Re: [openstack-dev] [Nova] FPGA as a resource
Hi all, We agree this is an ingesting direction we can do exploration on OpenStack and actually we already have such FPGA/GPU as a resource in our public research cloud called SuperVessel (https://ptopenlab.com). With modification of OpenStack from community, nearly all features mentioned by Roman Dobosz are supported, such as FPGA boards detection, scheduling, images management and user API layer. Here are some materials for you to reference [1][2]. Unlike CPU and GPU, right now FPGA accelerator is developed without common protocol above PCIe protocol. Along with huge kinds of FPGA chips, FPGA boards, it is hard to be managed. Passing through the FPGA as a general PCIe device to the application stack is a straight way to expose the FPGA acceleration ability in cloud while it is far way to make it easy used. It dose not help the accelerator detection, reload, reconfiguration, deployment in cloud. Good news is that, the FPGA community is actively promoting high level programming language, such as OpenCL. They hide nearly all hardware detail for the developers so it make the "accelerator" common in different hardware. It is reasonable to see the FPGA become a popular acceleration hardware. It also reasonable for OpenStack community begin the discussion to support the FPGA resource management. [1] http://openpowerfoundation.org/blogs/fpga-acceleration-in-a-power8-cloud/ [2] http://dl.acm.org/citation.cfm?id=2597929 Best regards, CHEN Fei From: Qiming Teng <teng...@linux.vnet.ibm.com> To: "OpenStack Development Mailing List (not for usage questions)" <openstack-dev@lists.openstack.org> Cc: Fei K Chen/China/IBM@IBMCN, Yong Hua Lin/China/IBM@IBMCN Date: 2016/04/06 12:25 Subject: Re: [openstack-dev] [Nova] FPGA as a resource Emm... finally this is brought up. We from IBM have already done some work on FPGA/GPU resource management [1]. Let me bring the SMEs into this discussion and see if we together can work out a concrete roadmap to land this upstream. Fei and Yonghua, this is indeed very interesting a topic for us. [1] SuperVessel Cloud: https://ptopenlab.com/ Regards, Qiming On Tue, Apr 05, 2016 at 02:27:30PM +0200, Roman Dobosz wrote: > Hey all, > > On yesterday's scheduler meeting I was raised the idea of bringing up > the FPGA to the OpenStack as the resource, which than might be exposed > to the VMs. > > The use cases for motivations, why one want do this, are pretty broad - > having such chip ready on the computes might be beneficial either for > consumers of the technology and data center administrators. The > utilization of the hardware is very broad - the only limitations are > human imagination and hardware capability - since it might be used for > accelerating execution of algorithms from compression and cryptography, > through pattern recognition, transcoding to voice/video analysis and > processing and all the others in between. Using FPGA to perform data > processing may significantly reduce CPU utilization, the time and power > consumption, which is a benefit on its own. > > On OpenStack side, unlike utilizing the CPU or memory, for actually > using specific algorithm with FPGAs, it has to be programmed first. So > in simplified scenario, it might go like this: > > * User selects VM with image which supports acceleration, > * Scheduler selects the appropriate compute host with FPGA available, > * Compute gets request, program IP into FPGA and then boot up the > VM with accelerator attached. > * If VM is removed, it may optionally erase the FPGA. > > As you can see, it seems not complicated at this point, however it > become more complex due to following things we also have to take into > consideration: > > * recent FPGA are divided into regions (or slots) and every of them > can be programmed separately > * slots may or may not fit the same bitstream (the program which FPGA > is fed, the IP) > * There is several products around (Altera, Xilinx, others), which make > bitstream incompatible. Even between the products of the same company > * libraries which abstract the hardware layer like AAL[1] and their > versions > * for some products, there is a need for tracking memory usage, which > is located on PCI boards > * some of the FPGAs can be exposed using SR-IOV, while some other not, > which implies multiple usage abilities > > In other words, it may be necessary to incorporate another actions: > > * properly discover FPGA and its capabilities > * schedule right bitstream with corresponding matching unoccupied FPGA > device/slot > * actual program FPGA > * provide libraries into VM, which are necessary for interacting between > user program and the exposed FPGA (or AAL) (this may be optional, > since user can upload complete ima
Re: [openstack-dev] [Nova] FPGA as a resource
