Jump to content

I am working on building a new PC for work first, and play second. I need a few bits of advice. I am not looking for a full build order of all parts, just advice on specific aspects of the parts.

 

I work a lot with VMs and on a day to day basis, I have sometimes up to 8 different virtual machines running at the same time on a single workstation compiling and processor intensive tasks running on them.

 

Current system requirements:

  1. >=64GB of Memory - It's probably overkill at this point, but I go through so much memory when working, I don't want to ever have to worry about it.
  2. Graphics card setup that can support a minimum of 5 monitors, possibly 4k mixed with 1440p monitors for output. For gaming, it will only ever run on a single 4k or 1440p monitor for resolution/performance. Monitors are arranged in a stacked arrangement, not surround.
  3. I do not do much video editing or rendering to need FirePro/Quadro grade graphics processing for any 3d/2d rendering operations. Some video encoding, but not enough to warrant a huge expense on a professional graphics solution, unless it is to push the large amount of displays. If someone has argument for this setup, I would gladly entertain the idea.

 

Existing hardware:

  1. PC will be built from the ground up and the only parts reused are peripherals (kb, mouse, etc) and monitors.
  2. I have a rack of lower end servers (10 linux servers) in my office. They all get used to perform longer operations to free up VM space on my workstation and for production purposes.

My first thoughts for this build is to split out work and play. Build a dual processor, high amounts of RAM for my VMs and just build a decent gaming rig for gaming and media consumption. The part that makes me hesitate is that with the amount of pixels I'm pushing through the multiple monitors, I'll need at least two higher end graphics cards not in xfire/sli to have enough inputs to plug in the monitors and run them. I currently run 2x GTX670s to push four 1440p displays, and can game without much of a problem on one of the 1440p displays. The graphics cards are not in xfire/sli or I lose display to one of the monitors (xfire/sli limits to one card which only has three outputs).

 

Questions:

  1. If I did go dual processor, does it hurt gaming and/or gaming compatibility?
  2. With this many displays, is it more sane for me to go something like an AMD FirePro W600 (~$500USD, up to 6 displays, max 4k res on each) or a W9000 (~$3400USD probably way overkill) vs 2x780ti (~$1400USD) or 2xTitan (~$2kUSD) configuration?
  3. If I get up and over 64GB in memory on some dual processor solutions, am I starting to get into where I'll need ECC memory and the costs associated with that?

I realize some of your first thoughts on this would be to expand out my rack of servers to allow me to reduce the cost of running so many virtual machines. This is a possible solution, but I've found the maintenance and the time spent having to run back and forth when formatting them costs me more than simply building a bigger PC to handle them in VMs. The cost of increasing RAM and Cores is less than multiple servers and their maintenance.

 

Another thought would be to simply build a larger server with these specs and farm out the VMs to that, lowering the specs I need on my operating machine. That is also a solution I've entertained, but my response to it is simply that I enjoy having all that RAM and cores at my fingertips rather than over a 1gbe or 10gbe LAN connection. Not a great explanation, but splitting it out would eventually lead me to have two really expensive systems (server [ram/cpu] and workstation [ram/graphics]) and then possibly another gaming setup in the other room.

 

So, can I accomplish this in one awesome rig? Should I split it out to workstation/gaming? Anything I've completely missed that you think should be addressed?

Link to comment
https://linustechtips.com/topic/106370-workstation-first-gaming-second-advice/
Share on other sites

Link to post
Share on other sites

Buying 1 system which is an all-rounder high end system will enable you to spend more on parts that you think you may need versus. buying two individual system where the balance may not be even 50:50 ratio so to speak.

Not a great explanation, but splitting it out would eventually lead me to have two really expensive systems (server [ram/cpu] and workstation [ram/graphics]) and then possibly another gaming setup in the other room.

Very true you're justification but in the end it's all up to your final decision.

 

ECC memory will definitely cost a tonne even for ram but if ECC memory isn't feesible get ram that is highly regarded for reliability if broken like Kingston or some other brand.

Just realized your acronym of VM's is --> virtual machines ( i'm stupid )

 

Running Geforce graphics cards limits you to 3+1monitors in a surround/multi-monitor set-up. But using AMD graphics cards you'll definitely get away with using 6 displays all at once for virtual machines.

Please become a member of the Linus Tech Tips forum, keep writing smug remarks & let us love you. Peace out.


<| Project M13 & Silverstream. Other DIY projects |>

Link to post
Share on other sites

  1. It shouldn't affect gaming one way or the other.

A relatively low-end workstation gpu or two to drive most of your displays and a high-end gaming gpu driving a gaming monitor.

Would ECC be such a bad thing, other than the cost? The added reliability might make it worth it for production. You may also want to consider a platform like LGA2011. Q3-2014 should see the release of Haswell-E cpu for this socket. The rumor is they will include 8 core / 12 thread options which may meet your needs.

80+ ratings certify electrical efficiency. Not quality.

