Page 5 of 16 FirstFirst ... 3 4 5 6 7 15 ... LastLast
Results 41 to 50 of 152
  1. #41
    TeamROG Moderator Array xeromist PC Specs
    xeromist PC Specs
    Laptop (Model)Dell Inspiron 15 7567
    MotherboardMSI x470 Gaming Plus
    ProcessorAMD 2600X
    Memory (part number)16GB Crucial Ballistix Elite 3600
    Graphics Card #1ASUS GTX 1080 Strix
    MonitorBenQ BL3200PT
    Storage #1Intel 600p NVMe
    CPU CoolerWraith
    Casecustom Antec 900
    Power SupplyCorsair HX1000
    Keyboard Logitech Orion Spark
    Mouse Logitech MX500
    Headset Plantronics 777 with Oregon Aero upgrade
    xeromist's Avatar
    Join Date
    Jul 2010
    Reputation
    344
    Posts
    8,023

    Quote Originally Posted by powerhouse View Post
    @Nodens: I hope this post doesn't get censored. I tried to post 3 links to 3 different forums which explain why VT-d is becomming popular.
    No worries. New posters linking outbound get put into our mod queue so we can check it out manually. It might take a few hours but if your post isn't spam it will show up as soon as we approve it.
    * Support disease research with Folding@Home *

    < < < Click the drop-down above my avatar for my PC specs!

  2. #42
    Super Moderator - Coderat Array Nodens PC Specs
    Nodens PC Specs
    Laptop (Model)ASUS GL702VW-GC004T
    MotherboardRampage V Extreme Edition 10
    ProcessorIntel 3930K C2
    Memory (part number)CMZ16GX3M4X2133C11
    Graphics Card #1ASUS STRIX-GTX970-DC2OC-4GD5
    MonitorASUS VG278H
    Storage #13x OCZ Vertex 3 240 (RAID5 - LSI 9361-8i)
    Storage #23x WD2003FYYS (RAID5 - LSI 9361-8i)
    CPU CoolerCorsair H100i + 4x Noctua NF-P12
    CaseCoolermaster HAF-X
    Power SupplySeasonic X-1250
    Keyboard Razer Anansi
    Mouse Logitech G502
    OS Windows 10 Pro x64
    Network RouterLinksys E4200v2 (running TomatoUSB/Toastman firmware)
    Nodens's Avatar
    Join Date
    Mar 2012
    Reputation
    266
    Posts
    4,389

    Quote Originally Posted by -Nax- View Post
    But understand me, i was using a good old 3820 with no problems with vt-d and changed it for a 4930K to gain 2 more cores and i lost vt-d.
    I understand perfectly my friend. But thing is the IB-E chips are new and once the issue is sorted out you will get it again. Next UEFI version should fix it from what I hear.

    It's not against you or ASUS and i know that vt-d support is a bonus which should'nt work with X79 chipset, but X79 is an old chipset, cards should not be so far from end of support and i am really afraid that my asus rampage iv gene finally never release a bios with vt-d.
    X79 is very far from end of support. Do not expect ASUS to drop support for these boards, it won't happen. Until Haswell-E which is more than a year away X79 is and will be Intel's flagship chipset. And even after the realease of X99, X79 will still be on top notch support for a while. Unlike other OEMs, that I won't name here for political correctness, ASUS ROG boards get constant updates. This and the engineering+manufacturing quality is what drew me to ROG, personally, in the first place.

    IB-E is new, it's hard to understand for me, Asus is not a small market actor, they should have Ivy-E ES for tests since long months (as they should already have Haswell-E ES now). Is it so hard to allow VT-D ? You give a beta bios for asus R4IV and it is a very good point, but i know that this is the reference working mainboard that's why i'am afraid for my RIVGene and RIVBE support.
    They did of course have ES CPUs for R&D but you are underestimating the R&D effort needed, the multitude of things on the todo list and the time needed to Q/A everything. A feature such as VT-d, is not on the top of the priority list due to the very small number of people it affects. They are fixing it though.

    For waiting for RIVBE, it was very long to wait, it was announced mi-september, news said that it should be release end of october, finally mid-november in US and in Europe mid-december ... Asus X79-Deluxe was annonce end of august with stock mid/end of september ! That's why i tought it was so long !
    Well there were several wrong expectations on release dates that came out of user to user rumors and speculation. The only real solid date on release of BE that was posted, was by Raja regarding the USA region and that had a "probably" attached to it. Still he was accurate enough with that date. The rest was all user generated and propagated.

    Haswell has a better chipset and better CPU performance, the only miss is 2 cores and less Pci Express Lane. It's hard to have the Premiun platform but one generation late ... with the maintstream ... even if performance are only 10% better (by core).
    The only thing better on Z87 is the native USB3 support. Other than that the chipset/platform is inferior. It's dual channel and has less PCI-e lanes. Haswell is faster on clock per clock comparison but still it's 2 cores short, has less cache, is dual channel and overclocks much worse unless you delid the chip. Overall the X79 is still faster. As far as being one generation late, this is how the tick tock Intel cycle works. Nothing we can do about it. But the Performance line is still the performance king.

    Quote Originally Posted by powerhouse View Post
    @Nodens: You asked the question whether people really need VT-d? Or mix it up with VT-x?
    I asked people here. Because if you search the threads there are lots of people on this forum who have actually made this misconception in the past. If they did so, then I could explain to them the difference as I have in the past.

    Well, I'm the author of the "HOW-TO make dual-boot obsolete using XEN VGA passthrough" thread over at the Linux Mint forum (http://forums.linuxmint.com/viewtopic.php?f=42&t=112013), which has had some 60,000 views over the past year. Here are two other popular threads on the same subject, just for reference:
    http://www.overclock.net/t/1205216/g...irtual-machine
    https://bbs.archlinux.org/viewtopic.php?id=162768
    That's great.

    An increasing number of people want to run Linux as their main OS, but still have a need for Windows to run specific software and games. VT-d and Xen/KVM (under Linux) are the key to do that without having to dual-boot.
    I am perfectly aware of this. As I am an avid supported of Linux, FOSS and I use PCI passthrough myself. Thing is that on this forum, very few people know what VT-d/IOMMU is, how it can be used and even less are interested in Linux (evident in our Software section were Me and Zygomorphic try to help people with Linux in general). The majority of threads in the past here about VT-d were people that thought it was the same thing as VT-x or that both are needed for taking advantage of accelerated virtualization technology. This is why I am asking, because I want to help people. This is a forum wtih about 87K accounts and an average of 800 people online at any time and yet we're still 6 people discussing this with proper knowledge and use of it...

    As you can see from the responses on these threads, and the sheer number of viewers, this IS A HOT topic.
    Unfortunately it is only a hot topic on Linux communities and among very technically adept power users. Had it been actually more mainstream we would have seen more related features about this and less restrictions (eg Nvidia non-Quadro cards PCI passthrough compatibility). Something I am really looking forward to as I am really interested in this technology. It does allow me to test my games engine code in VMs using the actual video adapters instead of virtual devices. The majority of users of VT-d though are on the enterprise level and use it mainly for passing through dedicated drive controllers or NICs to VMs on VM servers. This market segment is the initial target of IOMMU technology but I hope that this will change.

    With regard to the Asus Sabertooth X79 and some other boards, there is a confusion about which BIOS (if any) supports VT-d. I'm owning the Sabertooth X79 and a i7 3930K CPU and VT-d works with BIOS 1203. But here http://vip.asus.com/forum/view.aspx?...Language=en-us users report that VT-d support is broken in BIOS release 2003 and 2104.
    I do not have the specific board so I can not tell you more about its UEFI versions personally. But do understand that big revisions of the UEFI (like going from 1xxx to 2xxx, from 2xxx to 3xxx etc), change so much, that things may break or need disabling until they're Q/A'd. It's not just VT-d. For example from the late RIVE 4xxx have had a few reported issues but they're getting ironed out with each now version and soon proper VT-d support for IB-E chips will be provided as well.

    Yet, on my inquiry to Asus tech support I received a reply that says that VT-d is supported. What, are those users lying?
    I have no idea. There is always the factor of user error as well. What I can personally verify is that VT-d works perfectly fine with Xen on the Rampage IV boards when using Sandy Bridge-E CPUs and that it currently doesn't work for Ivy Bridge-E CPUs. I can also verify that the beta version I posted for RIVE works with IB-E chips as well.

    I have noticed that Asus has recently stepped up its efforts to iron out bugs that affect VT-d or AMD-V (AMD boards). However, in order to endorse specific Asus motherboards I need definite proof that they support VT-d or AMD-V (IOMMU). I would also like to test a newer BIOS on my board, but I don't want to brick my computer.

    Is there any way you or others here can help and put me in touch with those who do the VT-d testing (or the BIOS devs)?
    No and I'm afraid no one can help you in this regard. As far as I know the BIOS/UEFI engineers can not be contacted directly and they don't have any time to spend doing anything else than actually work on UEFI/BIOS. Even I, that I have been messing around with the UEFI etc and that I am collaborating with ASUS for RealBench development, have no direct way to contact or report anything to them.
    I am also sure that there is not a specific team that tests VT-d. VT-d is tested as part of the normal Q/A procedure just like any other feature.

    If you have any specific question perhaps you should post it here and someone @ASUS may get an answer for you.

    I would also be willing to do some BIOS testing, as long as I'm assured I will get a BIOS version I can fall back on where VT-d works. (Just to make it clear, I won't be able to run Windows or any other VM if VT-d fails, essentially turning my PC into a brick.)
    Public betas are available for download at the support site. Sometimes here you will find unofficial betas from Shamino that address a serious problem or are tweaked for benchmarking (only for the ROG branded boards). Internal betas are not available for anyone other than perhaps select major clients via other channels.

    I appreciate that Asus is taking this serious now and works on enabling VT-d. Is there a way to get an update on VT-d / AMD-V support across different Asus motherboards? That is, which boards and which specific BIOS releases have been tested for IOMMU support (VT-d or AMD-v)? Especially your desktop boards (I assume your server boards support IOMMU).
    I can not speak on behalf of ASUS but I imagine this to be hard to maintain and considering the actual market percentage interest in this, I do not think that anyone would bother to be quite frank with you. Rule of thumb is that VT-d should work on the latest boards with CPUs that support it, unless there's a temporary issue in UEFI or regression.
    Last edited by Nodens; 12-19-2013 at 07:15 PM.
    RAMPAGE Windows 8/7 UEFI Installation Guide - Patched OROM for TRIM in RAID - Patched UEFI GOP Updater Tool - ASUS OEM License Restorer
    There are 10 types of people in the world. Those who understand binary and those who don't!

    RealBench Developer.

  3. #43
    ROG Guru: White Belt Array
    Join Date
    Oct 2013
    Reputation
    10
    Posts
    75

    Damned ... i want to be an alpha bios test users for Rampage IV Gene and Rampage IV BE ...

    Few users of VT-D, i am not sure, it can't be an hasard than VT-D isn't present on maintstream K Processor like 4770K or 3770K ... i was sure that all these feature overclock and vt-d was for enthusiast X79 users.

    Well is it the same problems with 1150/1155 asus mainboard and non K processor ?

    VT-D is not priority for asus ... long time without solution for sabertooth ...

    Should i give up and buy an Intel BLKDQ87PG with an 4771 ?
    Last edited by -Nax-; 12-19-2013 at 09:37 PM.

  4. #44
    Super Moderator - Coderat Array Nodens PC Specs
    Nodens PC Specs
    Laptop (Model)ASUS GL702VW-GC004T
    MotherboardRampage V Extreme Edition 10
    ProcessorIntel 3930K C2
    Memory (part number)CMZ16GX3M4X2133C11
    Graphics Card #1ASUS STRIX-GTX970-DC2OC-4GD5
    MonitorASUS VG278H
    Storage #13x OCZ Vertex 3 240 (RAID5 - LSI 9361-8i)
    Storage #23x WD2003FYYS (RAID5 - LSI 9361-8i)
    CPU CoolerCorsair H100i + 4x Noctua NF-P12
    CaseCoolermaster HAF-X
    Power SupplySeasonic X-1250
    Keyboard Razer Anansi
    Mouse Logitech G502
    OS Windows 10 Pro x64
    Network RouterLinksys E4200v2 (running TomatoUSB/Toastman firmware)
    Nodens's Avatar
    Join Date
    Mar 2012
    Reputation
    266
    Posts
    4,389

    Doubt something like this could ever happen. In order to have an alpha testing program it has to provide useful data in a specific time window. Otherwise it's pointless. That means that participants would have to have some kind of verifiable technical expertise in order to ensure that testing and reporting methodology is correct something that is very hard to verify. My long experience with software development shows that 90% of end user bug reports are not bugs at all but user errorrelated or external factor depended (corrupted or messed up operating systems being the number culprit). Of course with software, one can set up a dump generation and logging mechanism that allows the developer to have a clear view of what's going on. You can't do that with UEFI/BIOS. Even if 5% of the reports are bogus it creates more problems than it solves as you take time to examine those reports in order to determine they are bogus/unrelated. With hardware this is even harder. Any possible testers would also have to sign NDAs and waivers to any claims to damage of data in the case any user tried to use an alpha (which is by definition unstable) in a production system. The whole process, paperwork processing, associated risks etc makes it not worth it for the developer at all.
    All the non open source projects that provide alpha testing programs do it solely for marketing and promotional reasons (building userbase/publicity) something that UEFI/BIOS development doesn't need.
    RAMPAGE Windows 8/7 UEFI Installation Guide - Patched OROM for TRIM in RAID - Patched UEFI GOP Updater Tool - ASUS OEM License Restorer
    There are 10 types of people in the world. Those who understand binary and those who don't!

    RealBench Developer.

  5. #45
    ROG Member Array
    Join Date
    Dec 2013
    Reputation
    10
    Posts
    18

    @Nodens:

    Thanks for your very detailed reply - this is really appreciated! I accept that developers should be able to work without being bothered by users. Also the other points are understood and fine with me, except the following:

    I can not speak on behalf of ASUS but I imagine this to be hard to maintain and considering the actual market percentage interest in this, I do not think that anyone would bother to be quite frank with you. Rule of thumb is that VT-d should work on the latest boards with CPUs that support it, unless there's a temporary issue in UEFI or regression.
    VT-d is an advertised feature or at least mentioned in the manual, and I would hope that Asus tests VT-d as part of their standard Q/A process for new BIOS releases. Actually it's more about not knowing what's going on that worries me.

    EDIT 5 January 2014: I checked the manual of my Sabertooth X79 as well as the RIVE manual - VT-d is NOT mentioned. The only place Asus mentioned VT-d support was in a BIOS release note. Sorry for my mistake. However, this BIOS release note led me to believe that VT-d is available in the Sabertooth X79 I bought, and indeed it was/is available with the old BIOS I use.

    I can live with Asus not supporting VT-d, there are other manufacturers that are more committed to it. Right now buying Asus is like playing roulette which is a little unnerving.

    A suggestion to Asus: If you are not committed to support VT-d or AMD-v, just delete it from your documentation. This would solve many problems. But seeing that Asus has already done steps to support AMD-v and is working now on VT-d support for Ivybridge, why not be more clear about it?

    Here is my question to Asus: Which of the latest Asus Sabertooth X79 BIOS releases supports VT-d with the SB-E 3930K (or 3820K or 3960X) CPU? How was it tested?

    Now two question to you, Nodens:
    1. Do you have some insight into how Asus tests and documents features like VT-d? Because all I'm looking for is reliable information.

    2. More than a year ago I had a real bad experience with Asus tech support. Do you know of a contact at tech support (or elsewhere at Asus) who would be able to provide intelligent answers regarding this subject (VT-d for PCI passthrough)? Preferably someone with Linux knowledge. A PM would be fine, if you don't want to share it with the whole world.

    Good to find someone who actually knows about passthrough etc. By the way, if you need passthrough support for Nvidia cards, check out KVM. If you are not afraid of tinkering with the graphics card, there is a way on how to make regular Nvidia cards be recognized as Quadro etc. "MultiOS" cards - see my how-to, at the beginning. The latter should also work for Xen.
    Last edited by powerhouse; 01-05-2014 at 11:23 AM. Reason: incorrect statement

  6. #46
    Super Moderator - Coderat Array Nodens PC Specs
    Nodens PC Specs
    Laptop (Model)ASUS GL702VW-GC004T
    MotherboardRampage V Extreme Edition 10
    ProcessorIntel 3930K C2
    Memory (part number)CMZ16GX3M4X2133C11
    Graphics Card #1ASUS STRIX-GTX970-DC2OC-4GD5
    MonitorASUS VG278H
    Storage #13x OCZ Vertex 3 240 (RAID5 - LSI 9361-8i)
    Storage #23x WD2003FYYS (RAID5 - LSI 9361-8i)
    CPU CoolerCorsair H100i + 4x Noctua NF-P12
    CaseCoolermaster HAF-X
    Power SupplySeasonic X-1250
    Keyboard Razer Anansi
    Mouse Logitech G502
    OS Windows 10 Pro x64
    Network RouterLinksys E4200v2 (running TomatoUSB/Toastman firmware)
    Nodens's Avatar
    Join Date
    Mar 2012
    Reputation
    266
    Posts
    4,389

    Quote Originally Posted by powerhouse View Post
    VT-d is an advertised feature or at least mentioned in the manual, and I would hope that Asus tests VT-d as part of their standard Q/A process for new BIOS releases. Actually it's more about not knowing what's going on that worries me.

    1. Do you have some insight into how Asus tests and documents features like VT-d? Because all I'm looking for is reliable information.
    Of course it's tested as part of the Q/A process. As far as I know everything is tested under Q/A. I believe I already said that. But it's not singled out like no feature is. Until the point were something is reported as broken in which case it gets fixed. I do not have insight on how exactly everything is tested under the internal Q/A processes. No one does and no one will as no company shares that stuff with the public.

    I can live with Asus not supporting VT-d, there are other manufacturers that are more committed to it. Right now buying Asus is like playing roulette which is a little unnerving.
    Firstly I do not believe that there is any single manufacturer who is "committed to supporting VT-d". Not a single one. No one will single out a feature for special treatment unless it's something the entire market depends on and directly affects sales. VT-d is FAR AWAY from being such a feature. Overclocking is such a feature.
    Secondly I do not see how buying ASUS is like playing roulette either. As I said everything works with SB-E chips and soon the issue with IB-E chips will be fixed as well. ASUS is known for having the absolutely BEST UEFI/BIOS engineers available in the industry. If a feature gets broken, on one particular UEFI/BIOS version (or maybe only on one particular board), well guess what..it happens. Everywhere. On every company and not only with UEFI/BIOS but software as well. How many times have you seen Linux packages reintroduce a bug that was there a year ago (what's called a "regression")? You can't avoid these things, they are part of development and happen everywhere and with everything. Sometimes you patch up one issue and break something else without noticing. It happens. All that is needed is someone to report so it can be fixed. Things working 99.99% every time only happens on the enterprise/server market and that is for 2 major reasons:
    a) No features that introduce instability are ever used (eg no overclocking).
    b) There is no fast update cycle. Every UEFI/BIOS revision will stay under Q/A as long as it takes. The consumer market won't wait, they need a fix for a major issue or new feature support here and now. What consumer would wait 3 months until they can use IB-E chips at all on their X79 boards? No one. In server/enterprise market, they would not even consider using a new cpu until it's out for an absolute minimum of a year. This huge difference in mentality on the consumer market ("I want the best new shiny thing and I want it now") which what causes faster update cycles which directly translate into more chances of not all features working immediately as they should. But like I said this happens everywhere on the consumer market and if anyone thinks otherwise, I can only attribute it into lack of experience/knowledge.

    A suggestion to Asus: If you are not committed to support VT-d or AMD-v, just delete it from your documentation. This would solve many problems. But seeing that Asus has already done steps to support AMD-v and is working now on VT-d support for Ivybridge, why not be more clear about it?
    There is nothing to be more clear about. The feature is supported on the boards and if it's not working it's a bug that will get fixed.

    2. More than a year ago I had a real bad experience with Asus tech support. Do you know of a contact at tech support (or elsewhere at Asus) who would be able to provide intelligent answers regarding this subject (VT-d for PCI passthrough)? Preferably someone with Linux knowledge. A PM would be fine, if you don't want to share it with the whole world.
    ASUS is no different than any kind of tech support. Do not expect first tier of tech support to be able to answer in depth engineering questions etc etc or have any clue about Linux unless you're talking to a company that deals mostly in the server market (eg Supermicro, LSI Logic). All ASUS non-server boards do not officially support Linux. That much is certain. So do not expect tech support to know or care about Linux. Your best bet at troubleshooting anything Linux related on these boards is talking to me or Zygomorphic here..
    Now with SteamOS coming out, things may change a bit on this department.. but not until it catches on.

    Good to find someone who actually knows about passthrough etc. By the way, if you need passthrough support for Nvidia cards, check out KVM. If you are not afraid of tinkering with the graphics card, there is a way on how to make regular Nvidia cards be recognized as Quadro etc. "MultiOS" cards - see my how-to, at the beginning. The latter should also work for Xen.
    I know, I already use KVM on one of my systems but I'd prefer Xen support for Nvidia for other personal reasons. And I am aware of the mods but they are not an option for me as my hardware is a business asset and any mod messes up the resale value of outdated hardware which is factored in my business plan. Meaning if I mod it, apart from issues that may arise that can prevent me from working (or can make me chasing an issue with my code that may happen due to the mod), I'll also not be able to sell it as easily or at the same price as stock when I require an upgrade. Recycling old hardware on one way or another is a standard practice for us indies. Although I do tend to get sentimental with some pieces which end up in my hardware museum heh. For example when I'll replace my RIVE on my main dev system, I will not sell it for sure. It's the best motherboard I have ever owned (And I have owned A LOT:d). It will move to my Linux server box and when it's really outdated, it will end up in honorary spot on my hardware museum
    RAMPAGE Windows 8/7 UEFI Installation Guide - Patched OROM for TRIM in RAID - Patched UEFI GOP Updater Tool - ASUS OEM License Restorer
    There are 10 types of people in the world. Those who understand binary and those who don't!

    RealBench Developer.

  7. #47
    ROG Member Array
    Join Date
    Dec 2013
    Reputation
    10
    Posts
    8

    I bought a 4960X and the Rampage IV Extreme Black Edition - Getting Xen errors about IOMMU support as well.The UEFI on this board is horrible - I know it's more for gamers but I do Xen at work so I wanted to use Xen with IOMMU PCI for the system at home.

    From my Xen dmesg:

    (XEN) I/O virtualisation disabled

    Does anyone know if there are patches coming out for this bios as well to enable VT/d and IOMMU support functions for the Black Edition?
    Last edited by sixtyeightmk2; 12-21-2013 at 12:54 AM.

  8. #48
    ROG Member Array
    Join Date
    Dec 2013
    Reputation
    10
    Posts
    18

    @sixtyeightmk2: It seems like Asus is working on it - see Nodens posts here.

    @Nodens: Once again I appreciate your insightful reply!

    1. VT-d is tested as part of the Q/A process: I take this statement at face value. So if I were to report a bug with VT-d and Xen, Asus would go and fix the bug (assuming it is BIOS related)? I was under the impression that I had no recourse because Xen/Linux is "not officially supported" (see next point).

    2. Linux is not officially supported: Funny, that was the wording of Asus tech support reply one year ago. There are different ways to read this phrase:
    a. Linux may or may not work, we don't care. Look for someone else to help you.
    b. We do test with Linux to make sure things work, but don't provide any support nor guarantees.
    c. We have "unofficial" support - perhaps for some enterprise customers - on a per case or contract basis, but not for end users.
    d. How do you spell "Linux"? What are you talking about? Oh well, let's formulate the answer to look politically correct.

    I was wondering where Asus stands.

    3. Firstly I do not believe that there is any single manufacturer who is "committed to supporting VT-d": Well, ASRock has repeatedly shown their commitment to IOMMU support (perhaps not in all boards). But I agree that the only way to be sure is to get a server board from a known company specializing on servers, such as Supermicro etc. Intel is also very clear about VT-d support, both on their CPUs and chipsets and on their boards - see http://www.intel.com/support/motherb.../CS-030922.htm.
    If you follow the link above you see what I mean. It would be really great if Asus could compile a similar table for its boards. Essentially Intel doesn't support VT-d in all boards, but it clearly lists which ones are suitable.

    Because of the general lack of information on IOMMU support, I started a "VT-d / AMD-v motherboard compatibility thread" over at the overclock.net forum here: http://www.overclock.net/t/1338063/v...e-motherboards.

    Anyone reading this post and having had a success with VGA passthrough on any board / BIOS / CPU combination, please post a short report. I will also be following this forum thread, so you can post here as well and I will update the list.

  9. #49
    Super Moderator - Coderat Array Nodens PC Specs
    Nodens PC Specs
    Laptop (Model)ASUS GL702VW-GC004T
    MotherboardRampage V Extreme Edition 10
    ProcessorIntel 3930K C2
    Memory (part number)CMZ16GX3M4X2133C11
    Graphics Card #1ASUS STRIX-GTX970-DC2OC-4GD5
    MonitorASUS VG278H
    Storage #13x OCZ Vertex 3 240 (RAID5 - LSI 9361-8i)
    Storage #23x WD2003FYYS (RAID5 - LSI 9361-8i)
    CPU CoolerCorsair H100i + 4x Noctua NF-P12
    CaseCoolermaster HAF-X
    Power SupplySeasonic X-1250
    Keyboard Razer Anansi
    Mouse Logitech G502
    OS Windows 10 Pro x64
    Network RouterLinksys E4200v2 (running TomatoUSB/Toastman firmware)
    Nodens's Avatar
    Join Date
    Mar 2012
    Reputation
    266
    Posts
    4,389

    Quote Originally Posted by powerhouse View Post
    1. VT-d is tested as part of the Q/A process: I take this statement at face value. So if I were to report a bug with VT-d and Xen, Asus would go and fix the bug (assuming it is BIOS related)? I was under the impression that I had no recourse because Xen/Linux is "not officially supported" (see next point).
    Yes, unless the product has reached end of life. For current products it should get fixed. ROG boards are premium and get faster fixes and better support right here in this forum (if you post a reproducible bug here about a ROG board, be sure it will reach the proper ears in a timely fashion). Reporting that VT-d is not working is a bug report. Trying to troubleshoot it on Linux with tech support though falls under the "Linux is not supported" thing. More on this below.

    2. Linux is not officially supported: Funny, that was the wording of Asus tech support reply one year ago. There are different ways to read this phrase:
    a. Linux may or may not work, we don't care. Look for someone else to help you.
    b. We do test with Linux to make sure things work, but don't provide any support nor guarantees.
    c. We have "unofficial" support - perhaps for some enterprise customers - on a per case or contract basis, but not for end users.
    d. How do you spell "Linux"? What are you talking about? Oh well, let's formulate the answer to look politically correct.
    It's a combination of a and c. For example I know of a particular case where an enterprise customer had a UEFI fix for TSC timer syncing so they could use it under Linux with their application which required an absolutely stable TSC timer (TSC can be used under Linux as a solo timer source if it syncs properly..unlike Windows). All companies treat special customers with special support via direct channels. It only makes sense doesn't it? If the customer spends a 5 or 6 digit number on hardware per year that earns them some special privileges. That happens everywhere in the industry

    I'm pretty sure that all major OEMs would love nothing more than ditching Windows and the license fees involved with it. But as long as the consumer market is ruled by Windows, that's what they officially support as supporting anything else doesn't justify the cost. You may notice that ASUS has a line of netbooks that come with Ubuntu now. Those of course have Linux tech support
    Let's hope that SteamOS will catch on and along with the general distaste for Windows 8/RT will help change the face of the consumer market. It is a decent chance for the first time in decades

    3. Firstly I do not believe that there is any single manufacturer who is "committed to supporting VT-d": Well, ASRock has repeatedly shown their commitment to IOMMU support (perhaps not in all boards). But I agree that the only way to be sure is to get a server board from a known company specializing on servers, such as Supermicro etc. Intel is also very clear about VT-d support, both on their CPUs and chipsets and on their boards - see http://www.intel.com/support/motherb.../CS-030922.htm.
    If you follow the link above you see what I mean. It would be really great if Asus could compile a similar table for its boards. Essentially Intel doesn't support VT-d in all boards, but it clearly lists which ones are suitable.
    Unless there's something I don't know about ASRock, in which case please share, it doesn't seem to me that they do anything different than ASUS that would classify them as "committed to VT-d" support. Some boards work, some don't. Others have UEFI versions that it's broken and then fixed again. I don't see any difference. I would assume "committed to VT-d" translates to making sure it works for every single board and UEFI/BIOS release for these boards something I don't see happening.

    Intel boards should be the most stable in this regard because Intel does not go out of their way in order to feature pack a board. The board is the chipset when it comes to Intel and they're the designers of both chipsets and CPUs so that's what should always work. Although I have heard of a story where it was broken for a few BIOS versions even on an Intel board (in one of the Xen mailing lists if I remember right).

    Anyone reading this post and having had a success with VGA passthrough on any board / BIOS / CPU combination, please post a short report. I will also be following this forum thread, so you can post here as well and I will update the list.
    Well you can add all Rampage IV boards for SB-E chips, currently, and you can add the RIVE board with that beta version posted here for IB-E chips. I've set up all 3 boards with VT-d and SB-E chips for myself or clients of mine. Unfortunately I have not tested the rest of the X79 boards personally.

    EDIT: One small bit of information that came to mind. The ASRock high end X79 boards perform much slower than the ROG boards on video adapter passthrough due to the latency penalty incurred by the dual PLX PCIe switches these boards use compared to the ROG boards which use none. This makes the ROG boards the fastest ones available for this.
    Last edited by Nodens; 12-21-2013 at 10:12 PM.
    RAMPAGE Windows 8/7 UEFI Installation Guide - Patched OROM for TRIM in RAID - Patched UEFI GOP Updater Tool - ASUS OEM License Restorer
    There are 10 types of people in the world. Those who understand binary and those who don't!

    RealBench Developer.

  10. #50
    ROG Member Array
    Join Date
    Dec 2013
    Reputation
    10
    Posts
    8

    One question comes to mind though - the Rampage IV Extreme is different than the Rampace IV Extreme Black edition - do you know if there is a Beta Firmware available with IOMMU support?

    The system I built was going to actually run Xen from the get go - if I had realized the lack of IOMMU support I would have gone with a server board.

    Since I'm running a Black Edition with 64GB of ram and a 4960X I wouldn't mind being a guinea pig if it means I can get PCI Passthrough working!

    Can you confirm if that beta version would work on the Black Edition?
    Quote Originally Posted by Nodens View Post
    Yes, unless the product has reached end of life. For current products it should get fixed. ROG boards are premium and get faster fixes and better support right here in this forum (if you post a reproducible bug here about a ROG board, be sure it will reach the proper ears in a timely fashion). Reporting that VT-d is not working is a bug report. Trying to troubleshoot it on Linux with tech support though falls under the "Linux is not supported" thing. More on this below.



    It's a combination of a and c. For example I know of a particular case where an enterprise customer had a UEFI fix for TSC timer syncing so they could use it under Linux with their application which required an absolutely stable TSC timer (TSC can be used under Linux as a solo timer source if it syncs properly..unlike Windows). All companies treat special customers with special support via direct channels. It only makes sense doesn't it? If the customer spends a 5 or 6 digit number on hardware per year that earns them some special privileges. That happens everywhere in the industry

    I'm pretty sure that all major OEMs would love nothing more than ditching Windows and the license fees involved with it. But as long as the consumer market is ruled by Windows, that's what they officially support as supporting anything else doesn't justify the cost. You may notice that ASUS has a line of netbooks that come with Ubuntu now. Those of course have Linux tech support
    Let's hope that SteamOS will catch on and along with the general distaste for Windows 8/RT will help change the face of the consumer market. It is a decent chance for the first time in decades



    Unless there's something I don't know about ASRock, in which case please share, it doesn't seem to me that they do anything different than ASUS that would classify them as "committed to VT-d" support. Some boards work, some don't. Others have UEFI versions that it's broken and then fixed again. I don't see any difference. I would assume "committed to VT-d" translates to making sure it works for every single board and UEFI/BIOS release for these boards something I don't see happening.

    Intel boards should be the most stable in this regard because Intel does not go out of their way in order to feature pack a board. The board is the chipset when it comes to Intel and they're the designers of both chipsets and CPUs so that's what should always work. Although I have heard of a story where it was broken for a few BIOS versions even on an Intel board (in one of the Xen mailing lists if I remember right).



    Well you can add all Rampage IV boards for SB-E chips, currently, and you can add the RIVE board with that beta version posted here for IB-E chips. I've set up all 3 boards with VT-d and SB-E chips for myself or clients of mine. Unfortunately I have not tested the rest of the X79 boards personally.

    EDIT: One small bit of information that came to mind. The ASRock high end X79 boards perform much slower than the ROG boards on video adapter passthrough due to the latency penalty incurred by the dual PLX PCIe switches these boards use compared to the ROG boards which use none. This makes the ROG boards the fastest ones available for this.

Page 5 of 16 FirstFirst ... 3 4 5 6 7 15 ... LastLast

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •