Upgrading the DGFF GPUs in the Precision 7530 & 7730

I'm interested. Let me know when you have it posted to marketplace.

 

Do you mind sharing a GPU-Z of the card, i'm waiting for a RTX 4000 and i've noticed some odd inconsistency between the RTX 5000 Max Q (Which i presume is what the RTX 5000 you received is, based on the information that the TDP is 80W and information on this page: https://www.notebookcheck.net/NVIDIA-Quadro-RTX-5000-Laptop-Graphics-Card.423752.0.html
reporting the following:

Model -------------------------- TGP (W) -- Base (MHz) -- Boost (MHz)
Quadro RTX 5000 Desktop 230 ---------- 1620 ------------ 1815
Quadro RTX 5000 Mobile - 150 ---------- 1350 ------------ 1770
Quadro RTX 5000 Mobile - 110 ---------- 1035 ------------ 1545
Quadro RTX 5000 Max-Q -- 90 ------------ 930 ------------ 1455
Quadro RTX 5000 Max-Q -- 85 ------------ 765 ------------ 1395
Quadro RTX 5000 Max-Q -- 80 ------------ 600 ------------ 1350

Also i find it odd that the RTX4000 (normal version) has 213 Texture Mapping Units vs 192 of the RTX5000 Max Q ????

RTX4000

https://www.notebookcheck.net/MSI-W...h-workstation-with-powerful-GPU.425383.0.html

RTX5000

https://www.notebookcheck.net/Razer...ation-with-a-Lot-of-Performance.445767.0.html

Besides that the FS graphics score is also 10% below that of the Razer blade, which shows that the one in balde must have been running at higher boost for whatever reason (tdp/thermals)..

 

Here you go:


I think that 213 TMU count is a sensor/detection bug; look at the shaders box directly beneath. The RTX 4000 has 2560, whereas the RTX 5000 has 3072 (both of which are correct). As for the Fire Strike score... My GPU adheres extremely strictly to the 80 W power limit, whereas Razer's might be a bit more lenient. I'm not entirely sure.

 

Thanks ...
That would mean that perhaps there will be very little to no performance difference between the 4000 and 5000 if they are both capped to 80W... You do get 16GB of VRAM though

 

This is even the case on the 7740 which has both GPUs capped at 110W. Benchmarks around/less than 5% improvement with the RTX 5000 over the RTX 4000. (That's for an extra $1,000...)
Unless you need the extra vRAM, or you're going to run a load that needs for example tensor cores only (AI stuff) and won't hit the power limit by maxing them out, there's not much point to getting the RTX 5000.
(@Ionising_Radiation got it at a steal of a price...)

 

Technically, so have all of you

I've ordered a hardware programmer, and I want to push this to 150 W (I'd rather thermal throttling than power limit throttling). As it is, the GPU runs at something like 60-65 °C, and there's plenty of thermal headroom for more. It's a downright travesty that a GPU with 20% more cores performs nearly equally.

 

I have few questions related to pushing GPU over the limits set by the maker.

I assume that you will need a bigger power adapter that 180w one which comes with the 75x0 systems. Actually is there any chance that the power limits will go past 80 w just by getting a bigger brick? Could Dell have some kind of system in place which adjust GPU limits according to the power provided?

What is the limit of motherboard or is it irrelevant? Will it be able to supply more than intended?

From what I have understood RTX 5000 use better heat pipe system. Will it benefit also CPU when then GPU is not in the use?

 

Hi, does anyone here have vbios for Quadro P4200 ?

 

Hi everyone; my RTX 3000 is for sale here. Accepting bids

 

I can't be sure about that. I don't have a 240/330 W power adaptor to use, so I can't give a definite answer, unfortunately.

Also no idea. I can only answer these when I try the hardware programming.

This I can answer with a definite no. The CPU side of the heatsink is as meagre as the original, meaning > 95 °C when under full load, with plenty of thermal throttling. That said, being in a tropical country might also be a factor. The GPU runs much, much cooler (even under load), as already mentioned.