Jump to content

Gerr

Member
  • Content count

    564
  • Joined

  • Last visited

Awards


This user doesn't have any awards

1 Follower

About Gerr

  • Title
    Member
  • Birthday 1972-08-04

System

  • CPU
    i7-4790k @ 4.6Ghz
  • Motherboard
    ASRock Z97
  • RAM
    16GB G.Skill Ripjaws-X 2133 CL9
  • GPU
    MSI GTX 970 Gaming 4G
  • Case
    FD ARC Midi R2
  • Storage
    2x250GB(RAID-0) Samsung 850 EVO
  • PSU
    Rosewill Capstone 750W
  • Display(s)
    Asus VG248QE
  • Cooling
    Noctua NH-D15
  • Keyboard
    Razer DeathStalker
  • Mouse
    Razer DeathAdder
  • Operating System
    Windows 10 Preview

Profile Information

  • Gender
    Male
  • Location
    Phoenix, AZ
  • Occupation
    Network Engineer

Recent Profile Visitors

844 profile views
  1. I ended up getting a used system off of eBay that should suit my needs... Xeon E5-2680 v2 (10c/20t-2.8Ghz) CM 212 EVO cooler x79 ATX mobo - unknown make/model 16GB DDR3-1866 ECC RAM - unknown brand. 240GB PNY SSD 1TB WD Black HDD Blu-Ray Combo drive GTX 1060 3GB by Asus Corsair TX650W PSU - unknown if Gold or Bronze. Corsair 100R case Windows 10 Pro 64 $625 shipped. I figure I'll sell the GTX 1060 and use that money to buy more RAM, 32GB min, trying for 48GB or 64GB, pending on mobo capacity. Outside of that, I thought it was a good deal and should be a good VM & Handbrake system.
  2. I have a spare copy/license for Windows Server 2016 Essentials. Can I use that for Hyper-V? I know nothing about the other options.
  3. I would love a Ryzen 7, but the prices of DDR4 are way too high, so sticking with a DDR3 solution at the moment.
  4. I am more worried about CPU cores than RAM. With 32GB RAM, I can reserve 8GB for the host giving 24GB for the VM's, which is 6x4GB, 12x2GB, or 24x1GB VM's. I think that should be plenty. When it comes to cores, I only have 4 of then, 8 threads with HT on. If I assign 1 core per VM and have 6 VM's running, I have already over provisioned my CPU cores. If I need a couple VM's that require 2 cores, I could quickly reach a bottleneck if those VM's are CPU intensive. Granted that my fast 3.6Ghz CPU will help with that, it's still a concern.
  5. I need to obtain several IT certifications over the next 2 years and most of my studying will be self studying at home. I need a home lab server that I can use to virtualize the different environments I will encounter. I will likely have to have a good 4-8 VM's running at the same time and while I'll try to use Linux as much as possible, some Windows 10 and even Windows Server VM's will be needed. My main question is if my current system is powerful enough, or if I should sell it and get something more powerful? My current system is a Xeon E3-1271v3(4c8t-3.6Ghz) CPU on an C226 mobo with 32GB ECC RAM(max). If that is not powerful enough, then for around the $550 range I could pick up a used HP Z420 with an 8-core 2.6Ghz Xeon E5 CPU and 64GB RAM or a HP Z620 with two 6-core 2.2Ghz Xeon E5 CPU's and 48GB RAM. I know those systems are actually older than my current one, but they are more powerful. I am up for other recommendations, but not a fan of rack mount systems as they are hot and loud.
  6. I need to build a lab server on the cheap. It doesn't need to be powerful, but I do need at least 8 cores and at least 32GB RAM as I will be doing a lot of virtualization with it, and would prefer 48-64 GB if possible. It won't be turned on except when I need it, so only a single PSU is needed. I have spare hdd's, so don't need those. Would also like a system that doesn't sound like a turbine engine when turned on. I would also prefer a tower system vs a rackmount system, but either would work. My initial thoughts would be to get an old Xeon system from the Sandy or Ivy bridge generation and get a motherboard that uses registered DDR2/DDR3 RAM as it's cheap. I would also like to keep the total cost at or under $500 if possible, but could go a little higher if required. Any suggestion on parts or a used system?
  7. My current bedroom system, which I use as a backup gaming system, has a i5-7500 CPU on a H270 mobo and a GTX 1060 6GB. I recently replaced the CPU in my main gaming rig and now have a spare i5-6600k that can OC to 4.6Ghz @ 1.32v. My initial plan was just to sell that CPU, but I thought of another option. I have a spare NZXT x41 AIO, so I thought I could buy a used Z170 mobo, put the 6600K into it, and replace my current 7500/H270 cpu/mobo combo with the 6600K/z170 combo which can overclock. Yes, I know the 6600K/Z170 is older tech, but it can overclock where the 7500 can't. Plus Intel secretly removed RST from the H270 chipset, so I can't use my 64GB SSD drive as cache for my 7200RPM game drive. And yes, I notice a difference! My question is would I get a good jump in FPS by switching from a 7500/H270 combo running at stock speed to a 6600K/Z170 combo running at 4.6Ghz, both with a GTX 1060 6GB? Just not sure it's worth the money or effort.
  8. Pfsense on Freenas

