New Build/ New GPU Bad Performance

tywalker_

Reputable
Jan 28, 2016
3
0
4,510
Hey guys this is my first time posting here. I have just bought a new MSI R9 390 gpu, and I installed it and everything ran just fine. I was getting very respectable frame rates then I came home from work to play it and I turned on MGS:5 and boom, 15-25 fps. Didn't matter if I changed the resolution to the lowest possible setting or changed any settings otherwise.

I also tried tomb raider the night before and ultra everything at 60fps constant. Again, drops to about 15-25 fps.

I tried uninstalling the drivers and reinstalling. I unplugged the psu and plugged back in. I removed the gpu and then put it back in. Finally at wits end, I put in my older GPU a gtx 760, and it was getting better framerates than the new card.

I forgot to add. I ran MSI Afterburner as well, card topping out at about ~75c. So I doubt that heat is an issue. It is idling at about 45c.

I keep forgetting to add more specific information. The weirdest thing it is doing, by far, is when I run the Tomb Raider benchmark, it's averaging at about 56 fps. I am running on a 1440 monitor, so that's why the framerate is a bit lower than what it should be for this card.

I'm not sure what else to try. I'd appreciate any help. Thanks.

Fractal Design Define R4
Asus A97m Pro4
G. Skill Ram 16gb
750w Bronze plus PSU
I7 4770s
 

SkyMembrane

Honorable
Dec 23, 2015
449
0
11,160
I see that you checked your GPU temperature, and it isn't too high for the card. Did you check your CPU temperature though? While it may have been good before, the 390 has a much higher TDP(Thermal Design Power), at 275W, compared to the GTX760 with 170W. This means more heat generated, hotter ambient air in the case and lower cooling capabilities for your CPU, which is located above the GPU. (Hot air is less dense, so it goes up).

The other thing is your PSU. I was going to rule it out because you had good FPS just before, but I won't. PSUs actual power output diminishes over time, for whatever reason the companies give. (It's because they aren't designed to last, but shhhht). I'd like to know the make and model of that particular PSU, but it's possible that if it was outputting 750W 2 or 3 years ago, now it's limit might be closer to 600W. If it happened to be a poorly rated model, like the CX series, it might have never actually been capable of outputting what the packaging said.

If your components are struggling to receive the correct amount of power they need, that'll throttle down the whole system, a lot.
 

tywalker_

Reputable
Jan 28, 2016
3
0
4,510
All components are new except for the CPU. Temps in MSI were similar to the GPU +/- a few degrees, maybe five at most.

My PSU is a CX version. Corsair CX 750M. Could the cables be the culprit?
 

SkyMembrane

Honorable
Dec 23, 2015
449
0
11,160


The problem is that Corsair doesn't make the CX, they use an OEM from Channel Well Technology, and it's cheap. Really cheap. In short, the OEM that Corsair used for that model used cheap parts and an iffy platform for its manufacture. The capacitors are mostly cheap not-so-good Chinese and Taiwanese versions, so honestly, it's very possible that your problem comes from a faulty PSU. The cables probably aren't the issue, but if you have other cables, changing them might be a good thing to try.

The CX series PSUs are regarded as very bad choices for any mid-high end builds, particularly when it comes to gaming, as it's very demanding in terms of power. The 760, which has a low TDP, ran fine, but your new R9 390 is probably more than it can handle at this point.

If nothing else is found to be the problem, a new PSU isn't THAT expensive.

The well ranked EVGA Supernova 650W G2 can be found for about 80$ on Newegg.
 

tywalker_

Reputable
Jan 28, 2016
3
0
4,510
I understand. So, though the box says, "Corsair CX750M" it is really a cheaper company that makes this PSU?

Would the best course of action be to run the card in a friends PC, and see what it does there?

I was just about to return the card, but seeing that it may be a PSU issue, I should probably test it more. Are there any tests I can run to isolate this issue otherwise?
 
Thermal throttling. 75°C is high. Try getting some case fans?

And don't listen to the guy who said CX is bad. CWT made a pretty decent power supply, and the capacitor (Capxcon) manufacturer is decent. It's used even on the widely acclaimed EVGA SuperNOVA G2 (on the modular board). CX750M isn't good for the budget you're at, but that doesn't mean it's outright bad.
 

SkyMembrane

Honorable
Dec 23, 2015
449
0
11,160


Trying the card in another computer with a more reliable PSU could be your best option. Verifying PSUs is kind of difficult, from personal experience. Installing it in another build that works with its own PSU and has a very similar or higher TDP would be the solution there.

You can never exclude the possibility of a defective GPU, though, to me, this PSU really isn't something I would recommend.

To answer your question, a lot of PSUs sold by big brands aren't made directly by them. Many will add their own modifications, like the cables, the casing and the fan, but many components remain the same. The same thing that applies to graphics cards. MSI buys base chipsets from Nvidia/AMD, then installs their own cooler/heatsink and then test the cards. They select the right ones and apply a factory overclock and sell the rest as normal cards. Not sure if that's the order they do it in, but it's something like that.
 

SkyMembrane

Honorable
Dec 23, 2015
449
0
11,160


Huh? 75C is high, but not enough to trigger thermal throttling on the CPU or the GPU

For his CPU,
4th Generation 22 nanometer: i7 4770K / i5 4670K (TDP 84W / Idle 2W)

Tcase (CPU temp) = 72C
Tjunction (Core temp) = 77C
Tj Max (Throttle temp) = 100C
Intel Temperature Guide
While his temps are higher than normal, they aren't even close to being hot enough to trigger throttling. (I do agree that adding case fans to increase the airflow would be a good idea, though. Maybe even an aftermarket cooler like an H7.)

And no way in hell that an R9 390 throttles when below 80C. Those cards aren't designed to run cool, and if he got 75C when testing with Afterburner, it's not what's causing issues.

The CX series is one of the most unreliable to ever be released by Corsair. While it works fine on low TDP builds, it's very possible that it's the cause of problems now that he has a card with over 50% more power consumption. I thought of thermal throttling at first, but if he's getting 75-80 on afterburner, it's too close to Tjunction recommended temperature to be an issue. For the GFX card, around 80C is usually where the GPU disables the turbo clock. He would see a dip in FPS, but not from 60 to 15-25.
 
CX isn't unreliable. If you try to load 1500W through a 750W PSU, EXPECT it to fail. If you put the CX750M through its paces with a simple i5 and a GTX 970, it should survive for a couple of years. Of course, again, this PSU is priced the same as a decent SeaSonic unit while performing at lower levels, and that's the only place it should get hell from. But you're mixing price/performance with simply performance, and that's where many make a mistake.

Again, this PSU isn't the best out there. It's not designed to be. It is designed to be functional, and to be left at that. And function it does well. It's just that Corsair expects you to pay a premium for a PSU that should cost around $50.
 

SkyMembrane

Honorable
Dec 23, 2015
449
0
11,160


Fine, my choice of word was incorrect. The PSU is fine for most low TDP builds, and the 750m is probably the best of the CX series. That being said, by process of elimination, let's look at what could be problematic.

[strike]The Motherboard[/strike] : Worked fine even the day before and doesn't seem to be warning of anything.
[strike]The CPU[/strike] : Again, it worked fine before the change and the OP doesn't mention anything about poor performances outside of games, which don't have the biggest load on CPUs.
[strike]The RAM[/strike] : Well it's ram...
[strike]Drivers[/strike] It worked the night before, and it would've had the same issues if the drivers weren't up-to-date.
The GPU : Could be defective and having issues, but the reviews of the R9 390 don't seem to question it's reliability. That being said, I don't exclude it.
The PSU : Not just the CX series, but PSUs in general are something to check when getting poor performances in games, because they don't necessarily cause crashes or warning if they don't output enough power. Not guaranteed to be an issue either, just one of the 2 things that seem to be going wrong.

I maintain my point, though, the temperatures mentioned by the OP for the GPU and later for the CPU aren't high enough to cause thermal throttling. However, if it goes higher than that in games like MGS:V, then it is likely to be the cause of the issue.