Search results

He did. Personally I wouldn't spend money on 5700, it's an upgrade sure, but barely. So unless he needs something like 2080/ti performance, and I do mean NEEDS, there shouldn't be a reason to buy anything lesser for a 0.5-1 worth of a GPU generation jump.

IMO for anything above 1600 mhz (or in other words above stock overclocking) will require adequate and gradual change of settings and NOT using someone elses settings which might or might not be stable for your card. When you say "the number 1 on his list" do you mean the first line of the list...

I haven't read the laws regarding that but will paraphrase Steve from GN. According to him the laws in the US and EU are that a customer has the right to service his GPU (and hardware in general) and repasting/repadding is part of that service. He even made a fuss when doing a tear down of some...

I was a bit surprised today to find a lack of disassembly guides for the vega 56 pulse. It should be straight forward, of course, but there is nothing specific, just a tear down of the vega nano pcb, disassembly of other Pulse model cards, nothing for the vega 56 specifically. Am I googling...

@5t3v0 from what I've gathered flashing gives little to no actual performance gain. Yes you could increase HBM frequency by 100mhz with the extra 0.1v but on reddit/here I read that the actual gains when it comes to gaming are very very little. Not that it's any significant risk, but it's hardly...

Yes, less power consumption and heat. Under full load/gaming this means you are less likely to hit the max temperature at which the card will thermal throttle (downclock itself to reduce heat).

3dmark firestrike/timespy and superposition are the usual suspects. The first two have tests in them that are more cpu heavy as well, but that won't skew the end result too much, so you'll still be able to compare to what other people post and determine if your card is running ok or faulty.

I upgraded my not-so-good PSU to a better one and got rid of the horrible coil whine. But I asked the local shop if I'd be able to return or swap for yet another if it didn't work. Didn't come to that but it's good to have assurance.

Question about temps: does the 85C thermal throttling limit apply to the measured Hotspot temperature (and others) as reported in HWinfo or the gpu temp as seen on Wattman? There's usually a >15C difference between the two and I've always adjusted fan curve for keeping hotspot (and others) under...

Didn't know, saw many reports of hynix cards capping at sub 900, like 850-860. But maybe later batches got better. Edit: siliconlottery... mine is Samsung capped at 940

It's probably samsung if he reached 955mhz stable. Also as per the lazy mans guide and many other sources that 950mv on the hbm is actually core voltage floor (minimum voltage for all performance states), hbm is hardlocked at 1250mv, that's why it's almost a must to overclock hbm to the maximum...

For 1080p 60hz it really is a bit of an overkill, but should you want to upgrade to 1440p/75hz or 1080p/144hz, this is the card. Too bad monitors aren't as cheap as we want them to be tho. My 144hz/1080p one cost 150 gbp.

The last 2 are attempts to post the same picture, uneditable for some reason. There is not much info left anyway, just a custom fan curve and +50% power limit. And as I said it's not a good idea to just copy settings but that's what he asked for. I mean, how high you can get on the HBM def...

Not sure if this is a good idea tho, but still...

I used the "Lazy mans guide to tweaking your GPU (rx vega 56 specific)" to get a general understanding of all the options, and what does what. The only difference is that I used Wattman insted of Overdrive tool, but the general guideline is the same.

Corsair CX550M. For me if I use Wattman interface only it also turns 0 rpm off and grays the button if I set a custom fan curve. But you can open the saved profile in notepad, there are many lines labelled "Feature ID =", the one that refers to the 0 rpm feature is Feature ID = "18". For this...

Yes, 2 fps below the refresh of the monitor (142). It wasn't that, same coil whine was present during benchmarks too (30-50 fps, 100% load). I'm pretty sure it was audible at 170-280W, louder at 220W, and very loud at 50%PL running 270-280W. Initially I thought it was only on high FPS too but...

I understand. That's why I specifically asked the local store if it was possible to return/swap if it wasn't ok after testing. They said it's totally fine so I went with it. But yeah, my old PSU with my old card was also dead silent, it's a matter of chance. Also I only decided to try this out...

I further identified that the noise(s) were coming from both the GPU and the PSU, almost similar types of buzzing/screeching, which initially caused me to believe it was only the GPU. Switching to a 750W seasonic platinum plus, and now both the card and the PSU are dead silent for coil whine...

950 Mhz is good, I also chanced Samsung but 940 is the max I can get. 945 starts artifacting, 950 crashing.