Re: [Engine-devel] CPU Overcommit Feature
- Original Message - From: Dennis Jacobfeuerborn denni...@conversis.de To: Doron Fediuck dfedi...@redhat.com Cc: engine-devel@ovirt.org, Andrew Cathrow acath...@redhat.com Sent: Wednesday, December 19, 2012 1:49:24 PM Subject: Re: [Engine-devel] CPU Overcommit Feature On 12/18/2012 07:33 PM, Doron Fediuck wrote: - Original Message - From: Dennis Jacobfeuerborn denni...@conversis.de To: Andrew Cathrow acath...@redhat.com Cc: engine-devel@ovirt.org Sent: Tuesday, December 18, 2012 7:59:26 PM Subject: Re: [Engine-devel] CPU Overcommit Feature On 12/18/2012 06:33 PM, Andrew Cathrow wrote: - Original Message - From: Dennis Jacobfeuerborn denni...@conversis.de To: engine-devel@ovirt.org Sent: Tuesday, December 18, 2012 12:30:34 PM Subject: Re: [Engine-devel] CPU Overcommit Feature On 12/17/2012 07:13 PM, Simon Grinberg wrote: - Original Message - From: Greg Padgett gpadg...@redhat.com To: engine-devel engine-devel@ovirt.org Sent: Monday, December 17, 2012 4:37:57 PM Subject: [Engine-devel] CPU Overcommit Feature Hi, I've been working on a feature to allow CPU Overcommitment of hosts in a cluster. This first stage allows the engine to consider host cpu threads as cores for the purposes of VM resource allocation. This wiki page has further details, your comments are welcome! http://www.ovirt.org/Features/cpu_overcommit Basically looking good. Hyperthread though is vendor specific. For AMD it's Clustered Multi-Thread while for Intel it's Hyper-Thread Official name is simultaneous multithreading (SMT) but no one outside of the academy will recognize that. in libvirt if I read it right it's attribute name='thread_siblings' So why not just call it threads. We'll have cpuSockets, cpiCores, and cpuThreads, should be clear when in CPU context. In the GUI just change hyperthreads to CPU threads. While in the tool tip explain that it's either AMD Clustered Multi-Thread or Intel Hyperthread Does this affect only the number of potential vCpus for the guests or does this also have an impact on the actual scheduling? So far I always disabled HT out of fear that a 2 vCpu guest might actually be scheduled to run in 2 threads of the same core but now I'm not so sure anymore. In the HT case does KVM know that two threads belong to the same core and will it only schedule its vCpus on distinct cores? Is there some documentation about this somewhere? This is about the maximum number of vCPUs we can give to a VM. If the machine has 32 Physical cores that are hyperthreaded then do we say the max number of vCPUs for a single VM is 32 or 64. If the actual scheduling of vCPUs cannot distinguish between threads and cores then why would you even want to limit yourself to 32 in you example? In that case the scheduling might end up being inefficient no matter how many vCPUs you assign to a guest so why restrict the user? (You might of course want to limit the user for policy reasons but that has nothing to to with the thread/core topic.) On the other hand if KVM does only schedule the vCPUs on distinct cores and extending the count from 32 to 64 implies that this distinction is to be disabled then this will have a performance impact for the guest. In that case I might want to limit the guests to just the 32 physical cores so two vCPUs of a single guest don't get scheduled on two threads of the same core. I've never really looked that closely into the scheduling issue but it might play a role here so I asked if someone had any pointers to infos about how exactly KVM makes its scheduling decisions. Regards, Dennis Dennis, first of all every virtual cpu is basically a qemu-thread which can run on any cpu-thread. The scheduling is done by the kernel scheduler, while we may control it using cpu pinning. ie- you may ask for specific vcpu to run on a specific thread which is from the OS point of view another core. Indeed there are cases where this is not recommended, but other cases where this will actually give you a performance boost, as L1 cache is being shared by the sibling threads. So it's really up to you to test your workload and decide id you wish to utilize it or not. We're giving you powerful tools, and you can decide if and how to use it. What I'm trying to get at is this: Isn't the Count threads as physical cores setting superfluous? If HT is disabled on the node this doesn't do anything anyway but if it is enabled what is to be gained by disabling this option? As far as I can see this makes the UI more complicated for no apparent reason. Regards, Dennis Hi Dennis. Let's take it one at a time; UI wise, this is a cluster level policy, and falls into the optimization tab. So it shouldn't
