Precision 7530 & Precision 7730 owner's thread

I've tried that before, didn't ever think it's because of too much power being pulled, more like because the CPU can't handle some of the graphic card processes anymore (i think it still manages some but might be wrong). Now that i check the figures: CPU + Ram pulls between 140-155W according to Hwinfo, and thats unaffected, but the GPU Power fluctuates between 30-94 when using Furmark at the same time; the 3d animation is very choppy. I think I've encountered something similiar while using 3dmark and analyzing the scores, every time i did something to boost the CPU the graphics score went down which efectively mean that i can't go above 16500p. If you really think that's because of the total power draw i will buy one that's more powerful, there are no other hardware issues but now i think that even as it is performance is being limited.
https://allegro.pl/oferta/dell-da330pm111-org-zasilacz-19-5v-16-9a-330w-7283657059
will this one be OK? (its an auctioning service from my country, for those of you wondering, totaly safe to enter)
I don't want it to run on an absolute limit, whole system is probably not as reliable as it should be. Are you sure there aren't any other limiting factors when it comes to how much this laptop can pull from the wall (within reason of course). I did the quick research myself and it should work but I don't want to buy a new AC adapter just to find out that there's something else still limiting power draw.

 

No idea, really. There may be something internal that limits the amount of power that can be fed to various components so grabbing a larger PSU may not help at all. I don't think that anyone has checked this in detail on the 7530. I do think that a 240W would be fine, 330W is totally overkill; though, if 330W is cheaper or around the same price then there is no reason why it wouldn't work.

I actually have mine attached to a 240W power supply sometimes because we have a whole lot of those at work; with the Quadro P1000 though, I'm not really pushing the power to the max.

When attempting to fully load my M6700 (both CPU and GPU), I just ran either Furmark or 3DMark which would put the GPU to max utilization, and then Prime95 (at a lower priority) to put the CPU to max utilization. In this case I could see that the CPU clock speed would be reduced if the GPU was fully loaded. This didn't happen before the Quadro P5000 was installed (or if it did, it wasn't nearly as extreme).

 

Yeah this 330W is as cheap as 240W and i'm also pretty sure it's an original one (there are a lot of cheap fakes around as far as I know). What I hope for is that limiting comes from weak ac power supply switching itself off and not from some kind of limiting by the PC itself as you said. I'll probably have the power supply by Tuesday so I'll let you know how it worked out. I'm also quite surprised that no one else tried all of these mods in one machine, together with vBIOS they account to a pretty serious performance gain so I think it's a risk worth taking (if there is any).

Did Dell actually said something official yet about performance difference between what is essentially marketed as the same GPU? For me this issue is also quite a big one and what is even worse - probably most of 7530 users don't even know they were kind of scammed. A lot of time passed and as far as I know Dell chose to pretend that problem doesn't exist.
Some explanation wouldn't be inappriopriate.

 

I have not seen anything from Dell regarding this. I plan to bring it up when the next iteration of systems launch (there's a Dell guy who usually shows up here to answer questions).

 

Couldn't wait till Tuesday so I've borrowed bigger AC Power Adapter from a friend, re-run the tests and this time everything was fine, no power fluctuations, and mostly no restricted CPU frequency (some throttling, not much). I played with XTU settings during the test - this is why some of the editable values may appear weird. Now i think that this cooling system in 7530 is REALLY impressive and the i9 in this chassis makes some sense after all



edit


As i thought graphics were restricted too by the lack of power from supply, gained even more (a few hundred pts) in graphics test in 3mark11 - https://www.3dmark.com/3dm11/13225182 and this is stock of course since you can't OC these GPUs or maybe there's something i don't know about.

 

Test with new released Cinebench R20.

Nope, I don't think this can be correct.

 

Might I ask if you could benchmark with 3DMark Fire Strike? Thanks much.

 

Sure, no problem I'll do both R20 and 3DMark - 3Dmark maybe later since i have a really bad metered connection and never downloaded it before.
Now that i think of the power consumption i feel really stupid, of course CPU has a power limit of 60 watts and Gpu of about 75 (and i knew about it). How come then all of the problems disappear when i connect a 240W power supply? And there's a significant increase in performance way above what we could call a statistical error. I've connected mine (the 180W power supply) back to check if it all wasn't just by accident and I'm back to power fluctuations and decreased GPU performance. Can someone explain what is happening and what is more likely an actual power consumption of my system?


edit: R20: after some runs it averages about 2900.

 

Which benchmark do you see this effect under? If it's Furmark, that's practically expected, because Furmark is a highly unrealistic, GPU-destroying power virus. It's the GPU analogue of Prime95's 'Small FFTs' test: they are designed to put your GPU and CPU respectively under immense loads.

You could try some gaming benchmarks for a more... realistic comparison.

 

The easiest would be buy a Kill-A-Watt. Those are damn cheap. Aka you can measure max power consumption from the wall. Test with both power adapters. Can you see battery drain if you use the smaller psu?
Cinebench R20 + Heaven simultaneous is a better stress test - find max power consumption than P95 and Furmark. None should run Furmark on their notebooks. It's a Virus.

From your previous post... "Under synthetic load from XTU my i9 happily sits at 4.2GHz all the time (i got bored after 15m or so)". 2900 is more like 3.6-3.7GHz on all cores. Either you reach Power cap and or thermal throttling.