I have had a long struggle with my Nitro RX 590. It did not fix the problem that the card before it had, crashing during benchmark tests. I have been perplexed by the fact that it works fine on my upstairs build (win 8.1 on SATA SSD, Asus F2 A85 V Pro, A10-5800k apu, 1050 watt Thermaltake psu, 32 gigs DDR3 1600). And for those new to my saga, the R9-380 Strix handles itself perfectly on the downstairs computer and never crashes ever.
The Nitro does not work fine on my downstairs build (win 10 on M.2 NVMe installed on mobo, Seasonic 850W Gold (and now 1300 watt), asus X470 pro mobo, Ryzen 5 2600X cpu), which is much more serious hardware.
When the Nitro ran so well upstairs on the old build, I thought for sure it was due to the power supply. I ran the numbers on my downstairs build and came up with it being a bit short. So I got a big honkin' 1300 Watt Seasonic. I plugged it in greatly hoping that it would fix my crashing issues. But it didn't. So my girlfriend and I reviewed everything that was different between the two builds. We eliminated the power supply because now they both have big honking power supplies:
Upstairs: Win 8.1, Asus F2 A85 mobo, older more feeble cpu
Downstairs: Win 10, Asus X470 mobo, newer more capable cpu
And I said well it seems to me unlikely the cpu is the problem if I'm going to make a big new move on this problem it will either be to resurrect win 8.1 on the downstairs build, or to RMA the Asus motherboard. Or to live with it. I am very tired of working on this build and do not relish putting in a new mobo or installing a new OS.
She said: You forgot to put on your list that upstairs you run with a computer only a few feet from one screen, whereas downstairs you often run two screens, and one is further away (very big DVI-D cable, about twenty feet long). Very clever. Not for nothing is her degree in computer science. (but she doesn't do the kind of stuff that solves these kinds of problems)
So I said well that's easy to test. I have been using a variety of benchmarks (Superposition, Passmark, Heaven) and though all three of these can crash the system, and at less than maximum settings, Superposition hits the hardest and fastest, as a rule. And when you work on these issues you get tired of running long benchmarks. So we ran a bunch of Unigine Superposition in EXTREME setting and determined:
If you boot up PC with both screens and immediately try to run the two on Unigine, it's nearly 100% certain that you will get a crash immediately upon starting the program.
If you boot up PC with only the screen closest to the PC, which is connected via HDMI, Unigine Superposition doesn't seem to crash the system at all. With reboots, without reboots, with repeated runs at EXTREME setting.
If you boot up only the screen furthest from the PC, connected via DVI-D, the chances of crashing on Unigine are about 50-50. Sometimes you can do a series of Extreme tests and not have a crash, running only the distant screen.
When I boot the PC with the screen closest to it, I am pretty much mimicking the set up on the upstairs computer. It seems stable. One screen, close to the PC. And that is certainly the most reliable and I might even say it never fails, though I don't have the energy for a thousand boots.
When you add in the screen at distance, either by itself, or paired with the close-by screen, the PC is more unstable during benchmarking and crashes either always (two screens starting benchmark immediately after a reboot) or erratically (one screen at distance).
So now the question is:
(My girlfriend says I should stop running benchmarks.)
Greg N
The Nitro does not work fine on my downstairs build (win 10 on M.2 NVMe installed on mobo, Seasonic 850W Gold (and now 1300 watt), asus X470 pro mobo, Ryzen 5 2600X cpu), which is much more serious hardware.
When the Nitro ran so well upstairs on the old build, I thought for sure it was due to the power supply. I ran the numbers on my downstairs build and came up with it being a bit short. So I got a big honkin' 1300 Watt Seasonic. I plugged it in greatly hoping that it would fix my crashing issues. But it didn't. So my girlfriend and I reviewed everything that was different between the two builds. We eliminated the power supply because now they both have big honking power supplies:
Upstairs: Win 8.1, Asus F2 A85 mobo, older more feeble cpu
Downstairs: Win 10, Asus X470 mobo, newer more capable cpu
And I said well it seems to me unlikely the cpu is the problem if I'm going to make a big new move on this problem it will either be to resurrect win 8.1 on the downstairs build, or to RMA the Asus motherboard. Or to live with it. I am very tired of working on this build and do not relish putting in a new mobo or installing a new OS.
She said: You forgot to put on your list that upstairs you run with a computer only a few feet from one screen, whereas downstairs you often run two screens, and one is further away (very big DVI-D cable, about twenty feet long). Very clever. Not for nothing is her degree in computer science. (but she doesn't do the kind of stuff that solves these kinds of problems)
So I said well that's easy to test. I have been using a variety of benchmarks (Superposition, Passmark, Heaven) and though all three of these can crash the system, and at less than maximum settings, Superposition hits the hardest and fastest, as a rule. And when you work on these issues you get tired of running long benchmarks. So we ran a bunch of Unigine Superposition in EXTREME setting and determined:
If you boot up PC with both screens and immediately try to run the two on Unigine, it's nearly 100% certain that you will get a crash immediately upon starting the program.
If you boot up PC with only the screen closest to the PC, which is connected via HDMI, Unigine Superposition doesn't seem to crash the system at all. With reboots, without reboots, with repeated runs at EXTREME setting.
If you boot up only the screen furthest from the PC, connected via DVI-D, the chances of crashing on Unigine are about 50-50. Sometimes you can do a series of Extreme tests and not have a crash, running only the distant screen.
When I boot the PC with the screen closest to it, I am pretty much mimicking the set up on the upstairs computer. It seems stable. One screen, close to the PC. And that is certainly the most reliable and I might even say it never fails, though I don't have the energy for a thousand boots.
When you add in the screen at distance, either by itself, or paired with the close-by screen, the PC is more unstable during benchmarking and crashes either always (two screens starting benchmark immediately after a reboot) or erratically (one screen at distance).
So now the question is:
- RMA the card, it should be able to handle all these configurations reliably
- Live with the situation and consider the dual screen setup to be something of a limiting factor to the build. If you want to push the system hard, gaming or benchmarking, use only one screen right next to the pc). The card should not be expected to run two screens during brutal benchmark tests.
- Pursue some other kind of fix
(My girlfriend says I should stop running benchmarks.)
Greg N