can a 550w seasonic gold psu handle a sapphire r9 390?
sapphire recommends a 750w psu. but psu calculators say that I should go for 400w-500w psus.
pls anyone
ok thank
>>51632620
You'll be fine.
I have a MSI 390x and my entire system (390x, 16 gigs ram, 4790k) has never gone over 450 watts load while gaming in 1080p. Fallout 4 on ultra makes me hover around 350-375 watts.
>>51633399
alright thanks
>>51632620
Pc parts picker is your friend
>Any AMD parts
>Less than 700w PSU
I just ran aida64 and checked the power draw, highest it spiked was 394w
>4790k
>R9 290
>Mpower Z97
Should be dandy
>>51635936
>Being this ignorant
>amd
>550w psu
>>51636161
Oh look, another retard who doesn't even bother to use google.
>>51635936
>me
>fx 6300
>r9 270
>430watt psu
>runs flawlessly
>>51632620
i'm using 2X 7850 with a 600w corsair
I have a 600w corsair 600m with r9 290 and i7 4770k, runs great
>>51632620
Word of advice, manufacturers always recommend much more than you actually need because they're envisioning the worst case scenario of some retard with a chink power supply bricking his system and blaming them.
somebody
call
This question really didn't need it's own thread.
>>51632620
you understand that the PSU doesn't power only the GPU right? What is the rest of your configurations? Any FANs? How many? FFS
>>51636188
There is a reason why I have a 750w psu.
>>51632620
I am not a robot
>>51638075
>peak load
Come back with average load instead - peak is pretty much irrelevant.
>>51638133
since they measure whole system power consumption with a cheap wall wattmeter no, it's not. "Peak" load is irrelevant for your PSU choice if by "peak" you mean the load spikes that only apply for fractions of a second (usuall less than a few ms) and are compensated by the secondary sides capacitors.
>>51632620
I'm using a 290x on 620w seasonic bronze so probably. My cpu and shit is oc'd as well have 16GB ram oc'd and hdd/ssd running as well as a bunch of shit plugged into every usb port
>>51638160
>"Peak" load is irrelevant for your PSU choice if by "peak" you mean the load spikes that only apply for fractions of a second (usuall less than a few ms)
Which is what I was indicating yes. Far too many people (both reviewers and /g/) cite those spikes as an indication of typical power draw of a card, which is not the case at all.
>>51632620
i'm running an overclocked R9 290, xeon e3 1230v3, generic other crap on 500W be quiet PSU since late 2013. Of course a good 500W PSU is more than enough for any hawaii card as long as you don't combine it with some other crazy powerhogging hardware (like an overclocked phenom II X6, FX 8 core, overclocked i7 5960X). The card will pull about 250W if you don't OC it. They recommend ridiculously overpowered PSUs so that your average user with his 15$ chiang liao whong PSU doesnt fry his PC.
>>51638252
yes, they arent directly related to average power draw, and a good PSU will have big enough capacitors on its secondary side to filter them out without dropping the voltage to shit. The pic you replied to doesnt show these peaks though, you actually need quite expensive equipment to even measure them.
the oscilloscopes tomshardware uses cost more than a "small car" (whatever igor wallossek means with that), and ht4u developed their own power analyzer with FPGA.
>>51638332
I do like to cite tomshardware for their gpu power draw testing. Take picture for example - most sites will claim the card pulls over 700w on its own which really is not true if one understand 1) how gpus work and 2) how psus work.
As I personally run a massively overclocked all AMD system I have a rather high quality 750w unit (super flower) and that is barely saturated for my current use, but it should serve me nearly forever (and some of the crazy builds i've got in mind for the future).
>>51632620
They reccomend much higher than needed because they factor that everyone will buy shitty discount psus like diablotek
>>51638459
i've never seen any site claim that peak values (the ones we talk about right now) are equal to power consumption. It gets a bit tricky with definitions from time to time, but most sites without the needed equipment have a different definition of "peak" loads. Like in highest sustained powerconsumption the card will reach in certain scenarios...
also i've never seen any site say that the 295 X2 consumes more than 500W. They test with shitty equipment and most of them list the consumption of the whole system, so if you have some high OC on your CPU you could reach ~700W with that card for your whole system...
>>51638532
>diablotek
>someone actually thought that would be a good name for a company shipping PSUs.
how retarded do you have to be to do something like that?
>>51638532
If you have seen /g/ build threads you'll realise why.
>>51638596
Hardocp do it, hence pic related. If you slash their supposed 95w claim off the number here that gpu is still pulling over twice its TDP.
>>51638641
well, that's about as retarded as you can go. Didn't they notice that the card would literally melt if it sucked up around 500W?
>>51638694
Possibly, but consider this:
>Hardocp
>>51636247
i had a chink psu.
piece of shit shat itself and took my hard drives with it when I plugged in my gtx 670.
>>51638641
hardOCP are confirmed nvidia shills, please don't post that shit here.
>>51638719
fuck them. they shilled nvidia and got butthurt when AMD refused them the nano.
>>51639210
Then this should be helpful.
>>51639269
Funnily enough for the fury release event AMD had they invited along one of the hardocp gang.
>>51639344
they did.
he didn't leave a positive impression though.
wasn't he bragging about how he hanged up on them when they told him the price for the nano?
>>51639344
is this an accurate representation or based on memes?
>>51639386
>wasn't he bragging about how he hanged up on them when they told him the price for the nano?
Kyle did. He also got so anally annihilated he deleted a nano build log on the hardocp forums because he didn't get a sample and a random pleb did (as AMD gave a bunch of nanos out for custom builds to showcase it).
This was the thread, but naturally it is long gone.
http://hardforum.com/showthread.php?p=1041770894#post1041770894
>>51639344
more of these tier chart?
>have 390x, 6700k, 16gb RAM, 4 hdd's, 1 ssd, and 3 case fans+2-fan AIO setup
>stock clocks under load I'm drawing 411w at the strip according to a Kill-a-Watt
>OC'd to 4.8ghz on CPU and 1185/1665@+63mv on GPU (can't remember what RAM is at, default XMP profile) it shoots up to 874w at the strip
>more than double the draw
This is why they recommend nearly double what a stock-clocked setup should need.
Corsair HX1000 from 8 years ago btw. Fucker's not even Energy Star rated, it's like a 60+tin PSU
>>51639488
Go read jonnyguru as a starting point - the chart is fiarly accurate. Seriously, anyone who wants to learn anything about power supplies should check out his site.
For lols his "death of a gutless wonder" are why youi should NEVER cheap out.
>>51639638
I don't have them, but i'm sure you could go to leddit and yoink them.
>>51633399
I can back this up. R9 290 on +50% power limit and OC'd pretty high, as well as a 4690K @ 4.3GHz (bad chip, sadly), drew just over 500W from the wall under 100% CPU+GPU load.
>>51639706
nice, thanks man!
I'm about to buy parts and assemble my first desktop PC after the finals. I really want to make sure the first experience is a good one.
So which 390 brand is the overall best? I know Sapphire is typically the go to brand, but there's a lot to choose from.
>tfw I almost bought a 970 for 5760x1080 gaming due to the sales.
>>51639776
I'll give you a hot tip: you can cheap out on nearly every part of a pc build to save dosh EXCEPT the psu - regardless of the wattage you need get a good unit.
Since i'm generous, in order or priority (most to least) this is how you should focus a build.
budget -> monitor resolution -> psu -> gpu -> cpu > everything else.
>>51639784
Whichever card has the best cooling, the top three 390's off the top of my head are sapphire's nitro, msi's gaming and powercolor's pcs+.
I personally like the look of the pcs+ a lot, but that is jsut me.
>>51639784
Sapphire is pretty highly regarded when it comes to AMD brands, can't really go wrong with them.
I've heard good things about MSI's coolers on the 390(X), they're fucking huge.
Asus typically develops their coolers for NVidia cards first, then swaps them over to AMD cards without changing anything, which sometimes causes it to not work as well as it should. As a result, Asus' R9 290(X) (and from what I've seen the 390(X) as well) run hot.
Haven't heard much about Gigabyte's 390(X) cards.
Basically, go with Sapphire.
Companies overestimate these kind of things all the time, not just relating GPU'S or electronics. It's to secure AMD's place with no liability in case something happens. I dub it "lawyer proofing".
>>51639804
Yea, I'm not really a fan of the red, but I've heard nothing but good things about MSI's. I'm coming from a Dual-X 280 though. Never heard much from powercooler.
>>51639824
I'd probably get Gigabyte's if they had the same triple fan set up they do on the 970, but the card just seems gimped with 2 when the 970 has 3. Call it brand loyalty or whatever, but my Sapphire card has done pretty well and is fairly quiet so I'll probably grab that whenever my refund for the 970 gets approved.
Friendly reminder to anyone in the market for an r9 390. Powercolor PCS+ model trades blows with Sapphire's Vapor-X model when it comes to cooling, noise, and overclocking headroom. Overclocking of course being dependent on the silicon lottery.
Also reminder that the dual 8 pins on Sapphire models don't guarantee any more overclocking than an 8+6 model.
>>51638075
Jesus the 390x and 980ti are such little power sluts
>>51639861
Yeah, when I first saw the dual-fan coolers on the Gigabyte cards I was pretty surprised.
Oh well, got a decent deal on a 290 Vapor-X, caught one just after the 390(X) launch happened so prices were going down fast.
>>51639871
Did Sapphire come out with a Vapor-X model 390 already? As far as I know they just have the 390 Nitro with or without backplate and the 390X Tri-X.
>>51639344
>TFW you just got a great deal on a Seasonic G-650 brand new, superior japanese capacitors
>>51639804
Thank you again, the priority list that I had in mind was:
(any 1080p 60hz monitor would do)
gpu=cpu->cooler->case->PSU-> everything else
I would have been horribly wrong.
>>51639907
I don't think they'll be releasing a TRIXXX or Vapor-X r9 390. Their nitro is like a bastard of the two so they're just sticking with one as far as I know.
>another Powercolor monster.
>>51639935
Powercolor makes some beastly cards, I'm really impressed that they managed to get an air-cooled DUAL 290X card working properly.
>>51639924
Yea I went the same way sans monitor. Which is why I have a CX600, which isn't as bad as a shitty off brand a lot of people get, but is pretty low end in the name brand arena.
>>51639877
To be fair they test the msi 390x which runs at eye-watering voltage out of the box to hit the staggering 1100mhz core clock it runs at.
>>51639824
>>51639907
>gigabyte
iirc their 300 cards are voltage locked which is an instant nyet from me.
>>51639924
Monitor dictates what gpu you need which in turn dictates what psu you need. given a gpu is the most power hungry part of a system its why I place a psu slightly above it in priority just so people become aware of how important it is.
>>51639935
>devil 13
Its cool, but pointless. Plus its not the Ares III
>>51639958
granted it easily got to 92-94C while benchmarking. But still. A monster. I have a Powercolor 290 and 390 and I can't recommend them enough.
Asus 390x or Asus 970?
>>51639986
For me it was between the Club3D RoyalAce (PCS+ rebrand with a different shroud) and Sapphire Vapor-X at the time. They were basically equally expensive, so I went with the one that was reportedly slightly quieter. No regrets.
>>51639974
>>51639971
I don't feel the urge to go over 1080p.
Should I choose a GTX 970 or a R9 390? (They are both around 400 Canadian $).
>>51639935
My theory is nitro will be the only flagship aircooler as an attempt to streamline cooler production with their zotac products.