On Wed, Apr 6, 2016 at 6:37 AM, Daniel P. Berrangewrote: > Most of the > work to support FPGA will be internal to nova, to deal with modelling > of assignable devices and their scheduling / allocation. I think the EPA blueprint already covers some of the FPGA device discovery & scheduling functionality needed. https://wiki.openstack.org/wiki/Enhanced-platform-awareness-pcie However, I was reading the Intel ONP whitepapers earlier, and found that Intel has dropped coverage of the EPA functionality in 2.1: -> https://download.01.org/packet-processing/ONPS2.1/Intel_ONP_Release_2.1_Reference_Architecture_Guide_Rev1.0.pdf EPA is mentioned in section "6.3 Enhanced Platform Awareness" in 2.0: -> https://download.01.org/packet-processing/ONPS2.0/Intel_ONP_Release_2.0_Reference_Architecture_Guide_Rev1.0-1.pdf Rayson == Open Grid Scheduler - The Official Open Source Grid Engine http://gridscheduler.sourceforge.net/ http://gridscheduler.sourceforge.net/GridEngine/GridEngineCloud.html > > Regards, > Daniel > -- > |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| > |: http://libvirt.org -o- http://virt-manager.org :| > |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| > |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| > > __ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] FPGA as a resource
Yes the project is quite new, we are still developing requirements at the OPNFV side. We will update with full description after the summit when we got all the inputs. Look forward to have a chat with you at the summit :) I think you are right that Nomad does not solve the problem per se in this thread, however it would be the natural next step. After having Nova being able to use FPGA as a resource, you will need to be able schedule FPGA resources for the VM. We also have a team based in Canada that specialized in FPGA related management that will participate in the Nomad development, we could have more deep dive chat when we got chance :) On Wed, Apr 6, 2016 at 6:28 PM, Roman Doboszwrote: > On Wed, 6 Apr 2016 16:12:16 +0800 > Zhipeng Huang wrote: > > > You are actually touching on something we have been working on. There is > a > > team in OPNFV DPACC project has been working acceleration related topics, > > including folks from CMCC, Intel, ARM, Freescale, Huawei. We found out > that > > in order to have acceleration working under NFV scenrios, other than Nova > > and Neutron's support, we also need a standalone service that manage > > accelerators itself. > > > > That means we want to treat accelerators, and FPGA being an important > part > > of it, as a first class resource citizen and we want to be able to do > life > > cycle management and scheduling on acceleration resources. > > > > Based upon that requirement we started a new project called Nomad [1] on > > Jan this year, to serve as an OpenStack service for distributed > > acceleration management. > > > > We've just started the project, and currently discussing the first BP > [2]. > > We have a team working on IP-SEC based accelerator mgmt, and would love > to > > have more people to work on topics like FPGA. > > > > We also have a topic on introducing Nomad accepted in Austin Summit [3]. > > > > You are more than welcomed to join the conversation : ) > > Thanks! I'll try to attend. > > Nevertheless, I've briefly looked at page of project Nomad, and don't > quite get it, how this might be related to cases described in this > thread - i.e. providing attachable, non trivial devices as a new > resources in Nova. > > -- > Cheers, > Roman Dobosz > > __ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > -- Zhipeng (Howard) Huang Standard Engineer IT Standard & Patent/IT Prooduct Line Huawei Technologies Co,. Ltd Email: huangzhip...@huawei.com Office: Huawei Industrial Base, Longgang, Shenzhen (Previous) Research Assistant Mobile Ad-Hoc Network Lab, Calit2 University of California, Irvine Email: zhipe...@uci.edu Office: Calit2 Building Room 2402 OpenStack, OPNFV, OpenDaylight, OpenCompute Aficionado __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] FPGA as a resource
On Wed, Apr 06, 2016 at 12:28:02PM +0200, Roman Dobosz wrote: > On Wed, 6 Apr 2016 16:12:16 +0800 > Zhipeng Huangwrote: > > > You are actually touching on something we have been working on. There is a > > team in OPNFV DPACC project has been working acceleration related topics, > > including folks from CMCC, Intel, ARM, Freescale, Huawei. We found out that > > in order to have acceleration working under NFV scenrios, other than Nova > > and Neutron's support, we also need a standalone service that manage > > accelerators itself. > > > > That means we want to treat accelerators, and FPGA being an important part > > of it, as a first class resource citizen and we want to be able to do life > > cycle management and scheduling on acceleration resources. > > > > Based upon that requirement we started a new project called Nomad [1] on > > Jan this year, to serve as an OpenStack service for distributed > > acceleration management. > > > > We've just started the project, and currently discussing the first BP [2]. > > We have a team working on IP-SEC based accelerator mgmt, and would love to > > have more people to work on topics like FPGA. > > > > We also have a topic on introducing Nomad accepted in Austin Summit [3]. > > > > You are more than welcomed to join the conversation : ) > > Thanks! I'll try to attend. > > Nevertheless, I've briefly looked at page of project Nomad, and don't > quite get it, how this might be related to cases described in this > thread - i.e. providing attachable, non trivial devices as a new > resources in Nova. I don't think it is really relevant in the immediate term. Most of the work to support FPGA will be internal to nova, to deal with modelling of assignable devices and their scheduling / allocation. Regards, Daniel -- |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] FPGA as a resource