 

Link to post
Share on other sites

 

  1. It shouldn't affect gaming one way or the other.
  2. A relatively low-end workstation gpu or two to drive most of your displays and a high-end gaming gpu driving a gaming monitor.
  3. Would ECC be such a bad thing, other than the cost? The added reliability might make it worth it for production. You may also want to consider a platform like LGA2011. Q3-2014 should see the release of Haswell-E cpu for this socket. The rumor is they will include 8 core / 12 thread options which may meet your needs.

 

Waiting for Haswell-E myself but by the time we hit Q3 of 2014 Intel might delay it to Q4 but thats a butt tonne of time to wait. I like your justifications so here are some cookies (#)(#)(#)(#)(#)(#)(#)(#)(#)

Please become a member of the Linus Tech Tips forum, keep writing smug remarks & let us love you. Peace out.


<| Project M13 & Silverstream. Other DIY projects |>

Link to post
Share on other sites

It sounds like a dual processor has a lot of benefit for me in a professional sense, and not much negative in a casual gamer sense. Anyone have any more details backing or refuting this?

 

As for running two different graphics cards on one motherboard: I've had a lot of issues with graphics cards. About 4-6 years ago, it was difficult to get a lot of games and/or graphics cards to properly use all of the monitors. They would disconnect, drop back into SLI/xfire randomly, the second card would completely disappear from the hardware list and not allow me to turn off/on xfire modes. Just a lot of problems. I've moved to nVidias graphics cards after many years of being an AMD guy, and I'm not sure I'll go back any time soon. Their drivers are much more stable and multi-monitor setups that don't use SLI but do use multiple graphics cards don't suffer from a lot of the problems I've experienced with AMD.

 

That said, keeping my system homogeneous with a single brand, type and model of graphics card is extremely important to me, even if it's "suppose to" work. I'd rather be safe with dual Titans or 780ti's then tearing my hair out dealing with different types, drivers, etc. for multiple cards. I have never used any of the quadro or firepro workstation cards before, but I would imagine they're *much* better at handling large amounts of displays. Their gaming chops aren't up to snuff though vs what I could get in a consumer grade gaming graphics sense. Since I will only utilize them for a lot of monitors and not 3d/2d rendering, I feel like I wouldn't be getting my monies worth.

 

Instead of ECC memory, since daily offsite backups are more important to me than immediate errors in data, I would probably side more towards expanding my NAS or a PCI-e SSD for the working/running copies of virtual machine images (I/O speed).

My reasoning for ECC memory is this, and no, it's not a bad thing... just an expensive thing. I was just curious if there's some amount of memory where the chance of memory errors (lots of ram, more chances) starts weighing into a much higher crash/error rate for the system as I routinely have processes running overnight for calculations and compilations:

 

 

 

“Do I Need to Get ECC Memory?”

To answer the question above, we have to figure out where memory errors come from first. There are two major causes of the so-called "soft" errors:
  • Naturally occurring radioactive isotopes (which emit alpha particles)
  • High energy cosmic rays from supernovas
Both of these events can change the value of data stored in a memory chip. These errors are called "soft" errors because they can be repaired by correcting the value of the memory bit, which is exactly what ECC does.

Chances for a single-bit soft error occurring are about once per 1GB of memory per month of uninterrupted operation. Since most desktop computers do not run 24 hours a day, the chances are not actually that high. For example, if your computer (with 1GB of memory) runs 4 hours a day, the chances of a single-bit soft error happening (when your system is running) is about once every six months. Even should an error occur, it won’t be a big issue for most users as the error bit may not even be accessed at that time. Should the system access the error bit, this little error won't result in a disaster either – the system may crash, but a restart of the system will fix that. That's why ECC memory is not a necessity for most home users.

Things are very different when it comes to workstations and servers. To begin with, these systems often utilize multi-gigabytes of memory, and they usually run 24/7 as well. Both of these factors result in increased probability of a soft error. More importantly, an unnoticed error is not tolerable in a mission-critical workstation or server – a system crash is only the smallest of worries. What really matters is the erroneous data itself – you can imagine the issues that can arise as a result of a soft error in bank systems or a flight control computer system. Therefore, ECC memory is definitely required for mission critical applications.

Finally, if you do need ECC memory, you'll have to buy a motherboard that supports ECC memory modules in addition to the ECC memory modules themselves. Without motherboard support (or memory controller support, to be more accurate), the ECC memory module is effectively the same as non-ECC memory.

From: 
http://www.newegg.com/product/CategoryIntelligenceArticle.aspx?articleId=126  
Link to post
Share on other sites

Quadros/Firepros won't handle many monitors any better than GTX/Radeons

What kind of VM's are you running?

If you ever need help with a build, read the following before posting: http://linustechtips.com/main/topic/3061-build-plan-thread-recommendations-please-read-before-posting/
Also, make sure to quote a post or tag a member when replying or else they won't get a notification that you replied to them.

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

×