Power consumption and efficiency in gaming: AMD vs Intel

[Tetras]

I have the RM750e what's the stats with that? I'd look it up myself,but you are clearly more knowledgeable than me.

hwbusters have a review of the newer model (2023) and it was pretty good, a couple of percent down on the RMx @ 2% load and slightly higher for being on, though that's for the 1000 watter so the lower ones might have less wastage.

The earlier model (2021) has a Cybenetics report and it got similar efficiency results to the RMx and around 5 watts for being on.

 

[ooStevo]

hwbusters have a review of the newer model (2023) and it was pretty good, a couple of percent down on the RMx @ 2% load and slightly higher for being on, though that's for the 1000 watter so the lower ones might have less wastage.

The earlier model (2021) has a Cybenetics report and it got similar efficiency results to the RMx and around 5 watts for being on.

I have the latest one,so that's good I assume? The fan is slightly audible which I was a little disappointed with,but I can live with it.

 

[Tetras]

so that's good I assume?

Yeah, it is fine.

 

[ooStevo]

It better be your the one that recommended me buy it lol. And the Tomahawk mobo
* The mobo is nice btw. Only wish the lights on it weren't so pathetically dim. I countered that with blue neon case fans. I'm Oliver twist goodnight

 

[Tetras]

It better be your the one that recommended me buy it lol. And the Tomahawk mobo

Recommend is a generous word, since I think you shared a list of 3, one of which was from a manufacturer with a history of exploding.

* The mobo is nice btw.

Cool!

I countered that with blue neon case fans. I'm Oliver twist goodnight

Nice!

 

[alec]

It may not be quite on topic. When I conteplated building a NAS best/cheapest/most power efficient solution for me is a second hand laptop + a USB3 HDD enclosure
No desktop build can beat a laptop in terms of idle power

 

[Blackjack Davy]

As much as I like his vids, he's such a douche sometimes... but he needs the clicks for the ad-rev.

intel stuff is super efficient when idle, it only gets hot and hungry when....(spoiler) you work it hard.

Salt-N-Pepa - Push It (Official Music Video)

REMASTERED IN HD!Official Music Video for Push It performed by Salt-N-Pepa.Listen now to "Push It"https://saltnpepa.lnk.to/PushItVDFollow Salt-N-Pepa:Faceboo...

www.youtube.com

So it only gets hot when it needs to actually do anything?! The copium is strong in this one.

 

[jigger]

So it only gets hot when it needs to actually do anything?! The copium is strong in this one.

The power of Atom cores sat idol. Like coiled springs ready to explode into action before getting cooked by the other cores.

 

[ooStevo]

It better be your the one that recommended me buy it lol. And the Toma

Recommend is a generous word, since I think you shared a list of 3, one of which was from a manufacturer with a history of exploding.


Cool!


Nice!

History of exploding lmao! Which one was that again?!

*Just remembered it was the Gigabyte one. You steered me away from it. Thank you kindly good Sir.

 

[Rroff]

It may not be quite on topic. When I conteplated building a NAS best/cheapest/most power efficient solution for me is a second hand laptop + a USB3 HDD enclosure
No desktop build can beat a laptop in terms of idle power

Some laptops have horrible idle consumption like ~30 watt idle! I've recently migrated my NAS setup to a mini PC powered by a J4125 processor - when all mechanical drives are spun down it draws sub 2 watt from the wall idle, unless Windows is being a pain and background activity keeping it higher which it far too often likes to do.

 

[jigger]

Some laptops have horrible idle consumption like ~30 watt idle! I've recently migrated my NAS setup to a mini PC powered by a J4125 processor - when all mechanical drives are spun down it draws sub 2 watt from the wall idle, unless Windows is being a pain and background activity keeping it higher which it far too often likes to do.

I’d question the resolution of your monitoring equipment at that point.

 

[Rroff]

I’d question the resolution of your monitoring equipment at that point.

The J4125 on the right mainboard is actually really power efficient and can be tuned down to about half the official idle power draw (the mini PC I'm using is also fanless so there is very little adding to the load when everything is fully idle).

I'm just using an old Zalman power monitor for measuring at the wall but it is accurate enough for long term averages, just lacks for ability to measure spikes accurately as it only updates at 2Hz.

 

[jigger]

The J4125 on the right mainboard is actually really power efficient and can be tuned down to about half the official idle power draw (the mini PC I'm using is also fanless so there is very little adding to the load when everything is fully idle).

I'm just using an old Zalman power monitor for measuring at the wall but it is accurate enough for long term averages, just lacks for ability to measure spikes accurately as it only updates at 2Hz.

I think it’s more than a little off TBH. Even spun down a few HDDs could easily pull over that alone.

 

[Rroff]

I think it’s more than a little off TBH. Even spun down a few HDDs could easily pull over that alone.

I'm using NVMEs mostly (SK Hynix Gold which has very low full idle power consumption in the mW range), with some Toshiba mechanical drives connected via USB for replication which pretty much go entirely to sleep when spun down - 0.6 watt typical active idle, but with the link down in the low mW range.

 

[jigger]

I'm using NVMEs mostly (SK Hynix Gold which has very low full idle power consumption in the mW range), with some Toshiba mechanical drives connected via USB for replication which pretty much go entirely to sleep when spun down - 0.6 watt typical active idle, but with the link down in the low mW range.

I think we can safely assume the 2 watt reading is way off.

 

[Rroff]

I think we can safely assume the 2 watt reading is way off.

I've put a fair bit of time dabbling with this including having known loads to test it against (end of the day wattage is just a product of voltage and current). And specifically choose the components for the purpose.

 

[jigger]

I've put a fair bit of time dabbling with this including having known loads to test it against (end of the day wattage is just a product of voltage and current). And specifically choose the components for the purpose.

I’m sure you have, but that reading is almost certainly wrong.

Assuming you was powering this system at mains voltage with zero conversion, it would still be over 2 watts of power use. 14nm quad cores aren’t that great for power efficiency and neither are the typical laptop brick PSU’s.

For me a 2 watt load at the mains would be 0.008097165991903 amps, so to get an accurate reading I’d need the ability to make accurate measurements to 7-8 decimal places.

Taking a bit of a guess, I’d hazard the Zalman is likely uncalibrated and measures at a resolution of 0.025-0.05.

 

[Rroff]

I’m sure you have, but that reading is almost certainly wrong.

Assuming you was powering this system at mains voltage with zero conversion, it would still be over 2 watts of power use. 14nm quad cores aren’t that great for power efficiency and neither are the typical laptop brick PSU’s.

For me a 2 watt load at the mains would be 0.008097165991903 amps, so to get an accurate reading I’d need the ability to make accurate measurements to 7-8 decimal places.

Taking a bit of a guess, I’d hazard the Zalman is likely uncalibrated and measures at a resolution of 0.025-0.05.

I'm not using a laptop here - it is a fanless mini PC, which comes with a fairly efficient wall wart style power adaptor, though I've swapped it out for a DIY UPS. I've tested the Zalman against known loads to be sure it is reasonably accurate.

 

[jigger]

I'm not using a laptop here - it is a fanless mini PC, which comes with a fairly efficient wall wart style power adaptor, though I've swapped it out for a DIY UPS. I've tested the Zalman against known loads to be sure it is reasonably accurate.

It’s not, It’s off by a factor of at least 3 and more likely a couple of factors of that. That system isn’t pulling 2 watts.

Adding a UPS into the mix would only increase the power draw. Going back to the external power supply would give better results.

 

[Rroff]

It’s not, It’s off by a factor of at least 3 and more likely a couple of factors of that. That system isn’t pulling 2 watts.

Adding a UPS into the mix would only increase the power draw. Going back to the external power supply would give better results.

It definitely isn't off by a factor of 3, I've tested the Zalman with known loads it is reasonable indicative even at low wattages.

The UPS isn't to improve efficiency just saying it is hooked up via that now - I'm using a DIY modified Eaton 3S Mini - I put quite a bit of time and testing into this setup including modifying the UPS with an ATTiny85 controller and writing custom software for it - part of which has been measuring the power draw under various different circumstances to make sure the UPS was suitable:

Spoiler

EDIT: PS we are talking about full idle wattage here where drives are spun down and the OS and hardware is in its lowest non-standby power state, not "active" idle which it is in most of the time, other than when unused for long periods, which is somewhere around 4-7 watt depending on factors like what Windows is doing.

EDIT2: Also if we were talking the N5105 which replaced it you'd be right even with a high efficiency power adapter that will pull around 7 watt from the wall with a lot of tweaking and nominally 10-12 watt idle which is an annoying step backwards.