On Wed, 6 Apr 2016 16:12:16 +0800 Zhipeng Huangwrote: > You are actually touching on something we have been working on. There is a > team in OPNFV DPACC project has been working acceleration related topics, > including folks from CMCC, Intel, ARM, Freescale, Huawei. We found out that > in order to have acceleration working under NFV scenrios, other than Nova > and Neutron's support, we also need a standalone service that manage > accelerators itself. > > That means we want to treat accelerators, and FPGA being an important part > of it, as a first class resource citizen and we want to be able to do life > cycle management and scheduling on acceleration resources. > > Based upon that requirement we started a new project called Nomad [1] on > Jan this year, to serve as an OpenStack service for distributed > acceleration management. > > We've just started the project, and currently discussing the first BP [2]. > We have a team working on IP-SEC based accelerator mgmt, and would love to > have more people to work on topics like FPGA. > > We also have a topic on introducing Nomad accepted in Austin Summit [3]. > > You are more than welcomed to join the conversation : ) Thanks! I'll try to attend. Nevertheless, I've briefly looked at page of project Nomad, and don't quite get it, how this might be related to cases described in this thread - i.e. providing attachable, non trivial devices as a new resources in Nova. -- Cheers, Roman Dobosz __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] FPGA as a resource
On Wed, Apr 06, 2016 at 07:34:46AM +0200, Roman Dobosz wrote: > On Tue, 5 Apr 2016 13:58:44 +0100 > "Daniel P. Berrange"wrote: > > > Along similar lines we have proposals to add vGPU support to Nova, > > where the vGPUs may or may not be exposed using SR-IOV. We also want > > to be able to on the fly decide whether any physical GPU is assigned > > entirely to a guest as a full PCI device, or whether we only assign > > individual "virtual functions" of the GPU. This means that even if > > the GPU in question does *not* use SR-IOV, we still need to track > > the GPU and vGPUs in the same way as we track PCI devices, so that > > we can avoid assigning a vGPU to the guest, if the underlying physical > > PCI device is already assigned to the guest. > > That's correct. I'd like to mention, that FPGAs can be also exposed > other way than PCI (like in Xeon+FPGA). Not sure if this also applies > to GPU. > > > I can see we will have much the same issue with FPGAs, where we may > > either want to assign the entire physical PCI device to a guest, or > > just assign a particular slot in the FPGA to the guest. So even if > > the FPGA is not using SR-IOV, we need to tie this all into the PCI > > device tracking code, as we are intending for vGPUs. > > > > All in all, I think we probably ought to generalize the PCI device > > assignment modelling so that we're actually modelling generic > > hardware devices which may or may not be PCI based, so that we can > > accurately track the relationships between the devices. > > > > With NIC devices we're also seeing a need to expose capabilities > > against the PCI devices, so that the schedular can be more selective > > in deciding which particular devices to assign. eg so we can distinguish > > between NICs which support RDMA and those which don't, or identify NIC > > with hardware offload features, and so on. I can see this need to > > associate capabilities with devices is something that will likely > > be needed for the FPGA scenario, and vGPUs too. So again this points > > towards more general purpose modelling of assignable hardware devices > > beyond the limited PCI device modelling we've got today. > > > > Looking to the future I think we'll see more usecases for device > > assignment appearing for other types of device. > > > > IOW, I think it would be a mistake to model FPGAs as a distinct > > object type on their own. Generalization of assignable devices > > is the way to go > > That's why I've bring the topic here on the list, so we can think about > similar devices which could be generalized into one common accelerator > type or even think about modeling PCI as such. I wouldn't specialize it to "accelerators" as we'll inevitably come across a need for other types of device assignment. We should just generalize it to track *any* type of host hardware device that is potentially assignable. Regards, Daniel -- |: http://berrange.com -o-http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :| __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] FPGA as a resource
