Jump to content

Multi-user Computer - How?

Okay here is the idea:-

1) U make a really powerful over-kill PC.

2) Connect several Stations( each including a Monitor, Mice, Keyboard & USB Hub).

3) Have multiple users access and work on things at the same time from their respective stations.

4) HOW DO I MAKE THIS HAPPEN???

 

Note:-

The operating system has to be Windows as these are 3D design workstations using software that only works on Windows. 

Link to comment
Share on other sites

Link to post
Share on other sites

Impossible

Intel 3570k 3,4@4,5 1,12v Scythe Mugen 3 gigabyte 770     MSi z77a GD55    corsair vengeance 8 gb  corsair CX600M Bitfenix Outlaw 4 casefans

 

Link to comment
Share on other sites

Link to post
Share on other sites

go read this

http://en.wikipedia.org/wiki/Multiseat_configuration

 

it's been explained in here

http://www.reddit.com/r/buildapc/comments/1wqbhu/can_two_people_use_the_same_computer_at_the_same/

 

anyway it's not really uncommon in my country, the problem is, setup probably lot's more complicated

I've seen and use these kind setup in all McDonalds, common problem for me

1. Resoultion is limited to 800x600/1024x800, and it only could be change from main computer

2. it doesn't support 3D Hardware Accelerator, so no flash, no 3D games, no Render

practically useless if you want make the client side a workstation except you doing multisheet/document

 

Although lot's of improvement (from software developer log)

with new launch intel/amd mini pc/pico pc/soc solution with very low power usage for based to mid usage (about 5-20w) this multiseat config no longer be an option nowdays, people also prefer staring with their tablet/phone on public resort.

 

you can try free trial here: http://www.miniframe.com/products/softxpand-2011-duo.html?resource=Technology

Link to comment
Share on other sites

Link to post
Share on other sites

You're statements are rather vague

 

But if you're referring to real-time editing the same piece of work through different connected workstations here is my answer:

In my experience accessing and editing the same thing/work from different machines is similar to  what you'd call cloud computing. And it's a great way of optimizing workflow when working with documents presentations etc. but that's it.

 

Working and accessing the file at the same time in 3D work is somewhat inefficient and impractical, this why in these kinds of jobs each team/person is assigned to do something specific. So that progress and ideas won't be conflicting each other real time and they just piece together the puzzle afterwards. If it doesn't work, discuss the things that were wrong or could be improved and do it all over.

 

But if you're creating this network of workstations to create a powerful RENDERING machine. Then that would make sense since this method has been used for decades.

 

So if it's for RENDERING it's the best idea(if you have not so powerful machines). But using it to edit the same file, per se a model of a Raid Boss in real time across different machines it would be better to assign each team/person a job then piece the puzzle together afterwards.

 

edit: the right term would be a Render Farm

 

Using one overkill PC with a multiseat config is not that popular nowadays(well at least not in my country). It's much better using different less powerful machines and creating a Render Farm when needed.

|CPU: Intel i7-5960X @ 4.4ghz|MoBo: Asus Rampage V|RAM: 64GB Corsair Dominator Platinum|GPU:2-way SLI Gigabyte G1 Gaming GTX 980's|SSD:512GB Samsung 850 pro|HDD: 2TB WD Black|PSU: Corsair AX1200i|COOLING: NZXT Kraken x61|SOUNDCARD: Creative SBX ZxR|  ^_^  Planned Bedroom Build: Red Phantom [quadro is stuck in customs, still trying to find a cheaper way to buy a highend xeon]

Link to comment
Share on other sites

Link to post
Share on other sites

I totally misunderstood the question then.  You have several other PCs connected to the main one.... not just peripherals then?

Link to comment
Share on other sites

Link to post
Share on other sites

If this is for a professional solution your best bet is probably going to be a VM cluster. One system running dual E5's with 128-256GB of RAM running HyperV - smaller PC's then act via RDP to access virtual Windows 7/8 Pro systems each allocated 4 cores, 16-32GB RAM ect. This can even be done with thin clients depending on the scenario.

Link to comment
Share on other sites

Link to post
Share on other sites

Okay here is the idea:-

1) U make a really powerful over-kill PC.

2) Connect several Stations( each including a Monitor, Mice, Keyboard & USB Hub).

