Grid 2.0

https://gridforums.nvidia.com/default/topic/599/hyper-v/



以下为原文




https://gridforums.nvidia.com/default/topic/599/hyper-v/

大家好，
这太荒谬了。
我们作为Citrix合作伙伴销售GRID Solutions，并希望使用GRID 2.0实施新安装。
问题是，没有经销商，硬件供应商，NVIDIA合作伙伴等可以告诉我们有关GRID 2.0的定价。
看起来营销一直比销售快。
我们在哪里可以获得有关GRID 2.0定价，许可证模型，分发等的可靠信息？
问候，马库斯



以下为原文

Hi All,

this is getting ridiculous.
We are selling GRID Solutions as a Citrix partner and want to implement new installations with GRID 2.0.
The problem is, no reseller, Hardware Vendor, NVIDIA Partner etc. can tell us about the pricing of GRID 2.0.
Looks like marketing has been faster than sales.

Where can we get reliable information about GRID 2.0 pricing, license model, distribution etc.?

Regards, Markus

大家好，
这太荒谬了。
我们作为Citrix合作伙伴销售GRID Solutions，并希望使用GRID 2.0实施新安装。
问题是，没有经销商，硬件供应商，NVIDIA合作伙伴等可以告诉我们有关GRID 2.0的定价。
看起来营销一直比销售快。
我们在哪里可以获得有关GRID 2.0定价，许可证模型，分发等的可靠信息？
问候，马库斯



以下为原文

Hi All,

this is getting ridiculous.
We are selling GRID Solutions as a Citrix partner and want to implement new installations with GRID 2.0.
The problem is, no reseller, Hardware Vendor, NVIDIA Partner etc. can tell us about the pricing of GRID 2.0.
Looks like marketing has been faster than sales.

Where can we get reliable information about GRID 2.0 pricing, license model, distribution etc.?

Regards, Markus

我认为这是一个“如果你独自在市场上就会发生什么”的案例。
关于定价，我同意Tobias的观点，我们无法用这些价格换成M60而不是K1 / 2。
现在的问题是K卡上会有EOL日期，然后当前的设置会发生什么？
我们是否需要提高当前客户的价格，还是托管合作伙伴需要提高价格？
我真的不明白如何在不添加“x2价格”的情况下说“x2性能”？
在我看来，如果定价加倍，你不会获得x2性能。
在这个版本中，我对NVIDIA感到非常失望。
重新思考，道歉并重新发布与现在相同的未经许可的模型的M卡。
只需我5美分。



以下为原文

I Think this is a case of "here's what happens if you are alone on a market" kind of issue.

Regarding the pricing I agree with Tobias here that there is no way that we can switch to M60 instead of K1/2 with these prices. The problem now is that there WILL be an EOL date on the K cards and what happens then with current set ups? Do we need to raise the prices for current customers or will the hosting partner need to take the hit?

I don't really understand how you can say "x2 performance" without adding "for x2 the price"? In my opinion you DON'T get x2 performance if the pricing doubles. I am rerally dissapointed with NVIDIA on this release.

Rethink, apologize and re-release the M cards with the same unlicensed model like you have today.

Just my 5 cents.

谁是您的Nvidia合作伙伴经理？
如果您有GRID 2.0系列的活跃机会，如果您是注册的NPN合作伙伴，您的PBM可以为您提供帮助。
我相信您知道，我们不直接销售，因此OEM和分销商设定街道定价，我们不提供建议零售价。
目前正在发运GRID 2.0的所有OEM都有硬件定价。
我最近与所有关键的OEM合作，为客户定价，但它仅适用于经过认证的机箱。
就在昨天，我有一个OEM客户经理告诉合作伙伴他们没有定价，并且必须从他们自己的价格表中发送详细信息，所以这可能是一个意识问题。
通过Ingram已经在NALA中分发（并且他们公布了他们的定价）
EMEA正在进行中，并且应该很快完成，但是EMEA中的一些OEM已经设置为直接或通过VAR销售GRID 2.0许可，有些已选择全球推迟软件推迟到自己的周期刷新。
该论坛不是讨论与OEM，分销商和VAR的商业关系的理想场所，正是这些组织为市场上的GRID 2.0组件设定了价格。



以下为原文



Who is your Nvidia Partner Manager? If you have active opportunities for the GRID 2.0 line, if you're a registered NPN Partner your PBM can assist you.

As I'm sure you're aware, we don't sell direct, so the OEM's and distribution set the street pricing, we don't provide an MSRP.

All the OEM's that are currently shipping GRID 2.0 have hardware pricing. I've worked with all the key OEM's recently to get pricing for customers, however it's only available for chassis's that are certified. Only yesterday I had an OEM account manager tell a partner they don't have pricing and had to send them the detail from their own price list, so it's possibly an awareness issue.

Distribution is in place already in NALA via Ingram (and they publish their pricing)
EMEA is in progress, and should be finalised shortly, however some of the OEM's in EMEA are already set up to sell GRID 2.0 licensing either directly or through VAR's, some have opted globally to delay software until a refresh in their own cycles.


This forum is not the ideal place to discuss your commercial relationships with OEM's, Distributors and VAR's and it is these organisations that set the price for the GRID 2.0 components in the market.

Grid 2.0是否支持Horizo​​n 6.1中的主机vmotion？



以下为原文

Does Grid 2.0 supports host vmotion in Horizon 6.1?

Grid 2.0是否支持Horizo​​n 6.1中的vmotion？



以下为原文

Does Grid 2.0 supports vmotion in Horizon 6.1?

出现了GRID2.0驱动程序的下一个问题。
新驱动程序（2015年8月之后）不允许降低最大显示尺寸和/或frl（即使对于K1 / K2）。
请参阅：https：//gridforums.nvidia.com/default/topic/258/nvidia-grid-vgpu/documentation-for-vgpu-configs/post/2087/#2087。
NVidia可以纠正这个问题吗？
M.C>



以下为原文

The next problem with GRID2.0 drivers appeared. New drivers (after 8/2015) does not allow to lower max display size and/or frl (even for K1/K2). See: https://gridforums.nvidia.com/default/topic/258/nvidia-grid-vgpu/documentation-for-vgpu-configs/post/2087/#2087.

Can NVidia correct this issue ?

M.C>

这从未得到过支持，因此新的解决方案堆栈按预期工作。
从不支持修改驱动程序，vGPU管理器或配置文件。
您可以通过Nvidia联系人（IBD，PBM或SA）提交功能增强请求。



以下为原文



This has never been supported, so the new solution stack is working as intended.

Modification of the drivers, vGPU manager or configuration files has never supported.

You could submit a feature enhancement request via you Nvidia contact (either IBD, PBM or SA).

不，这是我们正在与VMware合作的路线图项目。
这是一个极具挑战性的项目，因此短期内无法使用。



以下为原文



No, this is a roadmap item which we're working on with VMware.

It's an incredibly challenging project, and so it won't be available in the short term.

不，这是我们正在与VMware合作的路线图项目。
这是一个极具挑战性的项目，因此短期内无法使用。
同意，这肯定是一个重大挑战。
很高兴看到并行努力支持XenServer下的Xenmotion。



以下为原文



No, this is a roadmap item which we're working on with VMware.

It's an incredibly challenging project, and so it won't be available in the short term.
Agreed, it sure is a major challenge. It would be nice to see parallel efforts going in support for Xenmotion under XenServer, as well.

这确实是一个平行的努力。
我们都热衷于在虚拟机管理程序中运行动作。



以下为原文

It is indeed a parallel effort.

We're all keen to get motion working across the hypervisors.

这从未得到过支持，因此新的解决方案堆栈按预期工作。
从不支持修改驱动程序，vGPU管理器或配置文件。
您可以通过Nvidia联系人（IBD，PBM或SA）提交功能增强请求。
感谢您与我（我们）分享您的（NVidia）意见。
我真的很感激。
我有关于此的说明： 
所有虚拟化工作都基于专用（如GPU内存）或共享（如TDMA GPU共享）资源。
共享资源时，应该有一些共享的工具（限制）。
NVidia vGPU技术无法实现此仪器。
- 未能限制https://gridforums.nvidia.com/default/topic/258/nvidia-grid-vgpu/documentation-for-vgpu-configs/post/2087/#2087
- 未通电管理https://gridforums.nvidia.com/default/topic/378/performance-power-management-problem-on-shared-vgpu/ 
支持FRL的配置，甚至不像其他参数那样修改* .conf：
（ReleaseNotes＃31＃xe vm-param-set uuid = xxx platform：vgpu_extra_args =“frame_rate_limiter = 0”） 
向NVidia发送错误请求或功能请求对我来说是个问题。
- 我发送了CUDA技术的bug请求（nppiEncodeHuffmanScan_JPEG_8u16s_P3R（）的问题），响应时间为10个月。
- 忽略特征增强（如在Dom0或PV-DomU中启用CUDA（PAT / MTRR内存映射问题））。
-  NVidia网站充满了“不一致”，缺少许多信息 -  https://devtalk.nvidia.com/default/topic/546409/。
- 忽略对NVidia产品功能和限制的正确和经过验证的信息的直接电子邮件请求 -  https://devtalk.nvidia.com/default/topic/800942/gpu-accelerated-libraries/session-count-limitation-for-
nvenc-NO-麦克斯韦的GPU-与-2-nevenc-sessions- /后/ 4692848 /＃4692848
所以，我很悲观。
如果有人（大客户）有足够的能力提出增强请求（并且对IBD，PBM或SA的首字母缩写词有更多的了解:-)），我将非常高兴
M.C>



以下为原文


This has never been supported, so the new solution stack is working as intended.
Modification of the drivers, vGPU manager or configuration files has never supported.
You could submit a feature enhancement request via you Nvidia contact (either IBD, PBM or SA).


Thanks for sharing your (NVidia) opinion with me (us). I really appreciate this.

There are my notes about this:

• All virtualization efforts are based on dedicating (like GPU memory) or sharing (like TDMA GPU sharing) resources. When resource is shared there should be some instrumentation (throttling) of sharing. NVidia vGPU technology fails to implement this instrumentation.
  - fail to capping https://gridforums.nvidia.com/default/topic/258/nvidia-grid-vgpu/documentation-for-vgpu-configs/post/2087/#2087
  - fail to power management https://gridforums.nvidia.com/default/topic/378/performance-power-management-problem-on-shared-vgpu/
  
• Configuration of FRL was supported and even without modifying *.conf like other parameters:
  (ReleaseNotes #31 # xe vm-param-set uuid=xxx platform:vgpu_extra_args="frame_rate_limiter=0")
  
• Sending bug request or feature request to NVidia are problematic for me.
  - I sent bug request with CUDA technology (problems with nppiEncodeHuffmanScan_JPEG_8u16s_P3R()) and response time is 10 months.
  - Feature enhancement are ignored (like enable CUDA in Dom0 or PV-DomU (problems with PAT/MTRR memory mappings)).
  - The NVidia web is full of "inconsistencies" and many information missing -  https://devtalk.nvidia.com/default/topic/546409/.
  - Direct e-mail requests for correct and verified information of NVidia products function and restriction are ignored - https://devtalk.nvidia.com/default/topic/800942/gpu-accelerated-libraries/session-count-limitation-for-nvenc-no-maxwell-gpus-with-2-nevenc-sessions-/post/4692848/#4692848
  

So, I am pessimistic.
I will be very pleased if someone (big customer) has sufficient power to rise enhancement request (and is more knowledgeable about acronyms IBD, PBM or SA :-) )

M.C>

人们可能会担心，当K1和K2过时时，我们会被迫使用2.0许可模式。
对于一个人而言，我很难将这个商业模式包围起来，因为这是一个健全且计划周密的人。
我认为，@ nvidia的人应该非常小心，不要跟随像Oracle这样的公司的脚步，当谈到疯狂的许可政策时，那些模型只会让股东高兴，但肯定不会让客户满意。
那种生意并没有结交很多朋友。
GPU已经非常昂贵，如果还需要许可使用这种硬件，它可能会为大众消除GPU加速VDI。
当我们发布AMD多用户GPU时，我将更加关注AMD。
这个市场需要一些真正的竞争。



以下为原文

One could fear that when K1 and K2 are obsolete we are forced upon the 2.0 licensing model. I for one have a hard time wrapping my head around this business model as being sound and well planned. I think that the guys @ nvidia should be very careful not to follow in the footsteps of companies like Oracle when it comes to insane licensing policies, those kind of models only pleases shareholders but certainly not the customers. That kind of business does not make many friends.

The GPUs are already very expensive and if one further needs to license the use of this hardware it just might kill of GPU accelerated VDI for the masses.

I for one am going to take an even closer look at AMD when they release the AMD Multiuser GPU. This market needs some real competition.

同样在这里。
我正在使用Citrix HDX 3D为VDI环境准备PoC。
今天我拿到了M60卡的许可证费用：
GRID vPC永久许可证，1 CCU GRID SW 1（1），（3）375美元
GRID vWS永久许可证，1 CCU GRID SW 1（1），（3）$ 1.500
GRID vWS-Ext永久许可证，1 CCU GRID SW 1（1），（3）$ 6.000 
在Dollari的Stima costo supporto annuale obbligatorio：
GRID vPC基本SUMS第1年1CCU GRID SUMS 1（1），（2），（3）$ 75
GRID vWS基本SUMS第1年1CCU GRID SUMS 1（1），（2），（3）$ 300
GRID vWS-Ext Basic SUMS第1年1CCU GRID SUMS 1（1），（2），（3）$ 1.200
我的意思是。
30 vWS花了我45k + 9k / y subscrition ...........我的意思是..你认真吗????
与Nvidia的VDI对我来说已经死了。
取消NVIDIA所在的部分，继续推进AMD！
在这种情况下，我甚至不关心AMD多用户GPu的性能是否低于M60，因为54k我可以为每个用户提供两倍的硬件....



以下为原文

same here.
i was preparing a PoC for a VDI Environment with Citrix HDX 3D.
today i got the licence cost for M60 Cards:

GRID vPC Perpetual License, 1 CCU GRID SW 1 (1), (3) $ 375
GRID vWS Perpetual License, 1 CCU GRID SW 1 (1), (3) $ 1.500
GRID vWS-Ext Perpetual License, 1 CCU GRID SW 1 (1), (3) $ 6.000
            
Stima costo supporto annuale obbligatorio in Dollari:            
GRID vPC Basic SUMS 1st year 1CCU GRID SUMS 1 (1), (2), (3) $ 75
GRID vWS Basic SUMS 1st year 1CCU GRID SUMS 1 (1), (2), (3) $ 300
GRID vWS-Ext Basic SUMS 1st year 1CCU GRID SUMS 1 (1), (2), (3) $ 1.200

i mean. 30 vWS costs me 45k + 9k/y subscrition...........i mean..are you serious ????
VDI with Nvidia is dead for me. canceling the part where NVIDIA was in, going ahead with AMD!
i this case i dont even care if Performance of AMD Multiuser GPu is less of M60´s because with 54k i can put twice as much Hardware in for every user....

我同意ArminL！
nVidia正在杀死VDI！
有了这些价格，为什么要去VDI？
与真正的工作站一起使用会花费很少！
在价格相同或更低之前，但现在他们已经精神萎靡并想要杀死它。
我们已经完全部署了K1和K2的服务器，并且需要在明年刷新。
哦，好吧，它持续很好。



以下为原文

I agree with ArminL! nVidia is killing VDI! With prices like those, why go VDI? It will cost LESS to go with real workstations! Before it was par or lower price but now they've gone mental on prices and want to kill it. We had full deployments of servers with K1 and K2 and will need to refresh next year. Oh well, it was nice while it lasted.

我们有教育价格......还是太高了！
:(



以下为原文

And we had educational pricing... still too high! :(

教育价格似乎是许多供应商提供的标准40％折扣，包括许多产品购买的Citrix。
GRID 2.0许可证也有年度维护费用，因此许可证甚至不是永久性的。



以下为原文

Educational pricing appears to be the standard 40% discount that many vendors offer, including Citrix on many of its product purchases. There are also annual maintenance fees for GRID 2.0 licenses, as well, so a license is not even perpetual.

nVidia的价格高得离谱！
其实这次考虑去AMD了...



以下为原文

Those prices are ridiculously high nVidia! Actually considering going to AMD this time...

大家好，
看看M60 / GRID CCU牌照的价格，我们肯定会留在K1板......： -  /
因为我们还运行Linux桌面（VMware Horizo​​n），所以我在vGPU模式下添加了支持使用K1 / K2板的Linux桌面的投票（因为mcerveny在此线程的早期证明 - 不支持这只是一个纯粹的许可问题，它不是
技术问题）。
我们非常感谢的功能也是支持部署，具有“支持RDSH”的Linux服务器，支持3D，例如
能够将多个用户连接到1个Linux服务器，而每个用户可以同时运行3D应用程序。
问候，
spenatek



以下为原文

Hi all,

looking at the prices of M60/GRID CCU licenses, we will definitely stay with K1 boards... :-/

Because we also run Linux desktops (VMware Horizon), I am adding my vote for supporting Linux desktops with K1/K2 boards in vGPU mode (as mcerveny proved earlier in this thread - not supporting this is just a pure licensing issue, it is not a technical issue).

A feature really appreciated for us would also be a supported deployment with "RDSH-like" Linux server with 3D support e.g. the ability to connect several users to 1 Linux server, while every user could run 3D apps at the same time.

Regards,

spenatek