Hi Roman, You are actually touching on something we have been working on. There is a team in OPNFV DPACC project has been working acceleration related topics, including folks from CMCC, Intel, ARM, Freescale, Huawei. We found out that in order to have acceleration working under NFV scenrios, other than Nova and Neutron's support, we also need a standalone service that manage accelerators itself. That means we want to treat accelerators, and FPGA being an important part of it, as a first class resource citizen and we want to be able to do life cycle management and scheduling on acceleration resources. Based upon that requirement we started a new project called Nomad [1] on Jan this year, to serve as an OpenStack service for distributed acceleration management. We've just started the project, and currently discussing the first BP [2]. We have a team working on IP-SEC based accelerator mgmt, and would love to have more people to work on topics like FPGA. We also have a topic on introducing Nomad accepted in Austin Summit [3]. You are more than welcomed to join the conversation : ) [1] https://wiki.openstack.org/wiki/Nomad [2] https://review.openstack.org/#/c/284304/ [3] https://www.openstack.org/summit/austin-2016/summit-schedule/global-search?t=nomad On Wed, Apr 6, 2016 at 1:34 PM, Roman Doboszwrote: > On Tue, 5 Apr 2016 13:58:44 +0100 > "Daniel P. Berrange" wrote: > > > Along similar lines we have proposals to add vGPU support to Nova, > > where the vGPUs may or may not be exposed using SR-IOV. We also want > > to be able to on the fly decide whether any physical GPU is assigned > > entirely to a guest as a full PCI device, or whether we only assign > > individual "virtual functions" of the GPU. This means that even if > > the GPU in question does *not* use SR-IOV, we still need to track > > the GPU and vGPUs in the same way as we track PCI devices, so that > > we can avoid assigning a vGPU to the guest, if the underlying physical > > PCI device is already assigned to the guest. > > That's correct. I'd like to mention, that FPGAs can be also exposed > other way than PCI (like in Xeon+FPGA). Not sure if this also applies > to GPU. > > > I can see we will have much the same issue with FPGAs, where we may > > either want to assign the entire physical PCI device to a guest, or > > just assign a particular slot in the FPGA to the guest. So even if > > the FPGA is not using SR-IOV, we need to tie this all into the PCI > > device tracking code, as we are intending for vGPUs. > > > > All in all, I think we probably ought to generalize the PCI device > > assignment modelling so that we're actually modelling generic > > hardware devices which may or may not be PCI based, so that we can > > accurately track the relationships between the devices. > > > > With NIC devices we're also seeing a need to expose capabilities > > against the PCI devices, so that the schedular can be more selective > > in deciding which particular devices to assign. eg so we can distinguish > > between NICs which support RDMA and those which don't, or identify NIC > > with hardware offload features, and so on. I can see this need to > > associate capabilities with devices is something that will likely > > be needed for the FPGA scenario, and vGPUs too. So again this points > > towards more general purpose modelling of assignable hardware devices > > beyond the limited PCI device modelling we've got today. > > > > Looking to the future I think we'll see more usecases for device > > assignment appearing for other types of device. > > > > IOW, I think it would be a mistake to model FPGAs as a distinct > > object type on their own. Generalization of assignable devices > > is the way to go > > That's why I've bring the topic here on the list, so we can think about > similar devices which could be generalized into one common accelerator > type or even think about modeling PCI as such. > > > > All of that makes modelling resource extremely complicated, contrary to > > > CPU resource for example. I'd like to discuss how the goal of having > > > reprogrammable accelerators in OpenStack can be achieved. Ideally I'd > > > like to fit it into Jay and Chris work on resource-*. > > I think you shouldn't look at the FPGAs as being like CPU resource, but > > rather look at them as a generalization of PCI device asignment. > > CPU in this context was only an example of "easy" resource, which > doesn't need any preparation before VM can use it :) > > -- > Cheers, > Roman Dobosz > > __ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > -- Zhipeng (Howard) Huang Standard Engineer IT Standard & Patent/IT Prooduct Line Huawei Technologies Co,. Ltd Email:
Re: [openstack-dev] [Nova] FPGA as a resource