>>51640029
If you don't plan on ever going past 1080p then the 970. It's the better card if you're not considering the excessive amount of VRAM the 390 has. If you want to get into 1440p or a triple monitor set up you'll need more than that 3.5GB of ram though.
>>51640029
The 390 is faster at 1080p as it is and pulls away at higher resolutions.
>>51640029
They perform about equal at stock speeds, the 390 drinks more power but it should last for a while longer, having 8GB VRAM and seeing how AMD's cards usually age better than NVidia's.
The 970 on the other hand runs cooler and has more overclocking potential.
>>51639871
Are the VRMs actively cooled on that model?
>>51640089
If it's just their 290(X) cooler with a different shroud, then yes.
>>51640089
Yes. If https://www.youtube.com/watch?v=DRFeWEXBvPE is the same cooler as the 390 version.
>>51640103
>>51640109
Neato, I was leaning towards MSI for that reason.
>>51640089
>>51640103
>>51640109
>>51640124
http://www.kitguru.net/components/graphic-cards/zardon/powercolor-radeon-r9-390-pcs-review/2/
>>51640124
Sapphire cools the VRMs too, I think. I think they even make contact with the main heatsink.
>>51640089
yes sir they are. VRM has a heatsink right on them. VRM1 I've only seen reach 80C while running 1.3v and benchmarking. High 60's low 70's at stock voltage and benchmarking.
>when I was breaking it down to put the morpheous cooler on.
>>51640058
>>51640065
>>51640087
It seems like R9 390 would be better for me. I would like my graphic card to stay with the growing hardware demand for longer.
By the way, I want to go for a mini-ITX build, will the smaller case make the heat of R9 390 a more relevant issue?
>>51640203
The 390 puts out a lot of heat, and most third-party coolers are more than two slots thick, as well as very long. Space and cooling could be an issue.
>>51640203
>will the smaller case make the heat of R9 390 a more relevant issue?
Depends on your 390 of choice - the sapphire card is roughly 3 miles long but small cases handle long cards easily these days.
>>51640213
It really depends on how much room your case has but the Powercolor and Sapphire cards are pretty damn long so it might limit your card options.
I wish I could understand the fascination of ITX builds.
>>51640196
>morpheus
Based cooler. Reminds me to re-attach the memory heatsinks on my morpheus as they keep falling off (due to my own fault).
Actually if I can be fucked I will file down an old heatsink I bought as part of a project that resulted in a dead cpu and mobo and mount that onto the memory chips.
>>51640254
I've done a mITX build, it was definitely interesting and a pretty cool idea to put some pretty potent gaming power in such a small footprint. I can see why people do it, I just like to have a big case because muh cooling, muh airflow, muh thermal headroom.
>>51640213
>>51640215
>>51640254
smallest case I know off that actuall cools well and can handle a really long card is the Cooler Master Elite 130. Pics usually show it handling a long card.
Pic related is radeon 7990.
>>51640288
I can see that. I just like my PC to be a piece of 'art' in my room so the idea of getting some super small case to hide it away is foreign to me.
>I just like to have a big case because muh cooling, muh airflow, muh thermal headroom.
Pretty much the reason I bought my Define S and filled it with 140mm fans.
>>51640259
the thermal adhesive that comes with the morpheous cooler is complete garabge. I bought some 3rd party adhesive.
Actually ended up going back to the stock PCS+ cooler because it's stock fan curve means the fans are off 90% of the time, and my r9 390 is a terrible overclocker so the extra thermal headroom wasn't necessary. If the bearings in my stock cooler starts to die, then I'll switch over.
>my r9 290 was a much better clocker. Pic related is 290 with morpheous and stock 390.
I honestly REALLY wish a 3rd party made a blower style cooler but with heavily re-worked internal sink to allow for much better cooling. I personally love reference coolers. Not to mention vastly better xfire cooling.
>>51640342
Also forgot, muh 5.25" bays.
Looking at replacing my current case with an Enthoo Luxe sometime in the future.
>>51639269
>hardOCP are confirmed nvidia shills, please don't post that shit here.
what a bullshit Roy...
>>51640382
If I was insane i'd source some delta fans and mount them on the morpheus just to see what they are capable of. Noctua's industrial units are not a viable alternative.
>>51640513
I actually have those CM fans being powered off the GPU board itself using a special mini 4 pin to dual 4 pin PWM splitter
Problem with deltas is they'll probably burn up the GPU board if I tried.
Never mind, I meant a micro-ITX. I thought mini-would be between micro-ITX and ATX but I'm wrong.
Honestly I just want a smaller case because I don't have a loot of room in my bedroom. I don't want to overclock nor set up SLI/Crossfire. Are there any benefit for me to choose ATX?
>>51640603
Usually it's harder to find mATX boards with the same load of features you'd find on a fullsize ATX board.
Some also have worse power delivery systems than their ATX counterparts, meaning lower overclocking potential.
Besides that, it's a nice form factor for people that don't want a full-on tower next to their desk but don't wanna give up future PCIe expansion options/5.25" bays/the possibility of SLI/Crossfire in the future.
>>51640586
>Problem with deltas is they'll probably burn up the GPU board if I tried.
They will - they draw well over what the fan header is capable of providing. Even now I have the SP120's (which i'm looking to replace for better performance) on my morpheus wired into my cases basic fan controller.
>>51640710
sounds like mATX would be a good middle ground
I guess I will go with mATX, thanks man!
You need a 2KW PSU for the R9 390.
>>51640792
I heard a 600 W is enough.
>>51640779
There's some really neat mATX cases on the market, you're gonna have fun picking one.
Good luck with the build!
>>51640866
Just saw some Intel promo for those who work at Intel affiliated stores, you can get a i7-4790K for like 100 bucks.
What a steal!
>>51632620
I also bought a Sapphire R9 390 and I'm using it with a FX-8320 and a 750W PSU.
whenever i play games that stress both of them it shutsdown.
I'm afraid i'll have to get a new psu
>>51641321
What brand PSU did you get? 750W should be plenty for that combo.
>>51641453
I have a OCZ Fatal1ty. It's about 4 years old at this point iirc.
>>51640087
Not him but what do you mean the 390 ages better if the 970 overclock better? Wouldn't the extra potential be good in the long run? Or do you mean the VRAM?
How important is having a backplate to avoid sag in a GPU? They don't sell the Sapphire R9 390 with backplate where I live. I'm torn between one without backplate or a MSI.
>>51641506
The 8GB VRAM is a big factor, allows you to step up to a higher resolution in the future. Also AMD always manages to squeeze a fair bit more performance out of the GPUs over time with driver updates. Look at the 280X/HD7970, that thing's old as hell for a graphics card but it's still relevant performance-wise.
>>51641489
Might be a good idea to at least try a different PSU and see if the same thing happens.
>>51641540
It's mostly just for looks, with PCB protection in second place and added rigidity in third. The 390 Nitro's cooler is mounted to the PCB in a lot of different spots so it shouldn't flex the PCB all that much. I do vastly prefer the looks of a backplate over a bare PCB though, but that's just personal preference.
>>51641573
I see. How far along is the PCB in the 390? Is it a refined rebrand of an older card?
>>51639661
but that could be because your psu efficiency is shit, the more your parts consume the more its going to pull from the wall
>>51641700
Comparing pictures of the 290 Tri-X PCB, 290 Vapor-X PCB and 390 Nitro PCB, I'm going to guess it's an all-new design.
>>51639344
>Corsair Hx series
>not Tier 1
>>51632620
what's SOC on a video card??
>>51638043
what gpu do you have?
>>51642488
not him, but looks like a 290x seeing as it's 4GB.
>>51643070
I agree probably a vapor/tri-x