3) Have multiple users access and work on things at the same time from their respective stations.

4) HOW DO I MAKE THIS HAPPEN???

 

Note:-

The operating system has to be Windows as these are 3D design workstations using software that only works on Windows. 

sharing storage is no problem. what you're talking about would require a virtual machine for each user and probably a thin client to connect to. this is the sort of thing Kiosks do, it's more worth it to just build mid range PCs for each user.

Link to comment
Share on other sites

Link to post
Share on other sites

Impossible

IOMMU, totally possible

 

 

Okay here is the idea:-

1) U make a really powerful over-kill PC.

2) Connect several Stations( each including a Monitor, Mice, Keyboard & USB Hub).

3) Have multiple users access and work on things at the same time from their respective stations.

4) HOW DO I MAKE THIS HAPPEN???

 

Note:-

The operating system has to be Windows as these are 3D design workstations using software that only works on Windows. 

so,

step 1) ensure your hardware is AMD-Vi/Intel VT-d enabled. if amd, easy. its got like a 99% chance of having it. intel, check ark

step 2) ditch windows as the base OS. its time to dive deep into linux

step 3) you need libvirt, qemu-kvm, and virt-manager

step 4) setup linux host for PCI-Stub. you need to have various GPUs claimed by PCI-Stub.

step 5) create windows VM, get it set up and working. then clone it.

step 6) add devices to VM. GPU that is claimed by PCI-Stub, keyboard, mouse. remove VNC display, remove VGA display.

step 7) start VM back up, and it should now have its own keyboard,mouse, GPU, etc.

rough instructions, but there you go. multiple GPU accelerated VMs on a single power house system

Link to comment
Share on other sites

Link to post
Share on other sites

Okay sorry about not being very clear at first. This is my first ever forum post btw. This is the scenario. My friend has a makespace with things like 3d printers, laser cutters, can Mills and green screens. People who work on these machines often have to bring design files from their homes because we don't have any decent PCs at the makerspace. I thought we could save money by having a really powerful PC with several peripheral stations. This would allow people to work on their individual projects without the need of several workstation grade PCs. Windows is the only option because we use things like Corel, Autodesk & Adobe. Pls keep your answers coming. If it is cost effective we might actually do this. Thanks for the help so far.

Link to comment
Share on other sites

Link to post
Share on other sites

Also if you have any hardware recommendation Pls let me know. We need to look at the most cost-effective solution. Also take into account cost of maintenance.

Link to comment
Share on other sites

Link to post
Share on other sites

Okay sorry about not being very clear at first. This is my first ever forum post btw. This is the scenario. My friend has a makespace with things like 3d printers, laser cutters, can Mills and green screens. People who work on these machines often have to bring design files from their homes because we don't have any decent PCs at the makerspace. I thought we could save money by having a really powerful PC with several peripheral stations. This would allow people to work on their individual projects without the need of several workstation grade PCs. Windows is the only option because we use things like Corel, Autodesk & Adobe. Pls keep your answers coming. If it is cost effective we might actually do this. Thanks for the help so far.

 

honestly you have to do some comparison work. how complex are these designs? could they be ran on something like the 860K from AMD with a GTX 960 or some other cheap card. i dont much about what kind of hardware is really required to drive these machines, but it is possible to have multiple video cards in a single system and have multiple "workstations" based on that one set of hardware.

Link to comment
Share on other sites

Link to post
Share on other sites

As far as I know, you can do this with either VMware ESXi or Linux with KVM. I'm not familiar with the ESXi method. See this.

 

For KVM this is probably the best source of information. Read the first post to the get the idea, and the last 10-20 pages or so for updated methods. You will want to use libvirt with OVMF.

I believe the ideal hardware setup would be an E5 or better Xeon, server motherboard and AMD graphics cards. You won't need any kernel or other software patches.

 

This was my passthrough setup for a while and worked very well:

E5-2630 v2

Supermicro X9SRL-F

GTX 780ti

Link to comment
Share on other sites

Link to post
Share on other sites

As far as I know, you can do this with either VMware ESXi or Linux with KVM. I'm not familiar with the ESXi method. See this.

 

For KVM this is probably the best source of information. Read the first post to the get the idea, and the last 10-20 pages or so for updated methods. You will want to use libvirt with OVMF.