On Tue, 5 Apr 2016 13:58:44 +0100 "Daniel P. Berrange"wrote: > Along similar lines we have proposals to add vGPU support to Nova, > where the vGPUs may or may not be exposed using SR-IOV. We also want > to be able to on the fly decide whether any physical GPU is assigned > entirely to a guest as a full PCI device, or whether we only assign > individual "virtual functions" of the GPU. This means that even if > the GPU in question does *not* use SR-IOV, we still need to track > the GPU and vGPUs in the same way as we track PCI devices, so that > we can avoid assigning a vGPU to the guest, if the underlying physical > PCI device is already assigned to the guest. That's correct. I'd like to mention, that FPGAs can be also exposed other way than PCI (like in Xeon+FPGA). Not sure if this also applies to GPU. > I can see we will have much the same issue with FPGAs, where we may > either want to assign the entire physical PCI device to a guest, or > just assign a particular slot in the FPGA to the guest. So even if > the FPGA is not using SR-IOV, we need to tie this all into the PCI > device tracking code, as we are intending for vGPUs. > > All in all, I think we probably ought to generalize the PCI device > assignment modelling so that we're actually modelling generic > hardware devices which may or may not be PCI based, so that we can > accurately track the relationships between the devices. > > With NIC devices we're also seeing a need to expose capabilities > against the PCI devices, so that the schedular can be more selective > in deciding which particular devices to assign. eg so we can distinguish > between NICs which support RDMA and those which don't, or identify NIC > with hardware offload features, and so on. I can see this need to > associate capabilities with devices is something that will likely > be needed for the FPGA scenario, and vGPUs too. So again this points > towards more general purpose modelling of assignable hardware devices > beyond the limited PCI device modelling we've got today. > > Looking to the future I think we'll see more usecases for device > assignment appearing for other types of device. > > IOW, I think it would be a mistake to model FPGAs as a distinct > object type on their own. Generalization of assignable devices > is the way to go That's why I've bring the topic here on the list, so we can think about similar devices which could be generalized into one common accelerator type or even think about modeling PCI as such. > > All of that makes modelling resource extremely complicated, contrary to > > CPU resource for example. I'd like to discuss how the goal of having > > reprogrammable accelerators in OpenStack can be achieved. Ideally I'd > > like to fit it into Jay and Chris work on resource-*. > I think you shouldn't look at the FPGAs as being like CPU resource, but > rather look at them as a generalization of PCI device asignment. CPU in this context was only an example of "easy" resource, which doesn't need any preparation before VM can use it :) -- Cheers, Roman Dobosz __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] FPGA as a resource
Hi, On Tue, Apr 5, 2016 at 8:58 PM, Daniel P. Berrangewrote: >> As you can see, it seems not complicated at this point, however it >> become more complex due to following things we also have to take into >> consideration: I agree with this. Many kinds of resources could be considered, not just to augment processing, but also to provide input (or other forms of output), such as video devices, etc. regards, Gerard -- Gerard Braad β εζεΎ· F/OSS & IT Consultant in Beijing http://gbraad.nl gpg: 0x592CFE75 __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] FPGA as a resource
Emm... finally this is brought up. We from IBM have already done some work on FPGA/GPU resource management [1]. Let me bring the SMEs into this discussion and see if we together can work out a concrete roadmap to land this upstream. Fei and Yonghua, this is indeed very interesting a topic for us. [1] SuperVessel Cloud: https://ptopenlab.com/ Regards, Qiming On Tue, Apr 05, 2016 at 02:27:30PM +0200, Roman Dobosz wrote: > Hey all, > > On yesterday's scheduler meeting I was raised the idea of bringing up > the FPGA to the OpenStack as the resource, which than might be exposed > to the VMs. > > The use cases for motivations, why one want do this, are pretty broad - > having such chip ready on the computes might be beneficial either for > consumers of the technology and data center administrators. The > utilization of the hardware is very broad - the only limitations are > human imagination and hardware capability - since it might be used for > accelerating execution of algorithms from compression and cryptography, > through pattern recognition, transcoding to voice/video analysis and > processing and all the others in between. Using FPGA to perform data > processing may significantly reduce CPU utilization, the time and power > consumption, which is a benefit on its own. > > On OpenStack side, unlike utilizing the CPU or memory, for actually > using specific algorithm with FPGAs, it has to be programmed first. So > in simplified scenario, it might go like this: > > * User selects VM with image which supports acceleration, > * Scheduler selects the appropriate compute host