From: Rodrigo Barbosa rodrigob@suespammers.org
Btw, don't quote me on this one :) I'm only 90% sure of the hotswapping capabilities, and less than 50% sure about the price :)
There _are_ systems with hot-swap CPUs, memory and/or, PCI[-X] slots. They are _not_ commodity and pricey, and require OS-level support.
In fact, I believe Linux 2.6 has some support for hot-swap -- at least at the PCI[-X] level, if not memory and possibly CPU as well. It requires extensive support though for that type of "dynamic" flexibility.
-- Bryan J. Smith mailto:b.j.smith@ieee.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Wed, Jun 29, 2005 at 12:59:24PM -0500, Bryan J. Smith b.j.smith@ieee.org wrote:
From: Rodrigo Barbosa rodrigob@suespammers.org
Btw, don't quote me on this one :) I'm only 90% sure of the hotswapping capabilities, and less than 50% sure about the price :)
There _are_ systems with hot-swap CPUs, memory and/or, PCI[-X] slots. They are _not_ commodity and pricey, and require OS-level support.
In fact, I believe Linux 2.6 has some support for hot-swap -- at least at the PCI[-X] level, if not memory and possibly CPU as well. It requires extensive support though for that type of "dynamic" flexibility.
Memory hotswap ? Sounds fishy to me, at least on IA32{,e} platforms. Are you sure about that ?
[]s
- -- Rodrigo Barbosa rodrigob@suespammers.org "Quid quid Latine dictum sit, altum viditur" "Be excellent to each other ..." - Bill & Ted (Wyld Stallyns)
Bryan J. Smith b.j.smith@ieee.org wrote:
From: Rodrigo Barbosa rodrigob@suespammers.org
Btw, don't quote me on this one :) I'm only 90% sure of the hotswapping capabilities, and less than 50% sure about the price :)
There _are_ systems with hot-swap CPUs, memory and/or, PCI[-X] slots. They are _not_ commodity and pricey, and require OS-level support.
Sun EXX00 + series running UltraSparc II and above have pretty much hot swapable everything: cpu, memory, disks etc.
And they are cheap on ebay these days (8-24 way 400 Mhz US-II (64-bit), with anywhere from 2-14 GB memory and FC-AL IO etc. EX500).
On Wednesday 29 June 2005 23:23, Bruno Delbono wrote:
Bryan J. Smith b.j.smith@ieee.org wrote:
From: Rodrigo Barbosa rodrigob@suespammers.org
Btw, don't quote me on this one :) I'm only 90% sure of the hotswapping capabilities, and less than 50% sure about the price :)
There _are_ systems with hot-swap CPUs, memory and/or, PCI[-X] slots. They are _not_ commodity and pricey, and require OS-level support.
Sun EXX00 + series running UltraSparc II and above have pretty much hot swapable everything: cpu, memory, disks etc.
And they are cheap on ebay these days (8-24 way 400 Mhz US-II (64-bit), with anywhere from 2-14 GB memory and FC-AL IO etc. EX500).
If you need any kind of performance and reliability, stay away from those though. First, a E4500 with 8CPUs running Solaris compiles the tree for one of the products we work with in about an hour and a half. A single P4 3.2Ghz running Linux takes about 12 minutes. As a webserver and so on the picture is similar. The US-I/II frames only shine if you go big (64cpus) or if you have an app that takes a huge amount of memory. Anything thats IO isn't terrible but any PCIe device will blow it away. And reliability... figure in that the box is a few years old... Its just not the same as new... Besides that, a component failure will bring those designs down and they have never been very reliable in coming back up without you removing the failed component...
The redundancy and so that the US-I/II designs offer don't make up for the age of the system.
Peter.
On Wed, 2005-06-29 at 23:38 -0400, Peter Arremann wrote:
If you need any kind of performance and reliability, stay away from those though. First, a E4500 with 8CPUs running Solaris compiles the tree for one of the products we work with in about an hour and a half. A single P4 3.2Ghz running Linux takes about 12 minutes.
Build time is _never_ a good reference of server benchmark.
As a webserver and so on the picture is similar.
Some of the computational is relative. It all depends on the web application.
The US-I/II frames only shine if you go big (64cpus) or if you have an app that takes a huge amount of memory. Anything thats IO isn't terrible but any PCIe device will blow it away.
??? I would beg to differ. Especially versus most "commodity" PC servers.
If you have a set of applications or services that are well distributed, but require I/O or scale linearly in a NUMA platform, then the interconnect of a high-end EXX00 is better than many PC systems.
Of course, I would definitely spend my money on a 4-way HP DL585 instead a 8-16 way Sun EXX00, a 2-way HP DL365 or even Sun z2100 instead of a 4- way E450, etc...
