Graphics Card upgrade for M6800 ?

I have responded to your PM and we can continue the discussion in that "conversation".

 

Great work there mate! It's frontier guys like you and others that is not afraid to break stuff that is a great asset to NBR. Your work helps the community find out what works and what doesn't.

 

That sounds about right. Do you know if there is a similar cap on AC power? I've noticed that when running benchmarks, I'd cap at about ~210W from the wall, and every once in a while would get a "Unstable Voltage" limit on the GPU. My PS is a 240W, so it's running pretty close to max.

I'm wondering because once the GPU would fire up to about 99%-100%, my CPU would lock at 2.8GHz and wouldn't turbo anymore, well below the TDP limits.

I'm pretty sure I'll be able to flash it back into working order. My one concern is that the working ROM that I have was ripped with nvflash. Hopefully that is the full ROM.

Interestingly, the BIOS chip on the 1070 is a standard winbond chip, but I was looking at the 980M, and I can't find a BIOS chip anywhere on it.

 

I've heard 210W thrown around before. I'm not sure what the exact loss during AC/DC conversion is but I believe that the 240W refers to the maximum that can be drawn from the outlet, but after conversion loss, the system will end up with a lower amount.

The maximum power for an MXM card is supposed to be 100W. If the card uses no more than 100W is seems like there should be plenty of headroom to have the CPU running full speed (55W?) and other stuff like the drives, display, fans, etc. going... That is, if the system really can receive a full 210W DC power from the adapter. I don't have equipment to measure the power draw at the outlet; if you were measuring 210W on the AC side of the adapter then the system is probably receiving less than 200W.

I was looking at TechPowerUp and it shows the Quadro P5000 to have a TDP of 100W but the GeForce 1070 has 120W. I always wonder about the exact specs for some of these lesser-used cards in TechPowerUp's database, I've spotted errors in the mobile card specs there before. Anyway, when I was benching my P5000 it was reporting power throttling which I never got with the M5000M or K5000M, so it does seem to be pushing the power too the limit.

I hadn't tried loading both the CPU and GPU to see how that works yet, so I just did a test with 3DMark running in a loop (100% GPU load) and Prime95 running behind that. Just like you observed, my CPU can't clock up in this case, it was stuck bouncing between 2.1 and 2.6 GHz. As soon as I ended the 3DMark test, with Prime95 still running, the CPU went directly to max turbo speed.

So, at least the GPU gets power priority which I think is appropriate, it never clocked down while I had Prime95 running.

[Edit]
Tried Final Fantasy XV Benchmark which I had run in the past, but watched the system load this time...
This runs the GPU load up to a little over 90% and the CPU up to around 75% (all eight threads, good multi-threaded implementation). It runs visually smooth/great but the CPU/GPU situation is similar, GPU runs at a good turbo boost clock speed but the CPU is stuck at around 2.5 GHz at the top. The CPU might actually be the bottleneck here (why isn't the GPU load close to 100%?); it's normal for task manager to report below 100% when the CPU is actually fully busy if it hasn't engaged the turbo boost (I've seen this when capping the CPU speed for fan noise reasons in the past).

I also noticed a few days back that I'm only getting around 40 FPS in StarCraft II when there is a decent-sized fight going on-screen. I played with the graphics settings and nothing made any difference. Turns out that the CPU is the bottleneck here, too, it was pegging one core up to 100% while the GPU utilization was barely going above 20%. This game runs all of the core game logic in a single thread, though, it doesn't put multiple cores to good use. At least in this case, the GPU isn't running at full speed so the CPU is able to run up past 3.5 GHz.

 

Hi everybody.. I have been following the thread now a few weeks now. I’ve been trying to get a P5000 going in my M6800 during that time.. it’s been an emotional roller coaster ! I was wondering if you all might shed light on if I’m missing something?

I have a M6800 with the LVDS running Win 10. A 4810MQ cpu and recently updated BIOS to A25. 32Gb ram, 250Gb main SSD, and previously had the K4100m gpu. In Bios, fastboot it off (thourogh), and secure boot is off. Graphic switching is on.

I grabbed the vbios “iieeann” posted P5000v3.rom.. used NVFlash to load it.. which was sucessfull! It was recognized by the Device Manager, as well as the GPU-Z.. yea!

It seemed to be good to go. And I needed some sleep. I had not modded the .inf file yet.

So next morning I boot-up and .... and.... no boot!? No POST? The power button and HHD LED come on for a split second, the speaker makes a little pop, and the GPU fan revs for a second. Then I swear I heard it wisper "%\&+ you".

I switched back to the K4100, which booted.. but at the start up I get a screen saying the computer was shut down due to system becoming hotter than expected! #1004?? Keep in mind it’s been shut off all night. So it’s cold! I do the diagnostics thing and passed w/o an issue.

Out of shear curiosity I reinstalled the P5000 .. still no boot... then I start pushing buttons which I hit power.. oddly the 5 key and power booted it up!!? Wtf!

But after another shutdown it reverted to no boot.. I did monitor the temp during that time, never got above 45*C.

Any advice would be greatly appreciated...

Also, thanks! to all you smart people on here.. pretty neat to see people who are curious and then follow up with action.

 

What a pickle. I want to post something useful but I have no idea what to suggest here. "Conventional knowledge" is that if you have a P5000 with the vBIOS that you flashed then you should be good. (This from, we have like a whole two or three other people have tried it.) I'd suspect the motherboard, but if it is good with the K4100M installed, then it's gotta be either the GPU has some issue or some bizarre interaction between the motherboard and GPU.

Guess I'll ask some general questions...
Do you know what type of system your P5000 was pulled from? Do you know what vBIOS came on it (that might help answer the former question)? If you didn't save it, there should be a sticker on the card that has the vBIOS version labeled. I'm wondering if we can confirm that it matches mine or @DynamiteZerg's (which are different but both working) or if it is from a totally different source.

 

Well I was told it was new.. it was in a sealed pouch.

sticker says.. 86.04.3A.00.15

 

I'll add some pics... as it has a little different look than yours i see on page 21? Abd the error i get

Sorry its a bit messy as I haven't cleaned the thermal paste off yet.. not sure what to use.

 

Can't see your pics. My card on page 21 is a Quadro M5000M, not a Quadro P5000. You can see my P5000 pics in this post.

Your vBIOS matches the one that @DynamiteZerg had (judging from the first screenshot in this post). So it doesn't seem like a totally different card/source...

 

That sounds a lot nicer than "Guy who bricked his card."