I believe the ideal hardware setup would be an E5 or better Xeon, server motherboard and AMD graphics cards. You won't need any kernel or other software patches.

 

This was my passthrough setup for a while and worked very well:

E5-2630 v2

Supermicro X9SRL-F

GTX 780ti

how did you get the 780 Ti to passthrough without throwing error 43? or did you do the work around patches? i tried to pass in a 970 to a windows VM but didnt try any of the work around patches.

Link to comment
Share on other sites

Link to post
Share on other sites

how did you get the 780 Ti to passthrough without throwing error 43? or did you do the work around patches? i tried to pass in a 970 to a windows VM but didnt try any of the work around patches.

I know of two possible solutions:

 

1. Use an older version of the driver (340.xx i think). This may not be possible with the 970.

2. Remove hv_time and all other hyper-v flags. I recall hearing that newer versions of qemu may enable hv_time by default so it may need to be disabled explicitly.

 

I think you unfortunately take a 10% or so performance hit but better than not working.

Link to comment
Share on other sites

Link to post
Share on other sites

Okay here is the idea:-

1) U make a really powerful over-kill PC.

2) Connect several Stations( each including a Monitor, Mice, Keyboard & USB Hub).

3) Have multiple users access and work on things at the same time from their respective stations.

4) HOW DO I MAKE THIS HAPPEN???

 

Note:-

The operating system has to be Windows as these are 3D design workstations using software that only works on Windows. 

 

Impossible

 

 

Possible, virtual machines with remote connections from weaker PCs over LAN.

 

It is possible in other simpler ways, using Windows MultiPoint server on the main overkill PC, and then connecting something called Zero Clients, they look like the following, and would allow you to produce the following setup.

HP-t150t100.jpgconnect-many-box-300x219.png

There are multiple models, some cheaper than others, the cheapest model is I believe about $20 each and then they is also a more expensive model which has 4 USB ports, no PS/2 and it's own ethernet jack. However there is no need for the more expensive model unless you won't be using PS/2 or you need more USB ports. These thin/zero clients also only connect via regular USB ports. Although the software is expensive, and you need to buy individual CALs or user codes I believe.

 

EDIT: Also this is probably the most cost effective.

 

Gaming Rig - Excalibur - CPU: i5 6600k @ 4.1GHz, CPU Cooler: Hyper 212 Evo, Mobo: MSI Gaming M3 RAM: 16GB Corsair @2400MHz, GPU: EVGA 1060, Case: NZXT Phantom Full Tower (Red)

My Virtualization Server - Dell R710: 2x X5570s @ 2.93GHz with 32GB DDR3 RAM [Web Server, OSX, Plex, Reverse Proxy]

I love computers, gaming, coding, and photography! Be sure to quote me so I can respond to your post!

Link to comment
Share on other sites

Link to post
Share on other sites

I know of two possible solutions:

 

1. Use an older version of the driver (340.xx i think). This may not be possible with the 970.

2. Remove hv_time and all other hyper-v flags. I recall hearing that newer versions of qemu may enable hv_time by default so it may need to be disabled explicitly.

 

I think you unfortunately take a 10% or so performance hit but better than not working.

10% hit on performance is acceptable with VMs. i might have to look into disabling hyper-v flags. or see if xen works with it

Link to comment
Share on other sites

Link to post
Share on other sites

10% hit on performance is acceptable with VMs. i might have to look into disabling hyper-v flags. or see if xen works with it

My understanding is that this is due to nvidia being an ass and artificially disabling the driver if it sees virtualization flags and you aren't using some high end Quadro/Tesla. The platform may not matter.

 

I though of one more thing. I believe this flag needs to be set too:

  <features>    <kvm>      <hidden state='on'/>    </kvm>  </features>
Link to comment
Share on other sites

Link to post
Share on other sites

 

My understanding is that this is due to nvidia being an ass and artificially disabling the driver if it sees virtualization flags and you aren't using some high end Quadro/Tesla. The platform may not matter.

 

I though of one more thing. I believe this flag needs to be set too:

  <features>    <kvm>      <hidden state='on'/>    </kvm>  </features>

what version will those flags work on? like minimum. linux mint (distro i used for it) only had 2.0.0 in repos. i built 2.2.something from source and then libvirt broke so...

 

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×