Tag Info

Hot answers tagged

30

That benchmark is only comparing the speed of the native os to a single guest os. Hardly a real world test. I don't think I would put much weight on it. Most of the KVM camp argues that Xen requires too many interrupts and hops between kernel and user space but from most of the more real world benchmarks that I've seen that hasn't really been realized and ...


30

If you have concurrency set at 1, there's no difference between those two. It begins to matter when you have more than 1 request performed simultaneously. Let's look at an example of what I get on my localhost: ab -c 1 -n 1000 http://localhost/ will give: Time taken for tests: 3.912 seconds Time per request: 3.912 [ms] (mean) Time per request: ...


19

The typical experience for a general purpose server workload on a bare metal\Type 1 Hypervisor is around 1-5% of CPU overhead and 5-10% Memory overhead, with some additional overhead that varies depending on overall IO load. That is pretty much consistent in my experience for modern Guest OS's running under VMware ESX\ESXi, Microsoft Hyper-V and Xen where ...


16

"Performance" has many aspects. The n00bs measure the boot time of an OS, and say e.g. Windows 2012 is sooooooo great because it boots in 12 sec on real HD, maybe 1 sec on SSD. But this sort of measure not very useful: performance is equal to OS boot time, but the OS boots once a month so optimizing that doesn't make much sense. Because it's my daily ...


16

Here's some results comparing all major linux FSes with bonnie++ that you can use as a starting point. In terms of random seeks Reiser wins, followed by EXT4, followed by JFS. I'm not sure if this will correlate exactly to directory lookups, but it seems like it would be an indicator. You'll have to do your own tests for that specifically. EXT2 beats the ...


15

You say in your comments that you're evaluating a netboot / network root environment. The first thing you must realize is there is no such thing as "vanilla" - you're not going to run CentOS 5.10 right out of the box with zero changes (if you think you are you're deluding yourself: NFS Root is already at least Strawberry, verging on Pistachio). If you want ...


11

The performance of your SPDY setup will obviously depend on the implementation of the stack and you should know the processing cost vs. a 'regular' HTTP or HTTPS session, but.. Keep in mind that that the primary point of SPDY is to reduce the latency and overhead of rendering a typical webpage (as seen by your visitors), not to optimize your server-side ...


9

The "SATA = 7.2K RPM, SAS = 10/15K RPM" mind-set is strong, and (in my opinion anyway) where most of the "SAS is faster than SATA" thinking comes from. There are some slight differences between SAS and SATA drives, notably in their on-board caching algorithms (NCQ vs. TCQ). However, the performance difference of equivalently specced hard-drives will be ...


9

IOPs is the difference your looking for in the "speed". The simple way to explain the difference is that SATA is half duplex and SAS is full duplex. SATA drives are dumb and have to communicate with the controller for operations. SAS drives are smart and only requests and returns use the bus. Depending on your usage case, spending more may not gain ...


8

Microsoft has a tool called SQLIO which will push the IO to the limit. Just build a Windows VM with a disk on the correct data store and run SQLIO. Don't let the name fool you, it doesn't actually require SQL to be installed (it doesn't have anything to do with SQL, it was just built by the SQL team).


8

Yes. But that is not the question. The difference is normally neglegible (1% to 5%).


8

Install one system, boot it and check out the block layer statistics from /sys/block/${DEV}/stat e.g. /sys/block/sda/stat. Quoting from the documentation: The stat file consists of a single line of text containing 11 decimal values separated by whitespace. The fields are summarized in the following table, and described in more detail below: Name ...


7

I agree with most of what Andrew said, except that I would recommend Reiser4 or the older (but better supported) ReiserFS. As those tests (and the documentation for ReiserFS) indicate, it is designed for precicesly the situation you are asking about (large numbers of small files or directories). I have used ReiserFS in the past with Gentoo and Ubuntu ...


7

A lot depends on which VT-x features are supported by the processors you are concerned with. The initial implementations didn't provide measurable benefits and in most cases performance was actually degraded. However the recent H/W virtualization features related to Second Level Address Translation (SLAT) implemented by AMD as RVI\NPT with the Shanghai ...


