mainframe take about 50 fully loaded blade servers.
So you need to watch what you mean by save money. Sure Linux can be less expensive that Solaris. However, what happens when you find a bug in Linux? Solaris, you call up Sun and they fix it. Linux, you can try and fix it yourself or report it to the Open Source commuity, that then needs to fix if, but it may take them longer than Sun takes. It might not. but you never know.
I would strongly suggest that you be very carefull about how a change of an enviroment "saves" money. My background is in IBM mainframe enviroments. What I call small, but most would call "large" (1500-2000 mainframe MIPS). I call is small because we are a service originazation for envirments that have 5000-10000 mainframe MIPS. It is amazing how many shops "saved money" by migrating off the mainframe. They save money in hardware and in SOME instances software. However their support costs go through the roof. They don't even count their enviromental costs. The largest mainframe today fully loaded only draws about 18 Kw. The same is about 5 fully loaded blade servers from most vendors. The problem is to replace the largest fully loaded mainframe take about 50 fully loaded blade servers.
So you need to watch what you mean by save money. Sure Linux can be less expensive that Solaris. However, what happens when you find a bug in Linux? Solaris, you call up Sun and they fix it. Linux, you can try and fix it yourself or report it to the Open Source commuity, that then needs to fix if, but it may take them longer than Sun takes. It might not. but you never know.