TheQuentincc

I figured I should share the mod (along the memory IC replacement with K4ZAF325BM-HC14 or HC16) needed to upgrade RTX 3000 to 12GB on Precision 7730 and 7550.

I could sustain 115W load in cool environement with my 7550 and 10875H (8 cores) with liquid metal. Nearly reaching 4.3GHz max all-core with undervolting. I'm using throttlestop to undervolt the CPU, force a more "relax" performance state (speedshift EPP 196) for browsing and disable turbo while I'm not needing any performance (such as browsing). When requiring higher performance I usually limit the CPU turbo to 3.6~4GHz on all core using the FIVR tab and push the speedshit EPP to 128 ("balanced mode" I would say). For gaming 3.2GHz seems to be the sweet spot (maybe since you have a 6 core you might need a bit more frequency). Regarding heat output the main thing to be careful is the dGPU, you should always monitor which app is using the dGPU (even at idle it's a constant 15w heat output, you can enable the activity icon through the Nvidia control panel). I also have an XPS 9500 with a far inferior cooling that I'm dailying with an I7 10750H, disabled turbo, undervolting and speedshift EPP at 196. It's doing wonder for daily usage without noticeable slow down and a fan that spun like 5~10% of the time every few minutes at slow speed depending on the weight of the web page.

Hi, I have a Dell Precision 7550 that's missing the mecanism that hold the Eternet cable in place, my bottom case was stripped down so I need the whole mecanism. I also have an XPS 9500 that's missing the plastic led diffuser in the bottom case and I would like to find another one without buying a whole new bottom case. Thanks

To make the nvidia dGPU power off you need to be pretty strict. I found out the best way to power off the dGPU is to enable the "show dGPU activity icon in the task bar" from the Nvidia panel/desktop menu, this way you can check which program is using the dGPU. Then in both nvidia panel and windows "graphics" settings you can force the program to run with the iGPU instead of the dGPU. If no program is using the dGPU then it goes to "sleep mode" and basically consume no power so you can profit of a great battery life. -> this doesn't work if the video output is directly connected to the dGPU

I'm here again with my 7550, which share the same chassis as the 7560. I'd like to know if anyone knows the TDP of the Quadro RTX 5000 for the 7550 ? and for the RTX A5000 of the 7560 ? Thanks you edit : I'm looking to replace my RTX 3000 core with an RTX 5000 core but I'd like to know if it's the same TDP (80W), otherwise I would need to populate the missing power phase.

GM204 doesn't really support H265 decoding, it's not efficient and would still load quite a bit the GPU. Sadly there is no MXM-A GM206 card so your stuck with Pascal for your needs. Also I'd like to point out the HD4000 does support partial H265 decoding. Intel made a drivers when H265 was released and I was able to decode 1080p H265 video with my HD4000 rather than my CPU (which was a time was a 3610QM), from CPU at 90%/35W to not even 10%/10w, I was using MPC-HC back then. I know that the main issue is app support, I was able to get it working on like VLC when it was working on MPC-HC...

No not really, for sure it can be done, but it's not like the "good old days" 20sec CPU swap. Regarding the initial subject the memory upgrade can be done only if the actual vbios supports higher density memory (also means if the actual vbios was made when x memory density was available). So it's pretty hard on GDDR5, however on GDDR6 nearly all vbios supports the Samsung 16Gbit memory, that's including lower SKU like TU117, so we can upgrade almost all GPU except those with 16Gbit memory IC (like RTX5000/newer GPU). I was wondering in which case would someone wants (and pay "that much") to double their memory. Right now I'm doing this cause I can resell workstation more easily/at slightly higher price.

Well, not all ivy bridge CPU IMC can handle the 2133MHz memory frequency, that might be your issue. Or maybe the XMP you choose wasn't the right one, all I can say is that 2x8GB 2133MHz kit from Hynix and TeamGroup (which is tighter) were working on my M6700. Early GDDR6 cards with Micron GDDR6 are very prone to be defective. Code 43 either needs new memory or core reball (in my exp).

Hi, I wanna know if any of you guys could be interested paying a service to upgrade/replace the GPU memory of your workstation/laptop so you could double the capacity ? For now I'm offering such upgrade (on local France forum) for like 150€ the 6GB to 12GB upgrade (192bits bus width) and 200€ for the 8GB to 16GB upgrade (256bits bus width). It's mostly aimed at lower SKU like Quadro RTX 3000/4000 (low/mid tier series) cause Quadro RTX 5000 (high end series) is already at his max capacity, it applies to Turing and Ampere generation (Ada is already using high density memory IC). It also apply to gaming cards but it's a lot harder to find the memory strap resistor/boardview. I don't know if there is any needs to double the memory capacity on portable workstation, 6/8GB should already be plenty for many uses cases but I don't know I upgraded the RTX3000 of my Precision 7730 to 12GB anyway 😄 So what's your thoughts ?

did you measured voltage on the coils ?

There is nothing preventing you from using the bigger "Nvidia" heatsink, also the heatsink seems very similar to the 5680 one so you might be able to buy a Precision 5680 Nvidia heatsink with vapor chamber for good cooling capabilities.

Did it worked on yours, do you have any proof of performance ?

you can dump your own bios using "bios backup" : Download Universal BIOS Backup Toolkit - MajorGeeks Then yeah it's the same procedure as recent bios like the 2nd link

alright, does the Precision 7750 RTX 3000 got higher TDP vbios ? maybe it's possible to use a 7750 vbios on 7740 ?

Is the RTX 3000 supposed to be limited at 80W on Precision 7740 ? mine is capped at 80W with vbios 90.06.39.00.64 (maybe it's the 7540 vbios ?), I thought it was supposed to be limited at 100~110W.