7

Creating a block level access to virtual machine state, as opposed to a file level access will always be faster because there is a layer of abstraction removed. I would recommend the LVM approach. Don't forget, you can always backup the LVM volume just like a file. There isn't much difference between the two. LVM is also quite flexible in terms of ...


6

I personally would choose virtualisation based on usability, support, reliability and suitability for the virtual machines you are using. Xen networking data transfer rates seem to be as good as real hardware but I've also had some battles with Xen and vLans and multiple ethernet cards. I have no experience of KVM, but I would also suggest that you ...


6

Here's the way I sometimes test ram: first mount two tmpfs (by default tmpfs is half the ram): # mount -t tmpfs /mnt/test1 /mnt/test1 # mount -t tmpfs /mnt/test2 /mnt/test2 Check free memory and free space: # free total used free shared buffers cached Mem: 252076 234760 17316 0 75856 ...


6

The most important benchmarks are the ones you take against your application before making some changes, then re-running them, on your own hardware in your own server environment. This won't help if you're trying to choose whether to use framework X versus framework Y for your application or configuration, unless you are in a position to test your servers ...


5

Take a look at bonnie and bonnie++, they will give you pretty much information. http://www.coker.com.au/bonnie++/


5

What is reasonable depends on what kind of traffic your server is getting. A typical web server that handles about 1 million page requests per day (and that's quite a busy little server) will rarely ever need more than 50 concurrent connections, unless your page processing is really heavy. If your average page processing time (i.e. the time it takes from ...


5

Comparing Linode's Dual Quad-core Xeon CPUs vs your dual Atom D510 is like comparing a Lamborghini with a mobility scooter. The Atoms have a tiny percentage of the processing power of modern CPUs. In fact, their performance is even poor compared to CPUs from 6 years ago. They are not built for grunt, but for small low-power devices such as "netbooks". ...


5

On at least Linux, all synthetic benchmarking answers should mention fio - it really is a swiss army knife I/O generator. A brief summary of its capabilities: It can generate I/O to devices or files Submitting I/O using a variety different methods Sync, psync, vsync Native/posix aio, mmap, splice Queuing I/O up to a specified depth Specifying the size ...


5

I don't know of a tool that does this directly, but you could always just use tcpdump on the bsd gateway to sniff DNS requests and responses, and then compare requests to responses. The tcp dump would be something like: tcpdump -i interface 'udp port 53' -o dumpfile It is possible they would be tcp too, so you can capture both if you want. You can then ...


4

The question is what do you mean by flops? If all you care about is how many of the simplest floating point operations per clock, it is probably 3x your clock speed, but that is about as meaningless as bogomips. Some floating point ops take a long time (divide, for starters), add and multiply are typically quick (one per fp unit per clock). The next issue is ...


4

You will find very interesting information on the subject in the following presentation : http://www.xen.org/files/xensummitboston08/Deshane-XenSummit08-Slides.pdf The person who did this is a Xen expert but the comparison seems pretty fair.


4

I'll just add to all the answers above, by reminding you that LVM has a somewhat easy to use snapshot mechanism. This makes it pretty easy to backup or clone running VM's by simply making a snapshot, cloning or backing the VM up, and removing the snapshot. All without downtime.


4

The EULA for VMware prohibits publishing benchmarks: http://blogs.vmware.com/virtualreality/2009/03/a-big-step-backwards-for-virtualization-benchmarking.html That's why you don't see too many independent benchmarks.


4

What are the best possible ways to benchmark RAM (no-ECC) under linux / arm? RamSpeed is the only multiplatform memory benchmark tool I'm aware of. You might be able to compile it for arm, if supported: http://alasir.com/software/ramspeed/ If it's not supported, you might be able to benchmark using stream: http://www.cs.virginia.edu/stream/ref.html ...


4

If your app is processing a large amount of data, try following the data's path - if the input data is fed from the network, check for possible latencies, bandwidth limitations or transmission errors. You already checked disk I/O which otherwise would be a likely candidate for a bottleneck. Last but not least, since it is a highly multithreaded .NET ...



Only top voted, non community-wiki answers of a minimum length are eligible