Re: [Engine-devel] CPU Overcommit Feature
On Wed, Dec 19, 2012 at 09:53:15AM -0500, Doron Fediuck wrote: - Original Message - From: Dan Kenigsberg dan...@redhat.com To: Greg Padgett gpadg...@redhat.com Cc: engine-devel engine-devel@ovirt.org, vdsm-de...@fedorahosted.org Sent: Wednesday, December 19, 2012 3:59:11 PM Subject: Re: [Engine-devel] CPU Overcommit Feature On Mon, Dec 17, 2012 at 09:37:57AM -0500, Greg Padgett wrote: Hi, I've been working on a feature to allow CPU Overcommitment of hosts in a cluster. This first stage allows the engine to consider host cpu threads as cores for the purposes of VM resource allocation. This wiki page has further details, your comments are welcome! http://www.ovirt.org/Features/cpu_overcommit I've commented about the vdsm/engine API on http://gerrit.ovirt.org/#/c/10144/ but it is probably better to reiterate it here. The suggested API is tightly coupled with an ugly hack we pushed to vdsm in order not to solve the issue properly on the first strike. If we had not have report_host_threads_as_cores, I think we'd have a simpler API reporting only cpuThreads and cpuCores; with no funny boolean flags. Let us strive to that position as much as we can. How about asking whoever used report_host_threads_as_cores to unset it once they install Engine 3.2 ? I think that these are very few people, that would not mind this very much. If this is impossible, I'd add a cpuCores2, always reporting the true number, to be used by new Engines. We may even report it only on the very few cases of report_host_threads_as_cores being set. Dan. Hi Dan, Thanks for the review. I agree simply reporting cores and threads would be the right solution. However, when you have hyperthreading turned off you get cores=threads. This is the same situation you have when hyperthreading turned on, and someone used the vdsm configuration of reporting threads as cores. So the engine won't know the real status of the host. This is not surprising, as report_host_threads_as_cores means in blunt English lie to Engine about the number of cores. The newly suggested flag says don't believe what I said in cpuCores, since I'm lying. Next thing we'd have is another flag saying that the former flag was a lie, and cpuCores is actually trustworthy. Instead of dancing this dance, I suggest to stop lying. report_host_threads_as_cores was a hack to assist a older Engine versions. Engine users that needed it had to set it out-of-band on their hosts. Now if they upgrade their Engine, they can -- as easily -- reset that value. If they forget, nothing devastating happens beyond Engine assuming that hyperthreading is off. Please consider this suggestion. I find it the simplest for all involved parties. Dan. We need to be able to tell the difference. So this moves us to cpuCores2 suggestion. This is one possibility (cpuRealCores?), and the alternative is an indication of vdsm config (true/false) which may be removed in the future. I suspect over time cpu and cpu2 will confuse people. So I'd suggest having the boolean and removing it along with the vdsm configuration in the next ovirt version. ___ Engine-devel mailing list Engine-devel@ovirt.org http://lists.ovirt.org/mailman/listinfo/engine-devel
Re: [Engine-devel] CPU Overcommit Feature
On 12/17/2012 05:52 PM, Andrew Cathrow wrote: ... and let's not call this CPU overcommit feature. It's nothing like that - it's Hyperthread handling - Original Message - From: Simon Grinberg si...@redhat.com To: Greg Padgett gpadg...@redhat.com Cc: engine-devel engine-devel@ovirt.org Sent: Monday, December 17, 2012 1:13:03 PM Subject: Re: [Engine-devel] CPU Overcommit Feature - Original Message - From: Greg Padgett gpadg...@redhat.com To: engine-devel engine-devel@ovirt.org Sent: Monday, December 17, 2012 4:37:57 PM Subject: [Engine-devel] CPU Overcommit Feature Hi, I've been working on a feature to allow CPU Overcommitment of hosts in a cluster. This first stage allows the engine to consider host cpu threads as cores for the purposes of VM resource allocation. This wiki page has further details, your comments are welcome! http://www.ovirt.org/Features/cpu_overcommit Basically looking good. Hyperthread though is vendor specific. For AMD it's Clustered Multi-Thread while for Intel it's Hyper-Thread Official name is simultaneous multithreading (SMT) but no one outside of the academy will recognize that. in libvirt if I read it right it's attribute name='thread_siblings' So why not just call it threads. We'll have cpuSockets, cpiCores, and cpuThreads, should be clear when in CPU context. In the GUI just change hyperthreads to CPU threads. While in the tool tip explain that it's either AMD Clustered Multi-Thread or Intel Hyperthread Thanks in advance, Greg ___ Engine-devel mailing list Engine-devel@ovirt.org http://lists.ovirt.org/mailman/listinfo/engine-devel ___ Engine-devel mailing list Engine-devel@ovirt.org http://lists.ovirt.org/mailman/listinfo/engine-devel Thanks Simon and Andrew. I've moved the wiki page [1] (with a redirect at the old name), updated the terms within to not be vendor-specific, and will do the same with the implementation. [1] http://www.ovirt.org/Features/cpu_thread_handling ___ Engine-devel mailing list Engine-devel@ovirt.org http://lists.ovirt.org/mailman/listinfo/engine-devel
Re: [Engine-devel] CPU Overcommit Feature
On 12/18/2012 06:33 PM, Andrew Cathrow wrote: - Original Message - From: Dennis Jacobfeuerborn denni...@conversis.de To: engine-devel@ovirt.org Sent: Tuesday, December 18, 2012 12:30:34 PM Subject: Re: [Engine-devel] CPU Overcommit Feature On 12/17/2012 07:13 PM, Simon Grinberg wrote: - Original Message - From: Greg Padgett gpadg...@redhat.com To: engine-devel engine-devel@ovirt.org Sent: Monday, December 17, 2012 4:37:57 PM Subject: [Engine-devel] CPU Overcommit Feature Hi, I've been working on a feature to allow CPU Overcommitment of hosts in a cluster. This first stage allows the engine to consider host cpu threads as cores for the purposes of VM resource allocation. This wiki page has further details, your comments are welcome! http://www.ovirt.org/Features/cpu_overcommit Basically looking good. Hyperthread though is vendor specific. For AMD it's Clustered Multi-Thread while for Intel it's Hyper-Thread Official name is simultaneous multithreading (SMT) but no one outside of the academy will recognize that. in libvirt if I read it right it's attribute name='thread_siblings' So why not just call it threads. We'll have cpuSockets, cpiCores, and cpuThreads, should be clear when in CPU context. In the GUI just change hyperthreads to CPU threads. While in the tool tip explain that it's either AMD Clustered Multi-Thread or Intel Hyperthread Does this affect only the number of potential vCpus for the guests or does this also have an impact on the actual scheduling? So far I always disabled HT out of fear that a 2 vCpu guest might actually be scheduled to run in 2 threads of the same core but now I'm not so sure anymore. In the HT case does KVM know that two threads belong to the same core and will it only schedule its vCpus on distinct cores? Is there some documentation about this somewhere? This is about the maximum number of vCPUs we can give to a VM. If the machine has 32 Physical cores that are hyperthreaded then do we say the max number of vCPUs for a single VM is 32 or 64. If the actual scheduling of vCPUs cannot distinguish between threads and cores then why would you even want to limit yourself to 32 in you example? In that case the scheduling might end up being inefficient no matter how many vCPUs you assign to a guest so why restrict the user? (You might of course want to limit the user for policy reasons but that has nothing to to with the thread/core topic.) On the other hand if KVM does only schedule the vCPUs on distinct cores and extending the count from 32 to 64 implies that this distinction is to be disabled then this will have a performance impact for the guest. In that case I might want to limit the guests to just the 32 physical cores so two vCPUs of a single guest don't get scheduled on two threads of the same core. I've never really looked that closely into the scheduling issue but it might play a role here so I asked if someone had any pointers to infos about how exactly KVM makes its scheduling decisions. Regards, Dennis ___ Engine-devel mailing list Engine-devel@ovirt.org http://lists.ovirt.org/mailman/listinfo/engine-devel