    What I meant is that I have read that for ZFS to properly work, it wants direct access to the drives from what ever OS is implementing it. Thus what every OS you are using, FreeNAS, Linux, etc, to try to avoid virtualization.
  9. Pfsense on Freenas

    I thought ZFS doesn't virtualize well?
  10. $709.99 - EVGA GeForce GTX 1080 Ti (1480 MHz) US

    From what I understand, EVGA doesn't bin their chips, so any FTW card that doesn't pass at the higher clock speeds used in their FTW edition has their speeds reduced to Nvidia stock and labeled with FTW DT. You still get the higher end FTW cooler, but your speeds will be reduced. With the better cooler, GPU Boost will take the card back up to a decent level, but you are probably talking around the SC level and not FTW level.
  11. I figured I wasn't the first person to think of this, but it's nice to know that others do use this method. Just never seen a post or video about someone doing OC this way.
  12. I have been overclocking for a long time and have a new 7700K on the way, so started to think about the best way to overclock it and came up with a possible new method and want to know if my logic is sound and what people think. Normally, you would research the average OC a CPU hits, set your system there, and slow bump voltage & frequency until you reach either your thermal limit or it becomes unstable. My new way is different. Basically, since most CPU's are either voltage or temp limited, why not set those first and then bring up the frequency to see where it's stable limit is? Remember, it's voltage and not frequency that determines heat, so if you have those maxed out already, you could then determine the max stable frequency limit of your CPU, with some play in the temp and/or voltage limits as needed. Step #1a = after doing some research, determine both a soft cap and hard cap for your temps and voltages. For my 7700K, my soft caps will be 82* & 1.36v and my hard caps will be 86* & 1.4v. *there will be multiple factors in play here including your level of cooling, your comfort level threshold for both temp and voltage, and the type of stress test you use. With good cooling, you probably will reach your voltage limit before you hit your temp limits, but every CPU is different and that's why you have caps for both. Also, the type of stress test you use will make a difference as some tests push the CPU to an unrealistic level, so you have to account for that. I use OCCT's standard test, so I consider that max real world temps. But other tests like Intel's Burn In and the newer Prime95's will jack the temps up, so pick your temp caps accordingly. Step 1b = determine a short, medium, and long set of stress test time frames. My times will be 30m, 2h, 8h. *The short tests will be frequently used to determine quick temp & stability checks. The medium tests will be used to determine stability before changing a different setting. The long test will verify your final OC setting. I picked a 30m short test as that's enough time to see if a change is somewhat stable or not. My 2h medium test is enough IMO to check for decent stability. My 8h long test is fine as I never game longer. What you pick will determine how fast you reach your final OC. A safer but longer set would be 1h/4h/24h, but I figure that if my OC passes 8h, if I start to encounter come crashes, I can make small adjustments at that time. Step #2 = lock your CPU's frequency at stock and raise the voltage to your soft cap, locking it there, and see what your temps are on a short test. If they are too high, back the voltage down in small increments. If you still have headroom in temps, it's up to you if you want to start bumping your frequency or push your voltages towards your hard cap. Step #3 = once you have pushed your temps and/or voltages to around your soft caps, start increasing the CPU frequency until you encounter instability using short tests. After hitting a unstable frequency, back it down one notch and run a medium test. If it fails the medium test, you can either reduce the frequency one more level or increase the voltage a bit, pending you have cap room on the voltage and temps. Step #4 = once you have a stable OC that passes your medium test, you can start to make subtle tweaks before hitting the OC with your long test. I like to reduce my voltage until I hit an unstable setting using short tests, the bring it back up a bit and use my long tests to find the final stable OC value. If it fails, just bump the voltage up in small increments until it passes a long test. Step #5 = once you have your final OC values, go back into the BIOS and set your voltage to adaptive if you so desire so the CPU isn't running at max voltage all the time. Example of how this OC method MIGHT go on my 7700K... I have already set my soft and hard caps for temps and voltages, those being 82*/86* and 1.36v/1.40v. I have also set my test time frames to 30m/2h/8h. I will be using a 240 AIO in push/pull mounted on the front of the case as intake, so I should have good cooling. First I will lock my CPU's frequency to 4.4Ghz on all cores and then lock my CPU's voltage at my voltage soft cap of 1.36v. I then boot into Windows and run a 30m stress test and see where my temps are. If my temps are too high already, I will back the voltage off, but I presume my temps will be under my soft cap due to my good cooling, so I don't foresee being temp limited. I would then go back into the BIOS and bump the frequency from 4.4Ghz to 4.8Ghz, which is an average OC for this CPU and then boot into Windows and run a 30m test for stability. Since my voltage stayed the same, I shouldn't see any rise in temps, so I am only checking for stability. I would presume I am stable unless I get a really bad chip. Now I go back into the BIOS and raise the frequency 100Mhz at a time and run short tests until I encounter an unstable frequency. Lets say I am stable at 5.0Ghz, but unstable at 5.1Ghz. I would now test this 5.0Ghz frequency on a medium test. If it fails the medium test, I can either back it down to 4.9Ghz and redo the medium test, or slowly increase the voltage until it passes at 5.0Ghz. In this example, lets say it passes. Remember that your hard caps are levels you would accept running your CPU at, but would prefer if they were lower, so don't be afraid to use them. So in this case, I would set the frequency back to 5.1Ghz and set my voltage to my hard cap to see if it's stable. If it isn't, then 5.0Ghz would be my max stable frequency and I would start to reduce the voltage blow my soft cap to see how low I can go while remaining stable. However, if 5.1Ghz is stable at my hard cap voltage and my temps are still within limits, I would next start to slowly reduce the voltage to determine the lowest voltage setting where 5.1Ghz is stable. For this example, lets say 5.1Ghz is not stable even at 1.4v, so I would reduce the frequency back to 5.0Ghz and reduce my voltage to below my soft cap, lets say 1.35v and run a short test. I would continue to reduce the voltage until I hit an unstable level. For this example, say it's unstable at 1.34v, but stable at 1.35v. I would then try 1.345v and run a long test. If it passes the long test, that's my final OC value, but if it fails, I would bump the voltage back to 1.35v run it at a long test, which I presume it would pass. And that would be my final OC value. Now I would just to back into the BIOS and set the voltage to adaptive so it only gets this higher voltage when Turbo Boost is on. While the final part of this method really isn't any different the normal OC method, the early part is and I think this method will shorten the time it takes to reach your final OC value and might get you a higher value since you are testing your voltage/temp limits first. Thoughts?
  13. I ended up researching Windows ReFS and deciding that while not as good as ZFS, it was good enough for my needs, so going an all Windows 2016 Server Essentials build.
  14. Which level of 1080 Ti?

    Are there any 1080 Ti's that are binned that don't cost a fortune? And yes, I do plan on OC'ing. Wouldn't the Strix OC do better in that regard as chips not fast enough to get to the OC speeds just be sold as the regular Strix?
  15. I will be buying a new 1080 Ti this Xmas, but I am trying to decide which level of 1080 Ti to get. What I mean by level, using EVGA's line up for example, for around $750 you can get the SC2, or for around $800 you can get the FTW3 or SC2-Hybrid, or for around $850 you can get the FTW3-Hybrid or the FTW3 Elite. Are the gains, in FPS, sound, & thermals, worth the level jumps and where would be the sweet spot? To sum up, I am only looking at AIB cards with 2-3 fans, or single fan hybrids, so no FE blower type cards. The cards I am considering at each level are: Level 0 1080 Ti cards, all around $700... MSI Armor - would replace cooler with a NZXT G12(buy) & X41 AIO(already have). Level 1 1080 Ti cards, all around $750... EVGA SC2 Asus Strix non-OC Level 2 1080 Ti cards, all around $800... EVGA FTW3 EVGA SC2-Hybrid Asus Strix OC MSI Gaming X Trio Level 3 1080 Ti cards, all around $850... EVGA FTW3-Hybrid EVGA FTW3 Elite White Asus Poseidon My system's color scheme is white/black, so any cards that focus on other colors, like the MSI Gaming X regular, are not being considered. I am also not a fan of Gigabyte or Zotac, so avoiding those cards. Anyway, each level just is in the ballpark of $50, but are those jumps worth it? Right now, I am leaning towards either the SC2(L1) or Asus Strix OC(L2), but the 12Ghz memory on the FTW Elite(L3) is interesting as well as are some of those hybrid cards. Thoughts?
×