Question Are power supplies important after the say $120 dollar range

WINTERLORD

Distinguished
Sep 20, 2008
1,805
19
19,815
Just wondered is a good evga, seasonic, super Flower an corsair pretty much all the same like a 120 dollar 850 watt vs a 200 850 watt?

Also iv looked everywhere can't seem to find a psu that has a at 3. 1 standered or one that came out fairly recent like 2023 am on a budget but don't mind spending g extra if is worth iy

I have a good power supply but is 600 watts so idk if ima run into an issue with that or not 7900x3d and a 4070super

So was looking around in meantime as would be nice if need to get a new one for it to be modern /recent. I really like the psu from superpower thay list 9 different protection mechanisms where as all the others list anywhere between 5 and 7 like ocp ucp opp ect also is the atx 3.1 spec important at all for a amd system read somewhere that's a spec for intel I'd assume is for amd to idk but listed intel
 

Zerk2012

Titan
Ambassador
Just wondered is a good evga, seasonic, super Flower an corsair pretty much all the same like a 120 dollar 850 watt vs a 200 850 watt?

Also iv looked everywhere can't seem to find a psu that has a at 3. 1 standered or one that came out fairly recent like 2023 am on a budget but don't mind spending g extra if is worth iy

I have a good power supply but is 600 watts so idk if ima run into an issue with that or not 7900x3d and a 4070super

So was looking around in meantime as would be nice if need to get a new one for it to be modern /recent. I really like the psu from superpower thay list 9 different protection mechanisms where as all the others list anywhere between 5 and 7 like ocp ucp opp ect also is the atx 3.1 spec important at all for a amd system read somewhere that's a spec for intel I'd assume is for amd to idk but listed intel
Depends on the exact model number of the PSU.
 

WINTERLORD

Distinguished
Sep 20, 2008
1,805
19
19,815
It's a evga 600gd was bout 100 isn't that old few years and the 2080 the gpu is replacing basicly used bout same wattage will prolly be fine

However I'm asking for. Future refrence cause I would be running it at its limit I would suppose during heavy gaming wich idk don't think is good for a psu to run at limit but anyways won't hurt for awhile I assume

But looking at new ones if I end up getting a 850 or 1000watt id prefer a modern 2023 or 2024 model. Only reason I settled for such a low wattage to begin with is I had a 8700k wich was real energy effecient and was on sale

There's so many psus out there it's hard to sort through them all I might go for budget I mean is only so much I can spend but finding a new model has been difficult I read the Tom's 2024 psu guide just wondered if higher end or better power supplies increase longevity of components

Now if ur asking make an model of wich ones I'm looking at heck idk I'm just browsing tryin to find a list of recent models 2023+ and preferably

ATX 3.1 standered
 
Last edited:

WINTERLORD

Distinguished
Sep 20, 2008
1,805
19
19,815
So you don't think the psu could hold back clocks by like 100 or 200mhz slightly without tripping the psu protections

I'm just wondering cause when looking in forums an such I notice my cpu wich doesn't run anywhere near hot max 69degrees water cooled but with all the correct settings Iv compared to others in oc forums seems my cpu runs 100mgz slower on the 3dvache ccd and the non 3dvache ccd tops out at 5150 wich I believe is much lower then it should be the way I have it setup copied tested cussed occt for stability to use curve optimizer an yet I don't get the clocks everyone else gets so I'm just wondering (was like this on my 2080 rtx to if I'm hitting the limit on my psu an it is holding clocks back a tad without tripping any protections

Not really looking to upgrade my psu ad is still fairly new but somthin I might do in the future when finances are better but am seriosly wondering if the psu is somehow holding back the clocks hmjust a tad
 

Eximo

Titan
Ambassador
ATX 3.0 and 3.1 PSUs specifically meet a standard for operating above their ratings for short periods. That is to prevent the random trips that older PSU were having with the fast boost behavior of modern CPUs and GPUs.

Every CPU is a little different, and if you are comparing yourself against people that are posting benchmarks, they are often overclocking. Better coolers, lower ambient temperatures, etc.''

Seeking maximum clock speeds doesn't always result in the best performance from chips. As you get closer to instability, recoverable errors will start to slow things down.
 

WINTERLORD

Distinguished
Sep 20, 2008
1,805
19
19,815
well i prolly should of refrased my initial question, im really wondering if it is at all possable for the psu to basicle enough power but be right on the edge of underpowered, and then power everything fine but when reaching its limit cut some power just enough to change the clockspeeds of the cpu without actuialy tripping the over power protection

if so then a new psu starts to sound more sensable only reason its a thought is because my psu is only a few years old maybe 2 at best so dont really want to waste if that makes sense
 

Eximo

Titan
Ambassador
It wouldn't be the PSU doing that, they are pretty straightforward devices. To some extent there will always be voltage droop under sudden load changes, but that is what the main capacitors, and board capacitors are for. If they get drained too much, then the system would just shut down for under voltage or under current (possibly over current if the voltage dropped rapidly)

Now if the voltage started to go down, it is possible for the CPU/GPU VRMs to 'notice' and start backing off to maintain the voltage. Which would result in lower clock speeds.

As a hypothetical if a really good PSU maintained a consistent 12.1 volts under all conditions, and a lesser PSU would droop down to 11.8 or something, then to maintain the same power, the current draw would go up. This would put further load on the PSU and VRMs, this would raise temperatures slightly, and more power would be needed to compensate for the increased inefficiency. VRM feedback loop would possibly make adjustments, and the CPU/GPU control loop might pull back and reduce consumption so the voltage pulls back up.

Of course that is just theory, to really see something like that you would have to get an oscilloscope/logic analyzer/SPI or I2C communications on everything and understand the actual behavior.