Re: [Engine-devel] CPU Overcommit Feature
- Original Message - From: Dennis Jacobfeuerborn denni...@conversis.de To: Andrew Cathrow acath...@redhat.com Cc: engine-devel@ovirt.org Sent: Tuesday, December 18, 2012 7:59:26 PM Subject: Re: [Engine-devel] CPU Overcommit Feature On 12/18/2012 06:33 PM, Andrew Cathrow wrote: - Original Message - From: Dennis Jacobfeuerborn denni...@conversis.de To: engine-devel@ovirt.org Sent: Tuesday, December 18, 2012 12:30:34 PM Subject: Re: [Engine-devel] CPU Overcommit Feature On 12/17/2012 07:13 PM, Simon Grinberg wrote: - Original Message - From: Greg Padgett gpadg...@redhat.com To: engine-devel engine-devel@ovirt.org Sent: Monday, December 17, 2012 4:37:57 PM Subject: [Engine-devel] CPU Overcommit Feature Hi, I've been working on a feature to allow CPU Overcommitment of hosts in a cluster. This first stage allows the engine to consider host cpu threads as cores for the purposes of VM resource allocation. This wiki page has further details, your comments are welcome! http://www.ovirt.org/Features/cpu_overcommit Basically looking good. Hyperthread though is vendor specific. For AMD it's Clustered Multi-Thread while for Intel it's Hyper-Thread Official name is simultaneous multithreading (SMT) but no one outside of the academy will recognize that. in libvirt if I read it right it's attribute name='thread_siblings' So why not just call it threads. We'll have cpuSockets, cpiCores, and cpuThreads, should be clear when in CPU context. In the GUI just change hyperthreads to CPU threads. While in the tool tip explain that it's either AMD Clustered Multi-Thread or Intel Hyperthread Does this affect only the number of potential vCpus for the guests or does this also have an impact on the actual scheduling? So far I always disabled HT out of fear that a 2 vCpu guest might actually be scheduled to run in 2 threads of the same core but now I'm not so sure anymore. In the HT case does KVM know that two threads belong to the same core and will it only schedule its vCpus on distinct cores? Is there some documentation about this somewhere? This is about the maximum number of vCPUs we can give to a VM. If the machine has 32 Physical cores that are hyperthreaded then do we say the max number of vCPUs for a single VM is 32 or 64. If the actual scheduling of vCPUs cannot distinguish between threads and cores then why would you even want to limit yourself to 32 in you example? In that case the scheduling might end up being inefficient no matter how many vCPUs you assign to a guest so why restrict the user? (You might of course want to limit the user for policy reasons but that has nothing to to with the thread/core topic.) On the other hand if KVM does only schedule the vCPUs on distinct cores and extending the count from 32 to 64 implies that this distinction is to be disabled then this will have a performance impact for the guest. In that case I might want to limit the guests to just the 32 physical cores so two vCPUs of a single guest don't get scheduled on two threads of the same core. I've never really looked that closely into the scheduling issue but it might play a role here so I asked if someone had any pointers to infos about how exactly KVM makes its scheduling decisions. Regards, Dennis Dennis, first of all every virtual cpu is basically a qemu-thread which can run on any cpu-thread. The scheduling is done by the kernel scheduler, while we may control it using cpu pinning. ie- you may ask for specific vcpu to run on a specific thread which is from the OS point of view another core. Indeed there are cases where this is not recommended, but other cases where this will actually give you a performance boost, as L1 cache is being shared by the sibling threads. So it's really up to you to test your workload and decide id you wish to utilize it or not. We're giving you powerful tools, and you can decide if and how to use it. Doron ___ Engine-devel mailing list Engine-devel@ovirt.org http://lists.ovirt.org/mailman/listinfo/engine-devel