GTX480 to C2050 hack or unlocking TCC-mode on GeForce

[GottfrHW]

I’m not quite sure that i got the actual Status of modifying a GTX480/580 to be recognized as Tesla C2050/70/90 correctly.
Has there - besides Flashing the C2050 Bios onto the 480 and changing the Device PC ID - anything else to be done and
was anyone successfull in getting the so changed Card to work with full DP?

[cs0amc1]

I flashed my 590GTX (2x580) with the M2090 device id. The devices are showing in win7 64 as 2xTesla M2090s but one of them has a yellow exclaimation mark (device not working). However, the second device is now working in TCC mode.

[GottfrHW]

Good to hear. The main Q behind all that work is if a Non-Tesla card is - after whatever has been done - identified by the System as
Tesla-Card the DP-Performance has been increased as well. If it stays on the same Level then before (crippled to 1/8th of the comparable
Teslas for the 480/580/590s) for what all the work?

Would be great if the DP Perforrmance could be tested and the Results posted as well.

[seibert]

Good to hear. The main Q behind all that work is if a Non-Tesla card is - after whatever has been done - identified by the System as

Tesla-Card the DP-Performance has been increased as well. If it stays on the same Level then before (crippled to 1/8th of the comparable

Teslas for the 480/580/590s) for what all the work?

Would be great if the DP Perforrmance could be tested and the Results posted as well.

From what I’ve seen, none of the previous hacks have changed the DP performance because they were limited to changing the PCI-E identification of the card. Being able to use the TCC driver is a non-trivial reduction in kernel overhead on Windows 7, although Linux is low overhead all the time.

I doubt that anything will change DP performance short of a full firmware flash, and I give that 99% probability of bricking your GPU. :)

[sisyphus]

NVIDIA released a new driver 270.81, can you get 2 DMA engines with the new driver?

Good news, I was able to modify the straps so that my device shows up as 1091 (M2090). However, as there are no Windows drivers for the new cards, I cannot verify if anything has changed. Modifying the INF file for older drivers shows that nothing has really changed. (And I did not get 2 DMA engines, as the C2050 conversion that ijsfz got.

[zezinhocrack]

Hello friends,

It would be possible to use this technique to biosmod to release the 48 shaders GTX560, I’m trying biosmod but when I do the flash error “GPU override id mismatch” appears and I can not change the GTX560 (Non Ti)@ Ti, follows the screenshot:

http://img143.imageshack.us/img143/8746/testees.jpg

I need to be able to change the device id for the driver to recognize the missing shaders.

sorry for english

thanks

[lejeczek]

I flashed my 590GTX (2x580) with the M2090 device id. The devices are showing in win7 64 as 2xTesla M2090s but one of them has a yellow exclaimation mark (device not working). However, the second device is now working in TCC mode.

hi there,

apologies I bother you but was hoping you can help, I’m thinking of doing similar but with gtx580, another thing is I’ve not tampered with graphic cards this way before. I was hoping you could share with us novices some sort of a short howTo, simple instructions on how to flash a gtx into m2090?

if you get a chance to do that then many will be grateful!

thanks

lejeczek

[hktwinsen]

Could anyone can help me figure the soft straps from "06C0"gtx480 to"1080"gtx580.
I have a 512sp version of gtx480, and i already know alot of ppl had change the soft straps of 512sp gtx48 to gtx580.
Could anyone can help me figure it out what straps should i input. Thanks

[JMartinP]

Congratulations, you’re now running your secondary card in TCC! Don’t forget to make sure that both your cards are still functioning correctly though. Don’t forget to post a response!

Thank you for the awsome tutorial! I just managed to enable TCC on a GTX480 i had laying around. Have been stuck with Windows XP to avoid the WDDM overhead but now I can finally switch to Win7.

I’d really like to pull the same trick on the GTX580, which is my primary CUDA card. Did anyone have success running it as an M2090?

I still don’t fully understand how to figure out the correct soft straps but I’ll give it a go tonight.

