NVIDIA
Nvidia Grid M40 Driver for Windows
Hi I was wondering if there is a Driver for the GRID M40 for windows. I understand Vdi is not possible but i wanted to use the card for compute for encoding h.264 Thanks
Hi

I was wondering if there is a Driver for the GRID M40 for windows.
I understand Vdi is not possible but i wanted to use the card for compute for encoding h.264

Thanks

#1
Posted 12/21/2016 05:39 AM   
You can try standard quadro/grid (universal) driver (from standard download location http://www.nvidia.com/Download/index.aspx?lang=en-us -> "GRID" -> "GRID Series" -> ... -> 376.33-quadro-grid-desktop-notebook-win8-win7-64bit-international-whql.exe) and try Codec SDK. ([b]Update:[/b] I look to nv_dispwi.inf but support for GRID M40 and Tesla M10 are missing in driver package.) [i]My opinion: NVidia deny existence of GRID M40 board (never design, build, sell or support, but check Ebay). HPE delete all traces about product with part numbers "J0X20A 796120-001 797638-001". New release of this card (with new double sized memory chip) is branded "Tesla M10" ([url]https://gridforums.nvidia.com/default/topic/823/grid-boards/tesla-m10/[/url]) and it differs in "pci subsystem id".[/i]
You can try standard quadro/grid (universal) driver (from standard download location http://www.nvidia.com/Download/index.aspx?lang=en-us -> "GRID" -> "GRID Series" -> ... -> 376.33-quadro-grid-desktop-notebook-win8-win7-64bit-international-whql.exe) and try Codec SDK. (Update: I look to nv_dispwi.inf but support for GRID M40 and Tesla M10 are missing in driver package.)

My opinion: NVidia deny existence of GRID M40 board (never design, build, sell or support, but check Ebay). HPE delete all traces about product with part numbers "J0X20A 796120-001 797638-001". New release of this card (with new double sized memory chip) is branded "Tesla M10" (https://gridforums.nvidia.com/default/topic/823/grid-boards/tesla-m10/) and it differs in "pci subsystem id".

#2
Posted 12/21/2016 07:52 AM   
I have tried every driver, There is a driver that works there was a seller on ebay who had a screenshot of the M40 being used in Windows. God knows which version though i tried to message with no luck. Yes the ignorance of Nvidia repeatedly saying there is no such card, The card works in Linux with standard drivers for Compute I'm in the process of install CentOS to see. Great card for compute on paper, Low power draw quad gpu in a dual slot form factor. Has great potential for Video Encoding IMO.
I have tried every driver, There is a driver that works there was a seller on ebay who had a screenshot of the M40 being used in Windows. God knows which version though i tried to message with no luck.

Yes the ignorance of Nvidia repeatedly saying there is no such card, The card works in Linux with standard drivers for Compute I'm in the process of install CentOS to see.

Great card for compute on paper, Low power draw quad gpu in a dual slot form factor. Has great potential for Video Encoding IMO.

#3
Posted 12/21/2016 10:58 AM   
Encoding performance of GM107 is not good (only one encoder engine and Maxwell Gen1) and it is unsuitable even for GPU H.264 accelerated 64xVDI (M10 variant, as marketing glorify). Rethink: [list] [.][b]1x GRID M40 == Tesla M10[/b] is 4xGM107 (Maxwell Gen1) with [b]one[/b] encoder engine - for encoding ~ 235 FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 4*235 ~ [color="green"][b]940 FPS[/b][/color] / 225W / 2 slots / 4*5=20 SMM = 2560 CUDA cores[/.] [.][b]1x Quadro M5000[/b] is GM204 (Maxwell Gen2) with [b]two[/b] encoder engines - for encoding ~ 361 FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 2*361 ~ [color="green"][b]722 FPS[/b][/color] / 150W / 2 slots / 16 SMM = 2048 CUDA cores[/.] [.][b]1x Quadro M4000[/b] is GM204 (Maxwell Gen2) with [b]two[/b] encoder engines - for encoding ~ 361*80% FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 2*361*80% ~ [color="green"][b]578 FPS[/b][/color] (-20% performance drop estimated due to underclocking) / 120W / 1 slot / 13 SMM = 1664 CUDA cores[/.] [.][b]2x Quadro M4000[/b] - 2*2*361*80% ~ [color="green"][b]1155 FPS[/b][/color] / 240W / 2 slots / 2*13 SMM = 3328 CUDA cores - [i][b]It seems to be better and supported solution (FPS +20%, power +7%, price +30% (Quadro M4000 (ebay/new) ~ 2x$650=$1300, GRID M40 (ebay) ~ $1000))[/b][/i][/.] [.][b]1x Quadro P5000[/b] is GP104 (Pascal) with [b]two[/b] encoder engines - for encoding ~ 535 FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 2*535 ~ [color="green"][b]1070 FPS[/b][/color] / 180W / 2 slots (+ 8k HEVC and 10bit encoding) / 20 SMM = 2560 CUDA cores[/.] [.][b]1x Tesla P4[/b] is GP104 (Pascal) with [b]two[/b] encoder engines - for encoding ~ 535*70% FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 2*535*70% ~ [color="green"][b]749 FPS[/b][/color] (-30% performance drop estimated due to underclocking) / 50-75W / 1 slot (+8k HEVC and 10bit encoding)(FPS -20%, power -66%, price +80%) / 20 SMP = 2560 CUDA cores[/.] [.][b]2x Tesla P4[/b] - 2*2*535*70% ~ [color="green"][b]1498 FPS[/b][/color] / 100-150W / 2 slots / 2*20 SMP = 5120 CUDA cores (FPS +60%, power -33%, price +260%)[/.] [/list] NVidia should disclosure more detailed encoder/decoder benchmarks (with dependency on clock speed (and burst) and memory bandwidth). Refs: [url]https://developer.nvidia.com/video-encode-decode-gpu-support-matrix[/url] [url]https://developer.nvidia.com/nvidia-video-codec-sdk#NVENCFeatures[/url] [url]https://developer.nvidia.com/nvenc-application-note[/url]
Encoding performance of GM107 is not good (only one encoder engine and Maxwell Gen1) and it is unsuitable even for GPU H.264 accelerated 64xVDI (M10 variant, as marketing glorify). Rethink:
  • 1x GRID M40 == Tesla M10 is 4xGM107 (Maxwell Gen1) with one encoder engine - for encoding ~ 235 FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 4*235 ~ 940 FPS / 225W / 2 slots / 4*5=20 SMM = 2560 CUDA cores
  • 1x Quadro M5000 is GM204 (Maxwell Gen2) with two encoder engines - for encoding ~ 361 FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 2*361 ~ 722 FPS / 150W / 2 slots / 16 SMM = 2048 CUDA cores
  • 1x Quadro M4000 is GM204 (Maxwell Gen2) with two encoder engines - for encoding ~ 361*80% FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 2*361*80% ~ 578 FPS (-20% performance drop estimated due to underclocking) / 120W / 1 slot / 13 SMM = 1664 CUDA cores
  • 2x Quadro M4000 - 2*2*361*80% ~ 1155 FPS / 240W / 2 slots / 2*13 SMM = 3328 CUDA cores - It seems to be better and supported solution (FPS +20%, power +7%, price +30% (Quadro M4000 (ebay/new) ~ 2x$650=$1300, GRID M40 (ebay) ~ $1000))
  • 1x Quadro P5000 is GP104 (Pascal) with two encoder engines - for encoding ~ 535 FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 2*535 ~ 1070 FPS / 180W / 2 slots (+ 8k HEVC and 10bit encoding) / 20 SMM = 2560 CUDA cores
  • 1x Tesla P4 is GP104 (Pascal) with two encoder engines - for encoding ~ 535*70% FPS (lowlat, highperf, 1920x1080, YUV4:2:0, 8 bit) - 2*535*70% ~ 749 FPS (-30% performance drop estimated due to underclocking) / 50-75W / 1 slot (+8k HEVC and 10bit encoding)(FPS -20%, power -66%, price +80%) / 20 SMP = 2560 CUDA cores
  • 2x Tesla P4 - 2*2*535*70% ~ 1498 FPS / 100-150W / 2 slots / 2*20 SMP = 5120 CUDA cores (FPS +60%, power -33%, price +260%)

NVidia should disclosure more detailed encoder/decoder benchmarks (with dependency on clock speed (and burst) and memory bandwidth).

Refs:
https://developer.nvidia.com/video-encode-decode-gpu-support-matrix
https://developer.nvidia.com/nvidia-video-codec-sdk#NVENCFeatures
https://developer.nvidia.com/nvenc-application-note

#4
Posted 12/21/2016 02:44 PM   
I agree with the above and thanks for the effort although at less than 50% the cost of the others listed it offers great value (I got it for $400us), Also max power draw under compute is 130w Total as documented at STH. Also the encoder we use is a little different to NVENCODE, There are effects that utilise cuda and Resolve scales well with multi gpu. Card is working on CentOS with Nvidia Drivers. Resolve is seeing all 4 cards. Now to get windows compute drivers??
I agree with the above and thanks for the effort although at less than 50% the cost of the others listed it offers great value (I got it for $400us), Also max power draw under compute is 130w Total as documented at STH.

Also the encoder we use is a little different to NVENCODE, There are effects that utilise cuda and Resolve scales well with multi gpu.

Card is working on CentOS with Nvidia Drivers. Resolve is seeing all 4 cards.

Now to get windows compute drivers??

#5
Posted 12/21/2016 10:48 PM   
[quote="ideabox"]Also the encoder we use is a little different to NVENCODE, There are effects that utilise cuda and Resolve scales well with multi gpu. [/quote] True, CUDA is another approach. (but other chips have more CUDA cores too :-)) [quote="ideabox"]Now to get windows compute drivers??[/quote] [i]My opinion: NVidia will never publish drivers for GRID M40 but it is Christmas time and NVidia can give you driver as a present. You can try to add simple support to nv_dispwi.inf in the meantime (I do not own the card. I cannot test it but it can work). There is "theoretical" example for 376.33-quadro-grid-desktop-notebook-win8-win7-64bit-international-whql.exe to add in the right inf sections: [code][NVIDIA_Devices.NTamd64.6.1] %NVIDIA_DEV.13BD.110A.10DE% = Section111, PCI\VEN_10DE&DEV_13BD&SUBSYS_110A10DE [NVIDIA_Devices.NTamd64.6.3] %NVIDIA_DEV.13BD.110A.10DE% = Section121, PCI\VEN_10DE&DEV_13BD&SUBSYS_110A10DE [NVIDIA_Devices.NTamd64.6.2] %NVIDIA_DEV.13BD.110A.10DE% = Section112, PCI\VEN_10DE&DEV_13BD&SUBSYS_110A10DE [Strings] NVIDIA_DEV.13BD.110A.10DE = "NVIDIA GRID M40" [/code]Merry Christmas and Happy New Year ! [/i]
ideabox said:Also the encoder we use is a little different to NVENCODE, There are effects that utilise cuda and Resolve scales well with multi gpu.

True, CUDA is another approach. (but other chips have more CUDA cores too :-))

ideabox said:Now to get windows compute drivers??

My opinion: NVidia will never publish drivers for GRID M40 but it is Christmas time and NVidia can give you driver as a present. You can try to add simple support to nv_dispwi.inf in the meantime (I do not own the card. I cannot test it but it can work). There is "theoretical" example for 376.33-quadro-grid-desktop-notebook-win8-win7-64bit-international-whql.exe to add in the right inf sections:
[NVIDIA_Devices.NTamd64.6.1]
%NVIDIA_DEV.13BD.110A.10DE% = Section111, PCI\VEN_10DE&DEV_13BD&SUBSYS_110A10DE
[NVIDIA_Devices.NTamd64.6.3]
%NVIDIA_DEV.13BD.110A.10DE% = Section121, PCI\VEN_10DE&DEV_13BD&SUBSYS_110A10DE
[NVIDIA_Devices.NTamd64.6.2]
%NVIDIA_DEV.13BD.110A.10DE% = Section112, PCI\VEN_10DE&DEV_13BD&SUBSYS_110A10DE
[Strings]
NVIDIA_DEV.13BD.110A.10DE = "NVIDIA GRID M40"
Merry Christmas and Happy New Year !

#6
Posted 12/22/2016 02:43 PM   
The M40 was never released as a board with supported drivers for GRID or encode. Sometimes we issue not-for-resale samples to OEMs to evaluate cards for certain uses with development drivers. Those third-parties are contracted not to resell such prototypes.
The M40 was never released as a board with supported drivers for GRID or encode. Sometimes we issue not-for-resale samples to OEMs to evaluate cards for certain uses with development drivers. Those third-parties are contracted not to resell such prototypes.

#7
Posted 01/03/2017 02:43 PM   
Ok, found - search for NVIDIA-Quadro-M4000-Windows-354.56.zip or NVIDIA-Quadro-K2200-Driver-Windows-362.27.zip .... (site:hpe.com). It contains [b]NVIDIA_DEV.13BD.110A.10DE = "NVIDIA GRID M40"[/b] in "Display.Driver/nviss.inf". [i](NVidia part-numbers for "samples": 900-52405-0300-000, 699-22405-0040-100).[/i]
Ok, found - search for NVIDIA-Quadro-M4000-Windows-354.56.zip or NVIDIA-Quadro-K2200-Driver-Windows-362.27.zip .... (site:hpe.com). It contains NVIDIA_DEV.13BD.110A.10DE = "NVIDIA GRID M40" in "Display.Driver/nviss.inf". (NVidia part-numbers for "samples": 900-52405-0300-000, 699-22405-0040-100).

#8
Posted 01/04/2017 01:20 PM   
Scroll To Top

Add Reply