And reliability... figure in that the box is a few years old... Its just not the same as new... Besides that, a component failure will bring those designs down and they have never been very reliable in coming back up without you removing the failed component... The redundancy and so that the US-I/II designs offer don't make up for the age of the system.
Quoting Peter Arremann loony@loonybin.org:
If you need any kind of performance and reliability, stay away from those though. First, a E4500 with 8CPUs running Solaris compiles the tree for one of the products we work with in about an hour and a half. A single P4 3.2Ghz running Linux takes about 12 minutes.
Not defending Sun, just to put things into perspective for folks not familiar with Sun hardware (since Peter forgot to mention couple of facts).
E4500 is discontinued for some time now. It is rather old machine, much older that the above mentioned P4. I doubt that Peter's intention was to compare speed of Sun's server in general with the speed of latest-and-greates P4 systems. I hope that he simply forgot to mention how much older that E4500 is than the P4 box.
The processors in that thing are 400MHz or slower (not sure if 450MHz was option for E4500). Plus they are UltraSPARC-II (which is an older Sun's CPU design). So yeah, you can buy them cheap on ebay. Just as well as you can buy P-II and P-III systems even cheaper. And you could probably get VAX 11/780 for free (as long as you promise to take good care of it). People that buy E4500 nowdays are either buying them for spare parts, or they simply want redundant box exactly the same as the old box they already have. Or people that simply need some Sparc hardware around, but don't want to spend money on current line of Sun's servers and don't care if the system is slow as long as it does the job (my guess is Peter's company falls into this category, from what he wrote, otherwise they would be running their Sparc builds on something more recent).
---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.
On Thu, 2005-06-30 at 09:24 -0500, alex@milivojevic.org wrote:
Not defending Sun, just to put things into perspective for folks not familiar with Sun hardware (since Peter forgot to mention couple of facts). ... cut ...
Not only that, but people forget that SPARC is not sold by just Sun. SPARC is an IEEE standard licensed under "fair and non-discriminatory" terms. The SPARC ISA and most architectural details are freely available.
The _majority_ of my Solaris/SPARC experience in more recent years has been on Fujitsu/HAL solutions. I not only prefer the Fujitsu/HAL products over Sun, I not only prefer Fujitsu's "openness" to support SPARC solutions with non-Fijitsu networking, storage, etc... attached (whereas Sun likes 100% Sun equipment from a support perspective), but Fujitsu has taken over packaging design and fabrication of SPARCs for Sun itself. I.e., Sun used to design their own SPARC packages and TI was their foundary, but Fujitsu has almost always designed their own SPARC modules/ daughtercards, and fabbed them themselves. Sun even sells Fujitsu PRIMEPOWER products on their site now.
sun4u (UltraSPARC) is planned through UltraSPARC V, with UltraSPARC III (and lightweight IIIi) and IV available today. The 1-2 way UltraSPARC IIIi is more comparable to the cost/design of the Pentium III and IV, as well as the Xeon. The UltraSPARC III and IV, and their platform, are _clearly_superior_ to even OEM proprietary NUMA Xeon implementations (you really have to start looking at proprietary IA-64/Itanium II to get a good comparison).
But the design of the Opteron is why Sun is moving forth with it's move. The typical UltraSPARC NUMA/SBUS server architecture really doesn't offer much over Opteron in a 2-8 way. And to be honest, Solaris is probably the most "mature" OS for Opteron 200/800 right now, given its experience on partial mesh interconnected systems. Linux will get there, but it takes time and experience on a platform like SPARC (Windows never well, not even thanx to SGI's past, although brief, donations).
Bruno Delbono wrote:
Bryan J. Smith b.j.smith@ieee.org wrote:
There _are_ systems with hot-swap CPUs, memory and/or, PCI[-X] slots. They are _not_ commodity and pricey, and require OS-level support.
Sun EXX00 + series running UltraSparc II and above have pretty much hot swapable everything: cpu, memory, disks etc.
And they are cheap on ebay these days (8-24 way 400 Mhz US-II (64-bit), with anywhere from 2-14 GB memory and FC-AL IO etc. EX500).
I'm really sorry to start this thread again but I found something very interesting I thought everyone should ^at least^ have a look at:
http://uadmin.blogspot.com/2005/06/4-dual-xeon-vs-e4500.html
This article takes into account a comparision of 4 dual xeon vs. e4500. The author (not me!) talks about "A Shootout Between Sun E4500 and a Linux Redhat3.0 AS Cluster Using Oracle10g [the cluster walks away limping]"
Warm Regards,
-Bruno
On Friday 08 July 2005 17:34, Bruno Delbono wrote:
I'm really sorry to start this thread again but I found something very interesting I thought everyone should ^at least^ have a look at:
http://uadmin.blogspot.com/2005/06/4-dual-xeon-vs-e4500.html
This article takes into account a comparision of 4 dual xeon vs. e4500. The author (not me!) talks about "A Shootout Between Sun E4500 and a Linux Redhat3.0 AS Cluster Using Oracle10g [the cluster walks away limping]"
Bruno, you can make benchmarks say whatever you want - and this one seems to be one of the worst distortions I've seen in a while.
I know from personal experience that this isn't even close. dual xeon 2.0Ghz/4GB is about the same performance as a quad 450Mhz E4500/4GB in everything we're doing...
The TPC-C benchmarks (only official, impartial benchmarks I could find) are even worse: dual 3.6Ghz scored 63464 14 way 464Mhz E4500 scored 67103
Oracle has a word doc on their website talking about performance of a E4500 vs a Dell box (http://download-west.oracle.com/owsf_2003/Oracleworld2003.doc) where a quad xeon beats a quad E4500 by about 60% - and that was a quad 700Mhz P3 based Xeon...
Go search google for Xeon and E4500 and you'll see tons more of these benchmarks - and they all tell the same story...
Peter.
On Fri, 2005-07-08 at 19:33 -0400, Peter Arremann wrote:
Bruno, you can make benchmarks say whatever you want - and this one seems to be one of the worst distortions I've seen in a while. I know from personal experience that this isn't even close. dual xeon 2.0Ghz/4GB is about the same performance as a quad 450Mhz E4500/4GB in everything we're doing...
For what? That's the question.
The TPC-C benchmarks (only official, impartial benchmarks I could find) are even worse: dual 3.6Ghz scored 63464 14 way 464Mhz E4500 scored 67103
But there are over a _dozen_ TPC-C benchmarks -- some scale linearly on clusters, other prefer shared memory systems, some really taxi interconnect and do much better on NUMA+"true systems" interconnects.
The more you go the latter, the more P4 MCH "sucks." ;->
Oracle has a word doc on their website talking about performance of a E4500 vs a Dell box (http://download-west.oracle.com/owsf_2003/Oracleworld2003.doc) where a quad xeon beats a quad E4500 by about 60% - and that was a quad 700Mhz P3 based Xeon...
Which has the ALU/FPU equivalent of a 1.1-1.5GHz P4. Now the MCH of a P4 is certainly better, but still not a NUMA/"true system" interconnect. Which is why P3/P4 does _not_ scale well beyond 2 CPUs -- heck, the P4 really no better than P3, only some more throughput, but no less contention (actually more in many cases).
Which is why I still recommend dual-P3 servers today, especially refurbs, for the cost.
But here's the kicker ... "Simultaneously running 8 queries"
At that point, I'm _not_ taxing the interconnect at all. So there's no advantage to the NUMA/UPA architecture.
Go search google for Xeon and E4500 and you'll see tons more of these benchmarks - and they all tell the same story...
Of course, because they don't do a _full_ suite.
Compaq-Microsoft came out with a benchmark of the full TPC-C suite awhile back showing how Windows clusters beat a Sun shared memory system. What they didn't focus on, unless you read the entire article, is how the PC got _roasted_ -- by up to 10x -- on 4 of the 15 tests because they really stressed the interconnect, and the contention of memory access of threads.
looks like the 4500 came second to me by a fair margin, between 3x and 5x faster best case, not what Iwould call limping away.
Anyway that's enough of this thread, kill it now.....
Bruno Delbono wrote:
Bruno Delbono wrote:
Bryan J. Smith b.j.smith@ieee.org wrote:
There _are_ systems with hot-swap CPUs, memory and/or, PCI[-X] slots. They are _not_ commodity and pricey, and require OS-level support.
Sun EXX00 + series running UltraSparc II and above have pretty much hot swapable everything: cpu, memory, disks etc.
And they are cheap on ebay these days (8-24 way 400 Mhz US-II (64-bit), with anywhere from 2-14 GB memory and FC-AL IO etc. EX500).
I'm really sorry to start this thread again but I found something very interesting I thought everyone should ^at least^ have a look at:
http://uadmin.blogspot.com/2005/06/4-dual-xeon-vs-e4500.html
This article takes into account a comparision of 4 dual xeon vs. e4500. The author (not me!) talks about "A Shootout Between Sun E4500 and a Linux Redhat3.0 AS Cluster Using Oracle10g [the cluster walks away limping]"
Warm Regards,
-Bruno _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-
alex＠milivojevic.org -
Bruno Delbono -
Bryan J. Smith -
Bryan J. Smith <b.j.smith＠ieee.org> -
Peter Arremann -
Peter Farrow -
Rodrigo Barbosa