[JMartinP]

Ok I think I got it. My device IDs are:

GTX580	1089	1000010001001

M2090	1091	1000010010001

So I just have to switch bit 3 to 0 and bit 4 to 1 using AND and OR mask 0 respectively (real noob posting here so sorry if I’m stating the obvious):

OR mask 0:   -xx1xxxx xxxxxxxx xxxxxxxx xxxxxxxx

AND mask 0:  -xxxxxxx xxxxxxxx xx0xxxxx xxxxxxxx

I am a little bit confused about the changed AND mask in the GTX480->C2050 case:

No, in my case I just took the GTX 480 AND mask and put the bits to 0 that would later be OR’ed to 1 (through the OR mask). As far as I can tell, the firmware takes whatever the strap values are, applies the AND mask, and then applies the OR mask, so for case 0:

(((hard straps 0) & AND mask 0) | OR mask 0)

OR’ing a bit would set it to 1 regardless of what’s in the AND mask, right? What is the reason for changing the AND mask at all in this case?

EDIT: My device id was 1080, not 1089, so the masks turned out differently but it works. Now running my GTX 580 in TCC mode.

[longboy]

Now it’s time to change the softstraps. Let’s first do a binary comparison of the two PCI Device IDs:

GTX480	06C0	0000011011000000

C2050	06D1	0000011011010001

I read out my AND/OR masks and compare them to the masks of the C2050 firmware. Turns out AND/OR 0 are different, but AND/OR 1 are identical:

Hex:      AND mask 0  OR mask 0     AND mask 1  OR mask 1     CHECKSUM!

C2050     0x6FFC03FF  0x10000400    0x7FF1FFFF  0x80020000

GTX480    0x7FFC3FFF  0x00004000    0x7FF1FFFF  0x80020000

                                                (ignore first bit! always set to zero)

Softstraps in the firmware are applied over hardstraps on the card, where the AND and OR masks control how the hardstraps are modified. Masks should always be below or equal to 0x7FFFFFFF. The mechanism works as follows:

( ( [hardstraps] & [AND mask] ) | [OR mask] ) = final straps

In practice, the AND mask allows you to disable certain hardstraps while the OR mask allows you to enable specific straps.

So far, I’ve managed to figure out the functionality of a few of the strap bits (any additional information is welcome!).

straps 0:

          -xx+xxxx xxxxxxxx xx++++xx xxxxxxxx    

             ^                ^^^^

             |                ||||-pci dev id[0]

             |                |||--pci dev id[1]

             |                ||---pci dev id[2]

             |                |----pci dev id[3]

             |---------------------pci dev id[4]

- cannot be set, always 0

So in my case, it’s just a matter of ensuring that bits 0 and 4 of the PCI Device ID in straps 0 are set to 1. I can do this by adding the appropriate bits to OR mask 0. I’ll take the original OR mask for my GTX 480 - which has a few other bits set to 1 as well for who knows what, I’ll keep these just to be sure - and enable the appropriate bits for the PCI Device ID:

OR mask 0:

GTX480	  -0000000 00000000 01000000 00000000

NEW       -0010000 00000000 01000100 00000000

Be sure to figure this out for your own card. If you’ve figured this out, you can use nvflash to apply the new masks:

(--straps [AND mask 0] [OR mask 0] [AND mask 1] [OR mask 1])

>nvflash --index=1 --straps 0x6FFC3BFF 0x10004400 0x7FF1FFFF 0x00020000

nvflash will directly change the softstraps on the card’s firmware. This means that if you save the firmware back to a file, you should be able to see that the softstraps at the appropriate locations in the firmware have changed (including the checksums at 00000068).

The secondary card should now contain the modified firmware + modified softstraps, but in order to see the changes, you’ll now have to reboot your computer. The next boot in Windows will likely result in the secondary card being detected as new hardware, after which Windows will attempt to install the “appropriate” drivers (never works for me). Make sure that you immediately re-install the latest NVIDIA drivers (you don’t need the special Tesla drivers) and do another reboot after the installation is complete.

Run nvflash again to verify that the PCI Device ID has indeed changed:

> nvflash -a

NVIDIA Firmware Update Utility (Version 5.95)

NVIDIA display adapters present in system:

<0> GeForce GTX 480      (10DE,06C0,10DE,075F) H:--:NRM B:02,PCI,D:00,F:00

<1> Tesla C2050          (10DE,06D1,10DE,075F) H:--:NRM B:03,PCI,D:00,F:00

If all went well, you should now be able to enable TCC by using the nvidia-smi tool located at C:\Program Files\NVIDIA Corporation\NVSMI. Note that with the 270.32 (CUDA 4.0rc) drivers, nvidia-smi is broken for me (known issue) and I had to grab the nvidia-smi tool from 263.06 (Tesla) in order to get things running:

> nvidia-smi -g 0

==============NVSMI LOG==============

Timestamp                       :  03/20/2011  12:53:07 AM

Driver Version                  : 270.32

GPU 0:

        Product Name            : Tesla C2050

        PCI Device/Vendor ID    : 6d110de

        PCI Location ID         : 0:3:0

        Board Serial            : 6182738065

        Display                 : Not connected

        Temperature             : 46 C

        Utilization

            GPU                 : 0%

            Memory              : 0%

...

> nvidia-smi -g 0 -dm 1

TCC enabled for device 0

Keep in mind that you may have to reboot in order for TCC to be enabled. If everything goes smoothly, your Tesla card should not be showing up in the NVIDIA Control Panel, and the deviceQuery SDK sample should be outputting the following information:

Device 1: "Tesla C2050"

  CUDA Driver Version:                           4.0

  CUDA Runtime Version:                          4.0

...

  Device is using TCC driver mode:               Yes

Congratulations, you’re now running your secondary card in TCC! Don’t forget to make sure that both your cards are still functioning correctly though. Don’t forget to post a response!

P.S. If you’re interested in helping out figure out the meaning of the strap bits… one way of doing this is by trial and error: set all strap bits to 0 by using the AND mask (0x00000000), then add the appropriate PCI Device ID bits through the OR mask to make sure your card is still detected properly, and try settings a single bit to 1, reboot, perform a test/query (CUDA-Z or deviceQuery) and see what changes, then start over again and continue with the next bit.

EDIT: Added instructions for working softstraps modifications. Added softstraps documentation.

I is a very good guide. But I am not familiar with these. So I can’t do sofstraps part. My device id and target id below. How will I change these to use with nvflash? like

(--straps [AND mask 0] [OR mask 0] [AND mask 1] [OR mask 1])

>nvflash --index=1 --straps 0x6FFC3BFF 0x10004400 0x7FF1FFFF 0x00020000

gts450 0dc4 0000 1101 1100 0100

quadro2000 0dd8 0000 1101 1101 1000

[power-000]

I’m from CHINA，
you can search a software，name：1.BOIS-ID（ONLY CAN CHANGE ID3 AND ID4），2.NVROM（not by nVidia，very very powerful！ DIVE ID any to any）

The Virtual goods has be sold in B2C，G84 can Change to GTS450 ,GF104 to GTX560Ti and so on External Image ， All the INFO you can change by you want,but only the info.

English is poor,I hope it useful to you.

[mouhcuda]

Hy everyone.
I’m new in CUDa GPU programming.
I have a question if someone can be kind and answer ti it.
I have compared a time execution of a kernel that i have optimsed. the comparaison is between a TESLA C2050 and a E5640 2,66 GHz processeur (a single core implemetation) and i have obtained a speed up around 4000 !!! I just want to know if it is possible or, there is a mistake (what i think but didnt found it yet).
Thanks very much for any answer External Image

Kind regards,
Mo

[hpcgeex]

hello all,
this is very interesting post! i do have little dought, is it possible to convert gtx 690 or gtx 680 gpus into a tesla performer gpus by such tweaks?

any help will be appreciated well!

[vacaloca]

This technique will only work if the chipset of the two GPUs in question is identical. Furthermore, the softstraps have to be changed to match the GPU ID as mentioned… if the bits are too different, you won’t be able to. Bear in mind this is very dangerous territory and can brick your card. I’d suggest going for a GTX Titan… it has the DP performance, it’s PCI-E 3.0… so it’s faster than a Tesla k20X for CUDA.

[Dredok]

vacaloca:

but if you use a Titan card without the TCC trick, then you must use linux / other windows platforms in order to avoid the WDDM overhead :/

is there an easy way to fully utilize a Titan and avoiding the WDDM limitations when running in win7?

[vacaloca]

@Dredok

Without BIOS modifications like this threads mentions (if it even works, GTX Titan BIOS is about twice as large as Tesla K20), basically, no. As I do not have a spare GTX Titan to recover from a possible BIOS brick, I’m not really willing to experiment.

[Dredok]

thanks vacaloca, I guess I should install it on WinXP 64bits for avoiding WDDM overheads :/ (I’m tied to MS)

[mvrk]

Hi,

I would like to transform my GTX480 into a Quadro 6000 so that i can test VMware vSGA.

Can someone help me out please?

[ilyals]

Yes, same question as above.
I eager to convert (fake) my GF580 to any of vSGA supported card to test the new technology.

Is there any posibility to do that?
And if someone owns a Tesla M2070-Q bios rom?