with FPGA available, > * Compute gets request, program IP into FPGA and then boot up the > VM with accelerator attached. > * If VM is removed, it may optionally erase the FPGA. > > As you can see, it seems not complicated at this point, however it > become more complex due to following things we also have to take into > consideration: > > * recent FPGA are divided into regions (or slots) and every of them > can be programmed separately > * slots may or may not fit the same bitstream (the program which FPGA > is fed, the IP) > * There is several products around (Altera, Xilinx, others), which make > bitstream incompatible. Even between the products of the same company > * libraries which abstract the hardware layer like AAL[1] and their > versions > * for some products, there is a need for tracking memory usage, which > is located on PCI boards > * some of the FPGAs can be exposed using SR-IOV, while some other not, > which implies multiple usage abilities > > In other words, it may be necessary to incorporate another actions: > > * properly discover FPGA and its capabilities > * schedule right bitstream with corresponding matching unoccupied FPGA > device/slot > * actual program FPGA > * provide libraries into VM, which are necessary for interacting between > user program and the exposed FPGA (or AAL) (this may be optional, > since user can upload complete image with everything in place) > * bitstream images have to be keep in some kind of service (Glance?) > with some kind of way for identifying which image match what FPGA > > All of that makes modelling resource extremely complicated, contrary to > CPU resource for example. I'd like to discuss how the goal of having > reprogrammable accelerators in OpenStack can be achieved. Ideally I'd > like to fit it into Jay and Chris work on resource-*. > > Looking forward for any comments :) > > [1] > http://www.intel.com/content/dam/doc/white-paper/quickassist-technology-aal-white-paper.pdf > > -- > Cheers, > Roman Dobosz > > __ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Nova] FPGA as a resource
On Tue, Apr 05, 2016 at 02:27:30PM +0200, Roman Dobosz wrote: > Hey all, > > On yesterday's scheduler meeting I was raised the idea of bringing up > the FPGA to the OpenStack as the resource, which than might be exposed > to the VMs. > > The use cases for motivations, why one want do this, are pretty broad - > having such chip ready on the computes might be beneficial either for > consumers of the technology and data center administrators. The > utilization of the hardware is very broad - the only limitations are > human imagination and hardware capability - since it might be used for > accelerating execution of algorithms from compression and cryptography, > through pattern recognition, transcoding to voice/video analysis and > processing and all the others in between. Using FPGA to perform data > processing may significantly reduce CPU utilization, the time and power > consumption, which is a benefit on its own. > > On OpenStack side, unlike utilizing the CPU or memory, for actually > using specific algorithm with FPGAs, it has to be programmed first. So > in simplified scenario, it might go like this: > > * User selects VM with image which supports acceleration, > * Scheduler selects the appropriate compute host with FPGA available, > * Compute gets request, program IP into FPGA and then boot up the > VM with accelerator attached. > * If VM is removed, it may optionally erase the FPGA. > > As you can see, it seems not complicated at this point, however it > become more complex due to following things we also have to take into > consideration: > > * recent FPGA are divided into regions (or slots) and every of them > can be programmed separately > * slots may or may not fit the same bitstream (the program which FPGA > is fed, the IP) > * There is several products around (Altera, Xilinx, others), which make > bitstream incompatible. Even between the products of the same company > * libraries which abstract the hardware layer like AAL[1] and their > versions > * for some products, there is a need for tracking memory usage, which > is located on PCI boards > * some of the FPGAs can be exposed using SR-IOV, while some other not, > which implies multiple usage abilities Along similar lines we have proposals to add vGPU support to Nova, where the vGPUs may or may not be exposed using SR-IOV. We also want to be able to on the fly decide whether any physical GPU is assigned entirely to a guest as a full PCI device, or whether we only assign individual "virtual functions" of the GPU. This means that even if the GPU in question does *not* use SR-IOV, we still need to track the GPU and vGPUs in the same way as we track PCI devices, so that we can avoid assigning a vGPU to the guest, if the underlying physical PCI device is already assigned to the guest. I can see we will have much the same issue with FPGAs, where we may either want to assign the entire physical PCI device to a guest, or just assign a particular slot in the FPGA to the guest. So even if the FPGA is not using SR-IOV, we need to tie this all into the PCI device tracking code, as we are intending for vGPUs. All in all, I think we probably ought to generalize the PCI device assignment modelling so that we're actually modelling generic hardware devices which may or may not be PCI based, so that we can accurately track the relationships between the devices. With NIC devices we're also seeing a need to expose capabilities against the PCI devices, so that the schedular can be more selective in deciding which particular devices to assign. eg so we can distinguish between NICs which support RDMA and those which don't, or identify NIC with hardware offload features, and so on. I can see this need to associate capabilities with devices is something that will likely be needed for the FPGA scenario, and vGPUs too. So again this points towards more general purpose modelling of assignable hardware devices beyond the limited PCI device modelling we've got today. Looking to the future I think we'll see more usecases for device assignment appearing for other types of device. IOW, I think it would be a mistake to model FPGAs as a distinct object type on their own. Generalization of assignable devices is the way to go > In other words, it may be necessary to incorporate another actions: > > * properly discover FPGA and its capabilities > * schedule right bitstream with corresponding matching unoccupied FPGA > device/slot > * actual program FPGA > * provide libraries into VM, which are necessary for interacting between > user program and the exposed FPGA (or AAL) (this may be optional, > since user can upload complete image with everything in place) > * bitstream images have to be keep in some kind of service (Glance?) > with some kind of way for identifying which image match what FPGA > > All of that makes modelling resource extremely complicated, contrary to > CPU resource for example. I'd like to discuss how the goal
[openstack-dev] [Nova] FPGA as a resource
Hey all, On yesterday's scheduler meeting I was raised the idea of bringing up the FPGA to the OpenStack as the resource, which than might be exposed to the VMs. The use cases for motivations, why one want do this, are pretty broad - having such chip ready on the computes might be beneficial either for consumers of the technology and data center administrators. The utilization of the hardware is very broad - the only limitations are human imagination and hardware capability - since it might be used for accelerating execution of algorithms from compression and cryptography, through pattern recognition, transcoding to voice/video analysis and processing and all the others in between. Using FPGA to perform data processing may significantly reduce CPU utilization, the time and power consumption, which is a benefit on its own. On OpenStack side, unlike utilizing the CPU or memory, for actually using specific algorithm with FPGAs, it has to be programmed first. So in simplified scenario, it might go like this: * User selects VM with image which supports acceleration, * Scheduler selects the appropriate compute host with FPGA available, * Compute gets request, program IP into FPGA and then boot up the VM with accelerator attached. * If VM is removed, it may optionally erase the FPGA. As you can see, it seems not complicated at this point, however it become more complex due to following things we also have to take into consideration: * recent FPGA are divided into regions (or slots) and every of them can be programmed separately * slots may or may not fit the same bitstream (the program which FPGA is fed, the IP) * There is several products around (Altera, Xilinx, others), which make bitstream incompatible. Even between the products of the same company * libraries which abstract the hardware layer like AAL[1] and their versions * for some products, there is a need for tracking memory usage, which is located on PCI boards * some of the FPGAs can be exposed using SR-IOV, while some other not, which implies multiple usage abilities In other words, it may be necessary to incorporate another actions: * properly discover FPGA and its capabilities * schedule right bitstream with corresponding matching unoccupied FPGA device/slot * actual program FPGA * provide libraries into VM, which are necessary for interacting between user program and the exposed FPGA (or AAL) (this may be optional, since user can upload complete image with everything in place) * bitstream images have to be keep in some kind of service (Glance?) with some kind of way for identifying which image match what FPGA All of that makes modelling resource extremely complicated, contrary to CPU resource for example. I'd like to discuss how the goal of having reprogrammable accelerators in OpenStack can be achieved. Ideally I'd like to fit it into Jay and Chris work on resource-*. Looking forward for any comments :) [1] http://www.intel.com/content/dam/doc/white-paper/quickassist-technology-aal-white-paper.pdf -- Cheers, Roman Dobosz __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev