RTX 30 Series Discussion - Part 2

Gpu is unfortunately, and probably the most prone to failure part in a pc, from my own experience at least, had 8800gtx, 8800gt, 780ti sli (both failed) ati 5850, 6970 crossfire (one fan failed) all had miserable death, meanwhile I only had 2 ram sticks failed (design defect), 1 portable hdd failed, a motherboard fail, over the years.

I read that asus 3080 uses 6 mlcc, so was wondering did I make a bad choice, oh and also asus 3080 tuf was cheaper too from my place of purchase.

1 Like

Jesus hath returned. Or are the eyes a subtle reference to the god of thunder?

1 Like

This guy… Calling most other people openly stupid usually isn’t a sign of intelligence or at least wisdom (unless you are Einstein and make a joke).

Cards just shouldn’t crash at their default boost. Not everybody has or is willing to install OC tools to clock things down themselves, many just want and expect a system that runs. And imho a (slight) slashback might be healthy, particularly for Nvidia. Hopefully there is some learning effect that holding back drivers from OEMs to properly test the cards right until the last second means more trouble than good.


He seems like a giant a**jack… :upside_down_face:


Makes me laugh lol 2020

The model I’ve gone for should be over 20% faster in the workloads I’ll be mainly using it for (ultra high resolution VR @75hz). It was 19% over a 3080 in all results anyway (factory overclock), plus should take another 4% from a custom overclock. Plus you should discount 1080p results from comparisons for 8KX use which I haven’t even done from the review, and will be dragging down the score in generic benchmark tests like usermark.


Not sure about the channel but here’s some Strix 3090 (not OC it seems, but that shouldn’t change much anyway) benchmarks comparing to a (FE?) 2080Ti?

This (or the TUF) card is the one I’ll be going for once they’ve been beta tested by first-movers and are widely available… :slight_smile:

As mentioned earlier, I’ll be buying a 3090 because I know myself well enough to know that if I bought a 3080 I would replace it with a 3090 before the 4000 series are out anyway… :wink:


Let me comment a bit more on why I am sceptical about the current benchmarsks and how they extrapolate to VR.

  • The typical gaming benchmark uses one game and cranks up the resolution and/or the level of detail, special Fx, etc. All of these aspects affects mostly rasterizing part (pixel shaders). The FOV at which the game is rendered is fairly low (compared to the FOVs used in VR), so geometry perf is not that affected.

  • VR game renders much larger FOVs (at least in high FOV headsets, e.g. Pimax) and the geometry (vertex shader) performance becomes more important. The high FOV also stresses the CPU more, because it has to simulate and process more objects (which are visible).

  • The most demanding VR apps today are the simulators, which are pretty much CPU limited, for both the simulation and the rendering. I.e. they neither do manage to run the game nor feed the GPU.

The brute force which 3080 and 3090 bring may help, but I do not expect it to help as much as benchmarks show, especially as they seem to tend mostly the first point of the list above.

I am looking forward to tests for DCS, E:D, MSFS (once it is supporting VR), hopefully folks here, who are getting the new cards, will provide them soon.


Those are all interesting points, thank you for making them.

I agree that in sims that are CPU limited the gains may well be very slight, but thankfully I don’t play those sims.

For the rest, I guess time will tell. However I do want as much performance as is
available as VR as you say is just so demanding in a whole host of ways. X% here or there makes a very real and tangible difference in the experience.


Not sure if this has been shared (DigitalFoundry 3090 FE Review):

I like how he compares performance relative to various cards:

  • 3080 (9-16% difference)
  • 2080Ti
  • Titan RTX
  • 1080Ti (not shown but mentioned)



Yeah, on the Eurogamer site version you can select which cards you want to compare, at what resolution and it shows you the live replay of the benchmark run with the graph tracking them all. Really cool.


Nice. Didn’t even check as I’m surfing using the phone right now… :wink:

I’ve quoted You in my post… :+1::upside_down_face:

1 Like

Here are first 3090 bench’s compared with 2080ti oc on the 8kx


Jup. Still going for the ASUS ROG Strix 3090 (OC):upside_down_face:


Will probably go for 6900XT 16GB or 3080 20GB and ofc a Zen 3 CPU with 64GB memory.


I have the card right now,its not crashing at all… so dont worry